May 2000, Volume 6, Number 5: On The Cover

May 2000, Volume 6, Number 5
Cover Art By: Arthur Dugoni
ON THE COVER
6
On the Net
Parsing the Web Richard Phillips

In this tour de force, Mr Phillips describes the parsing rules for SGML,
HTML, and XML. He then shares and demonstrates three powerful Delphi
classes for programmatically gathering any Web information.
FEATURES
13
Delphi at Work
18
Dynamic Delphi
Delphi in the Office Ron Loewy

Step by step, Mr Loewy demonstrates how to construct Office 2000 addins with Delphi. Now you can use COM to extend the capabilities of your
favorite Microsoft Office tools, such as Word, Excel, and Outlook.
Dynamic Forms Ron Gray
Need to create a form or dialog box on the fly? How about adding new
controls even database controls at run time? The applications
are endless, and Mr Gray shows us how its done.
1 May 2000 Delphi Informant Magazine
23
The API CAlls
30
Greater Delphi
Active Directories Simon Murrell

Microsofts Active Directories is to directories what ADO is to databases.
Its also at the foundation of Windows 2000. All of which makes it
important to read Mr Murrells guide to ADSI for Delphi developers.
VisiBroker 3.3 for Delphi Eric Whipple
With VisiBroker 3.3, and its IDL2PAS compiler, Delphi can now
create CORBA clients based on existing interfaces, written in other
languages. Mr Whipple provides the details and demonstration
client and server applications.
DEPARTMENTS
2
5
34
36
Delphi Tools
Newsline
Best Practices by Clay Shannon
File | New by Alan C. Moore, Ph.D.
Delphi
T O O L S
FraserSoft Announces GenHelp

FraserSoft announced the
latest version of GenHelp, the
companys component Help

file generator. GenHelp can
New Products
and Solutions
import Pascal and C++ header

files to produce the basis for
component Help files. Help
files created by GenHelp
adhere to the look and feel of
Borlands Help files and integrate fully into the IDEs of
Delphi and C++Builder.
GenHelp employs a full graphical interface to speed up the
creation of Help files.
FraserSoft
Price: US$50
E-Mail: pete.fraser@frasersoft.clara.net
Web Site: http://www.frasersoft.clara.net/
genhelp
Extended Systems Releases RPM Server and Crystal Reports Driver for Advantage
Database Server
Extended Systems, Inc.
released RPM Server (Remote
Procedure Middleware) and
Crystal Reports Driver for
Advantage Database Server.
RPM Server is a middle-tier software solution that allows application developers to move intensive
database processing off the client
application and onto the database
server. RPM Server complements
Advantage Database Server, the
companys SQL client/server software, by acting as a workhorse
for thin clients. The combination
provides an alternative to traditional two-tier database applications, where client applications
perform much of the applications
custom database activity. RPM
Server offers a way to deploy
and maintain code at a single
point, providing greater stability
to the application and increased
efficiency for end users.

The distributed architecture
of RPM Server is designed to
allow future enhancements that
will provide connected middletier processing for mobile device
platforms, such as Palm OS and
Microsoft Windows CE. This
technology provides the framework for developers to expand
their business applications to
include the handheld computer
market. Rather than building
traditional two-tier client/server
database applications, developers
can build centralized business
rules and processes. RPM Server
enables developers using Delphi
to create true middleware applications.
Crystal Reports Driver for
Advantage Database Server provides developers with a native
interface for Seagate Softwares
reporting tool (works with Crystal

Reports Driver version 6 and version 7MR1).
The Advantage Crystal Reports
Driver utilizes the Advantage
StreamlineSQL engine to provide
access to FoxPro DBFs and
Advantage Database tables. With
the release of this new driver,
Extended Systems now offers
ODBC-free access to the Advantage Database Server through
Crystal Reports and moves all
SQL processing to the server.
This is a free tool available
to all Advantage developers and
requires Advantage Database
Server 5.6 ( version 2.6 client).
Extended Systems, Inc.
Price: Call for pricing.
Phone: (800) 235-7576 x5030
Web Site: http://www.
AdvantageDatabase.com
SkyLine Tools Imaging Releases Doc-to-Net 6.0

SkyLine Tools Imaging
announced Doc-to-Net 6.0, its
Internet document displaying
application and image conversion
tool. With Doc-to-Net, users can
send documents over the Internet
without downloading a plug-in.
This CGI application transforms
a scanned .tiff to a .png, .gif,
or .jpeg, anti-aliases it on the
fly, and streams it through the
browser. Added to this application
are zoom, pan, and scroll features,
as well as rotating and inverting
controls. Doc-to-Net offers scaling (resizing), has been engineered
to perform at high speed, and is
royalty free.
Features include a COM object
for ASP pages that will return the
number of pages in a multi-page
.tiff; the ability to recognize and
return sizes of varied size pages
within a multi-page .tiff; fewer
DLLs required; the ability to specify the size of an image; the UNC
path name feature, allowing files
on different computers within a
system to be accessed by the same
server; and password protection.
For photographic images, Docto-Net supports .bmp, .pcx, and
.tga, and presents them as .jpeg,
.png, or .gif files. Image correction
tools offered for documents are

also provided for images. The
Doc-to-Net software is Windowsbased and is recommended for use
with Windows NT and Internet
Information Server.
Developers who wish to utilize Doc-to-Net as a part of
an independent programming
application can leverage it as an
add-on to SkyLine Tools Corporate Suite.
Price: US$599
Phone: (800) 404-3832
Web Site: http://www.imagelib.com
Delphi
T O O L S
The Imaging Source Ships TX Text Control 7

The Imaging Source is shipping TX Text Control 7, its word
processing software, in reusable

component form. TX Text Con-
New Products
and Solutions
trol is a 32-bit programming

component that includes a lowlevel API and an OCX that
allows developers to add to their
applications sophisticated text
formatting and display capabilities typically seen in large word
processing programs. TX Text
Control 7 is royalty free; programs created with TX Text
Control can be shipped to an
unlimited number of customers
without any additional costs.
The Imaging Source
Price: From US$439 for TX Text Control
Standard.
Phone: (877) 898-2875
Web Site: http://www.textcontrol.com
Blinkinc Announces DeltaPatch 1.2
Blinkinc announced DeltaPatch 1.2, a multi-platform

change distribution system that
makes it possible to create a
small change or patch file
to update databases, documents,
and programs.
Instead of replacing an entire
database after modifying a handful
of records, or replacing a master
document after changing only a
few words, companies can e-mail a
patch file of the changes. Programmers can send a small patch file
of the changes to their software
instead of redistributing the entire
system every time they make
a minor update. Remote sales
people can download changes to
customer and stock databases
more frequently and in less time.

DeltaPatch also offers a solution to the problem of remotely
updating entry-level computer
users by creating self-applying
patches that can automatically
locate and update virtually any
type of file. Patch files will
only update existing users and
the Apply program may be
freely distributed, which makes
it possible to safely post patch
files on the Web for download.
DeltaPatch allows companies
to distribute product improvements, beta-test software, and
bug fixes.
DeltaPatchs Build program
uses an intelligent algorithm to
compare old versions of files
or directories to their new versions and create a patch file

of the differences. A wizard
guides users through a pointand-click process, and there is
no limit to the size or number
of files that can be compared.
The Apply program uses the
patch to update the destination
system to the latest version
under Windows NT/95/98/3.1
or DOS. The Apply program
can back up all files as they are
updated, so that application of
a patch can be reversed.

announced ImageLib
Combo@TheEdge 5.0, a programming tool that provides 18
VCL components that enable the
adding of images, image processing, and multimedia to applications in Windows 3.1/95/NT.
Museum-quality images can be
scanned, rotated, flipped, sized,
zoomed into, and converted to
.jpeg, .png, .tiff, .gif, .pcx, .bmp,
.ico, .wmf, .cms, and .scm. This
package includes 16- and 32-bit
versions. A VCL/DLL with source
code is available for Delphi developers.
Combo@TheEdge features
TWAIN support, a zoom tool,
flipping of images, text over
image, vertical and horizontal

scrolling text messages, a thumbnail manager, and a video frame
grabber.
Combo@theEdge provides
25 royalty-free flexible image
correction and manipulation
tools. Included are image cropping, scaling, and rotating, as
well as image manipulation
and special effects, including
mosaic, wave, ripple, and a
fisheye polar effect. Programmers can grayscale an image
or add and reduce colors,
brighten, sharpen, and increase
the contrast of an image.
Multimedia formats supported
are .avi, .mov, .rmi, .mid, and .wav
with BLOb support. Scrolling text
and vertical credit messages can

be put into an application, and
colors, fonts, and font sizes can be
manipulated.
Combo@TheEdge enables the
cutting, copying, and pasting of
images to and from the Windows
Clipboard and enables all image
formats with 1-, 8-, 16- and
24-bit dithering. Optional RAD
tool bars are available for image
manipulation.
(Note: Unisys may require a
royalty for applications developed
using gif and tiff LZW.)
Blinkinc
Price: US$299
Phone: (804) 784-2087
Web Site: http://www.blinkinc.com
SkyLine Tools Announces ImageLib Combo@TheEdge 5.0

Price: US$199
Phone: (800) 404-3832
Web Site: http://www.imagelib.com
Delphi
T O O L S
New Products
and Solutions
New Wave Software Offers SPI 2.5

New Wave Software, Inc.
announced SPI 2.5. SPI (Software
Piracy Intervention) adds a protective layer to software products
that allows only those end users
who legitimately purchased the
software to install it. No matter
how many copies of a protected
program get distributed illegally,
only the valid purchaser can
install the product.
SPI includes features such as passive and active encryption protec-
tion. SPI-protected products can

be distributed over the Internet,
on CD-ROM, or on diskette.
Another feature of SPI is that it
can remotely disable stolen prepackaged software; one phone call
to New Wave can disable one,
two, or thousands of stolen software products, rendering them
useless.
The version 2.5 upgrade adds
Locking General Purpose Registers, an Uninstaller feature,
and built-in e-commerce. The

Locking feature adds strength
to the six internal General
Purpose Registers by allowing
the developer the option of setting them to Read Only, Write
Only, Increment Only, Decrement Only, or Read/Write by
clicking a radio button.
The Uninstaller feature allows
the end user to move an SPI-protected product to another computer. The end user runs the SPI
Installer on the computer where
the product was first installed. SPI
disables the original installation
and resets the license, allowing the
end user to install the protected
product to another computer.
The built-in e-commerce feature
allows vendors to sell products
online without the use of shopping cart software; customers
can simply download a product
and select Purchase when they
run the SPI Installer. A dialog
box appears where the end user
enters their credit card information, which is encrypted and
transferred to CyberCash or
Authorize.Net for real-time
authorization. Upon authorization, the SPI-protected product is
installed and ready to run.
New Wave Software, Inc.
Price: Call or visit Web site for pricing
information.
Phone: (800) 920-9283
Web Site: http://www.nwspi.com
Quest Announces Schema Manager 3.0 and Data Manager 3.0

Quest Software, Inc.
announced Schema Manager 3.0
and Data Manager 3.0. Schema
Manager is a comprehensive
solution that allows users to
create, track, and deploy schema
changes throughout an application lifecycle. A new feature
quickly compares multiple schemas and database objects for
large applications and ERP systems, such as Oracle applications, eliminating guesswork
when applying patches.
Data Manager allows users
to accurately move and subset
data between databases for testing and reporting. Important
new features include support
for SQLoader, an import/export
method available for data migra4 May 2000 Delphi Informant Magazine
tions, and support for SQL

Server, enabling DBAs to migrate
Oracle data to SQL Server for
simplified data manipulation.
Quest Software, Inc.

Price: Call for pricing.
Phone: (949) 754-8000
Web Site: http://www.quest.com
News
L
May 2000
Delphi 5 Update Pack Available

An Update Pack for Delphi 5
is now available from the Inprise/Borland Web site at http://www.
inprise.com/devsupport/delphi/
downloads/index.html#Delphi 5.
Customers have other options to
get the Delphi 5 Update Pack.
US customers can obtain the
update CD by calling (800)
457-9527. US customers can
also order the CD online at
http://shop.borland.com/shop/
proddtl/0,1070,1-523,00.html.
The Delphi 5 Update Pack CD will
be shipped for US$6.95 (shipping
charge) beginning at the end of
March. The update CD contains
all updates, including the ADOExpress Update and the most recent
BDE update, BDE 5.1.1.
To order the Delphi 5 Update
Pack CD outside the US, please
contact the Inprise office that
serves your country.
Informant Communications Group Launches ComputerBookstore.com

Elk Grove, CA Informant
Communications Group, Inc.
(ICG) announced the opening
of ComputerBookstore.com, a
full-service technical bookstore
catering to IT professionals.
ComputerBookstore.com features a wide variety of computing, gaming, certification, training, science, programming, and
business books from all the major
book publishers, including IDG,
Macmillan, SYBEX, Osborne/McGraw-Hill, Coriolis Group,
Wrox Press, Warner, Microsoft
Press, John Wiley, AddisonWesley, Prentice Hall, and
Random House. In addition,
ComputerBookstore.com will
also make it easier to locate
many hard-to-find titles from
smaller, independent book publishers, such as Wordware, Manning, and Informant Press.
Computer professionals and
novices alike are looking for a comprehensive, competitively-priced
source for IT-related books and
training materials, said Mitchell
Koulouris, president and CEO
of Informant Communications
Group. ComputerBookstore.com
is the perfect resource for experts
in information technology, programming, database development, and related technologies
and industries.
In addition to the vast selection of books, training materials, videos, and documentation,
ComputerBookstore.com guarantees a savings of up to 41per-
cent off the suggested retail

price for featured on-sale items.
ComputerBookstore.com offers
a focused shopping experience
that places the emphasis on
technical books and training
materials without the noise and
distraction of non-related items.
Shoppers on the World Wide
Web have become increasingly
selective about the places they
shop, said Koulouris.
ComputerBookstore.com is
the best place to shop on the
Web for technical books and
related materials whether youre
a CIO at a major corporation
or a beginner just getting up to
speed on your home computer.
For more information, visit
http://www.ComputerBookstore.com.
Corel and Inprise Merger to Create Linux Powerhouse

New York, NY Corel
Corp. and Inprise/Borland Corp.
announced theyve entered into
a definitive merger agreement.
Upon completion of the merger,
the combined organization, called
Corel, will be a Linux powerhouse, offering a single source
for end-to-end solutions, featuring a range of productivity applications, development tools, and
professional services for all major
platforms. The valuation for the
entire transaction is approximately US$2.44 billion.
In 1999, the two companies
had total revenues of US$418
million and currently have over
US$200 million in cash. The
merger will be accounted for as a
purchase transaction under Canadian GAAP and is expected to be
accretive to Corels cash earnings
per share before the amortization
of goodwill.
Upon completion of the merger,
Inprise/Borland will operate as
a wholly-owned subsidiary of
Corel. Dr Michael Cowpland will
remain as president, CEO, and a
director of the corporation. Dale
Fuller, Inprise/Borlands interim
president and CEO, will be
appointed as chairman of Corels
board of directors. The operations
of the combined entity will be
headquartered in Ottawa, with
the Inprise/Borland operations
remaining in its current Silicon
Valley locations. The combined
businesses will have a presence in

over 100 countries.
Under the terms of the
agreement, Inprise/Borland shareholders will receive 0.747 Corel
common shares for each share
of Inprise/Borland common stock
held. As a result of the merger,
Corel expects to issue approximately 53.7 million common
shares in the aggregate, in
exchange for Inprise/Borlands
outstanding shares.
Based on the closing price of
US$20.00 per share of Corel as
of February 4, 2000, this represents a value of US$14.94 per
share of Inprise/Borland, giving
a US$2.44 billion valuation for
the entire transaction, on a fully
diluted basis. Upon closing of
the transaction, Inprise/Borland
shareholders will own approximately 44 percent of Corel,
with the balance being held
by Corels current shareholders.
The boards of directors of both
companies have approved the
transaction.
The merger will offer training,
education, and migration paths
so customers can fully exploit
the power of Linux-based Internet solutions. Both companies
will also continue to provide support for open standards, ensuring
compatibility across Linux, Windows, and Solaris platforms and
applications.
Corels work on the Linux
operating system grew out of its

earlier efforts developing software
for the UNIX operating system.
With the release of WordPerfect
8 for Linux in December 1998,
Corel established itself as a software developer for the opensource operating system. Corel
also developed the first Linux
operating system (OS) built specifically for the desktop.
Inprise/Borland provides tools
to create enterprise applications
for the Linux operating system.
Most recently, Inprise/Borland
announced a free download of
JBuilder 3 Foundation, a pure
Java development environment
for Linux; Kylix, planned to be
one of the first rapid application
development (RAD) tools for the
Linux platform, scheduled to be
available in mid-2000; and a free
download of the Linux Just-InTime (JIT) compiler.
The merger is subject to
certain customary conditions,
including shareholder approval
from Inprise/Borland and Corel,
compliance with the Hart-ScottRodino Antitrust Improvements
Act, and certain other regulatory
filings and approvals. The transaction is expected to close in
the late spring and is expected
to be tax-free to Inprise/Borland
shareholders.
For more information, visit
http://www.inprise.com or
http://www.corel.com.
On the Net
HTML / XML / SGML / Parsing / Internet
By Richard Phillips
Parsing the Web

Three Classes for Grabbing HTML/XML Information
recently bought one of the new digital satellite dishes and ran across an interesting
challenge figuring out just what was on, and when. DirectTV provided a Web site
with a basic search engine, but the Web-based search was slow and very narrowly
focused. As a typical programmer, I knew I could provide a better, more powerful UI, if I
could just figure out how their Web search engine worked.
A quick scan of the HTML source pointed out
a relatively simple search form that I could easily
duplicate, but the HTML results came back in
a mildly complicated HTML table. Brute force
code would have been simple enough to construct to parse through the table, but Id been
looking for a reason to build a more general
parser, so off I went. If Id known just how lax
the HTML rules are, and just how many hacks
there are, Id have just stuck with the brute force
method and saved myself a lot of agony, but
since Im here now ...
The Basics
To put together a parser for HTML, an understanding of the rules is required. HTML originated
as an SGML-like syntax, and over time has grown
to fit more closely within the confines of SGML.
These days the syntax is described within an
SGML Data Type Definition (DTD), bringing it
into a reasonably well-understood and managed
domain. Given that SGML now establishes the
underpinnings of HTML, the parser should apply
the SGML syntax rules as a starting point. This
also allows for the consideration of some simple
extensions that allow parsing of XML.
Therefore, the parser is built to work on SGML in
general, with specific handlers for exceptions and
extensions that occur in HTML and XML. The
rules for SGML are straightforward, and provide
five basic constructs that we care about:
elements,
attributes,
comments,
SGML directives, and

everything else.
Elements are the facet of the SGML content with

which we are most concerned, and around which
the parser is established. Elements have a start tag,
content, and an end tag, for example:
<TITLE>HTML Parsing</TITLE>
where TITLE is considered to be the name of the

tag. Element names are case-insensitive. So we start
with the following parsing rules:
Element start and end tags are surrounded by
< and > characters.
Element end tags are denoted by the /
character immediately following a <.
Element content is surrounded by start and
end tags.
HTML Extensions
In HTML, we immediately note that there are

exceptions to these rules. For some elements
most notably <P> the end tags may be omitted,
even though the element may have contents. This
offers perhaps the most annoying challenge of the
HTML parsing rule set, because there are several
methods by which the element may be terminated.
To start with, we note another syntax rule from
SGML: elements may not span. That is, if an
elements start tag is contained within another elements start and end tags, its end tag must also
appear there. Put simply, if we encounter an end
On the Net
tag, all omitted end tags are considered closed back up to the
matching start tag. Also, by observation (I couldnt find a formal
rule for this in the HTML specification), virtually all elements with
optional end tags close when they encounter another of themselves,
<LI> and <OPTION> being fine examples. Further, the HTML
reference material does indicate that <P> elements that omit their
end tags are terminated by block elements. The HTML DTD must
be consulted to determine which elements are considered block elements. Unfortunately, all of this prevents us from using a general rule,
and requires that we become concerned with the HTML DTD.
A quick consideration of the DTD is therefore in order. The DTD calls
out which elements must have end tags, and which may omit (or are
forbidden to have) end tags. For example, the DTD fragment for <P> is:
Comments
SGML also provides that its content may include comments. Comments are of the form:

The <! is a markup declaration open delimiter, and indicates that an

SGML directive is to follow. Comments are specifically denoted by
-- following the open delimiter (white space is not allowed between
the <! and the --, but is allowed between the closing -- and >).
Further, a comment may contain < and > characters. Comments may
not include other comments.
indicates a comment; --> terminates a comment.

and > are ignored while parsing a comment.
<!-<
<!ELEMENT P - O (%inline;)* >
The important thing to note here is - O. The - indicates the start tag
is required, and O means the end tag may be omitted. Compare this
to the fragment for <BR>:
<!ELEMENT BR - O EMPTY >
where - O EMPTY indicates that the start tag is required. Since the
element is EMPTY, however, the optional end tag is now expressly
forbidden. In the fragment for <P>, the (%inline;)* shows the
legal contents of the P element. In this case, %inline; refers to a
list of elements defined earlier in the DTD. Notably missing from
the %inline; list is <P> itself. A perusal of the other ambiguous
elements reinforces the observation that in general, an element may
not immediately contain itself (although this ambitiously general rule is
certainly not guaranteed to remain valid for future releases of the HTML
DTD). In the same way that %inline; is defined, so there exists a list
name %block;, which contains the list of block elements.
This leads to another set of parsing rules:
<P> with an omitted end tag is terminated when a block element
is encountered.
Elements are terminated when another element of the same name
is encountered.
Elements are terminated if a parents end tag is encountered; no
spans are allowed.
Attributes
Attributes represent the various properties of elements. By definition, attributes appear in name/value pairs within the start tag of
the element. For example, in:
<BODY bgcolor="#FFFFFF">
the BODY element has an attribute name bgcolor, and the attribute
has a value of #FFFFFF. Double quotation marks or single quotation
marks are required to delimit the value, unless the value contains
only letters, digits, hyphens, and periods. If the value contains single
quotation marks, it should be delimited with double quotation
marks, and vice versa. Attribute names are case-insensitive. Also
worth noting is that not all attributes have a value. For instance, the
NOWRAP attribute of the <TD> element.
Attributes appear within an elements start tag
Attributes are delimited by a space character (ASCII 32)
Attribute values are delimited by "or '
The remaining SGML directives are denoted by the beginning

markup declaration open delimiter <!. To further complicate things,
comments may exist within the directives delimited by --.
<! denotes an SGML directive (if its not a comment)
Comments within the directives are delimited by -Lastly, we consider what remains. Content of elements not contained
within a start tag, end tag, or comment is considered by the parser to
be PCData (parsed character data).
Store text not included in element start/end tags or comments
as PCData.
XML Extensions
As mentioned before, XML is also derived from SGML. While

HTML is basically a DTD described within and using SGML,
XML is a subset of SGML capable both of representing data
and containing other DTDs of its own. XML also demonstrates
that those working on the standards in the programming community actually do learn from the mistakes of those that went
before them. For instance, the rules of containment are much
more formal in XML than they are in HTML, making parsing a
great deal simpler. This means that while a DTD may be included
in an XML document for syntax checking purposes, it isnt necessarily required for the actual parsing of the XML content, as it
is for HTML.
Knowing this, we can add two more rules and provide DTD-less
XML parsing as well. For one, empty elements in HTML are
simply called out as such in the DTD with their end tags forbidden
(<BR> for example). If an element in XML is to be empty (that is,
it will have no content), its start tag may have a / just before the
closing >, indicating that no content and no end tag will follow.
Additionally, XML directives may appear with the ? character
rather than !.
Empty elements in XML may be terminated by a / just before the
> in the start tag, e.g. <partno/>.
Additional directives appear using ?, instead of the ! character.
Everything Else
Any items encountered in the content that are not contained in

element start or end tags, comments, or DTD items are considered
by the parser to be PCData. The content of elements fits this bill,
as do carriage returns and line feeds encountered by the parser. This
leads to the final parsing rule:
On the Net
<HTML>
<HEAD>
<TITLE>Example HTML</TITLE>
</HEAD>
<BODY>

<H1>Example HTML</H1>
Plain old text right out there in the middle of the document.
<P>Text contained within in paragraph</P>
<B>Unordered List</B>
<UL>
<LI>Item #1
<LI>Item #2
<LI>Item #3
</UL>
</BODY>
</HTML>
Figure 1: Sample HTML.
Any content located outside of start/end tags, comments, or

DTD items is PCData.
Additional Considerations
It is also worth noting that syntax errors and occurrences of browsertolerated HTML inconsistencies are frequently encountered, and as
such, should not raise exceptions except in extreme cases. Instead, a
warning should be noted and parsing should continue if possible.
The Parser
The goal of parsing the SGML content is to place the data it represents into a form more readily accessible to other components.
Because SGML calls out a hierarchical structure, a hierarchy is
probably the most accurate way to store the parsed content. With
that in mind, the parser is built from two primary classes, and a
third supporting class:
First and foremost is the THTMLParser class. Its Parse method
accepts the content to be parsed, and places the processed results
in the Tree property.
Next is the TTagNode class in which the parsed results are contained. This class is a hierarchical storage container with Parent
pointing to the TTagNode that contains the current node, and
Children containing a list of children immediately owned by the
current node.
TTagNodeList is provided as a list container for a collection of
TTagNode objects, typically produced by a call to the GetTags
method of the TTagNode class.
A Simple Example
Consider the sample HTML shown in Figure 1. The parser would

produce from the HTML a hierarchy that can be visualized as in Figure
2. Each of the boxes in the tree represents a TTagNode instance.
<root>
HTML
HEAD
NodeType
Caption
HTML elements
Text
Comments
SGML/XML/DTD
directives
nteElement
ntePCData
nteComment
nteDTDItem
Element name
!
! or ? and directive
Figure 3: TTagNode node types.

procedure Button1OnClick(Sender: TObject);
var
Elements : TTagNodeList;
Counter : Integer;
begin
Elements := TTagNodeList.Create;
HTMLParser1.Tree.GetTags('LI',Elements);
for Counter := 0 to Elements.Count - 1 do
ListBox1.Items.Add(Elements[Counter].GetPCData);
Elements.Free;
end;
Figure 4: Using GetTags.
Each node has a NodeType property that indicates what type of node
it is. All node types except ntePCData also have text in the Caption
property that provides more information about the nodes contents.
See the table in Figure 3 for details.
For example, the HTML node in Figure 2 has a NodeType of
nteElement, while the comment tag is of type nteComment, and the
PCData nodes are of type ntePCData.
The content or text for each HTML element is contained in a node
in its Children list. For example, the TITLE node in Figure 2 has
a PCData node whose Text property contains Example HTML.
The GetPCData method of a TTagNode returns the PCData text for
all children of the node for which its called. Note, this method is
recursive and will return the PCData text for all nodes in the tree
beneath the node upon which its called.
Retrieving Elements
The GetTags method of a TTagNode will return a list of all children

that match an element name. If '*' or '' is specified as the element
name, then all children will be returned. Note that this method is
recursive. The result list is a TTagNodeList.
The code in Figure 4 illustrates how the
GetTags method is used to collect a list of all <LI> elements in the
HTML from Figure 1 and insert their contents into a list box. The
process is as follows:
1) Create a container for the results, i.e. the Elements list.
2) Call the GetTags method, passing the desired element name and
the container.
3) Iterate through the container, placing the text for each element
in a list box.
4) Destroy the container.
BODY
TITLE
PCDATA
Node Contents
Comment
H1
PCDATA
PCDATA
PCDATA
PCDATA
LI
Figure 2: Hierarchical representation of

parsed HTML.
PCDATA
TTagNodeList has two methods that offer another approach

(assuming that a result set has already been acquired):
GetTagCount and FindTagByIndex. GetTagCount
UL
returns a count of the occurrences of an element
name. FindTagByIndex returns the index within the
list of the specified occurrence of an element
name. For instance, the statement:
LI
LI
PCDATA
PCDATA
ShowMessage(Elements.FindTagByIndex('li',1).GetPCData);
On the Net
were it included in Figure 4, would display the text for the second
occurrence of the <LI> element in the Elements container. This can
prove exceptionally useful for locating a specific tag from target
HTML content. For example, if the third <TABLE> in the HTML
contained the desired data, the following code would make quick
work of locating the root <TABLE> node:
HTMLParser1.GetTags('*',Elements);
Node := Elements.FindTagByIndex('table',2);
if Assigned(Node) then
begin
// Perform processing on <TABLE> node.
end;
Working with Results as a Hierarchy
The procedure in Figure 5 provides yet another method for accessing

the contents of the Tree (albeit a more brute force approach). In this
example, the HTML content is assumed to be fairly straightforward:
<TABLE> elements contain <TR> elements, which contain <TH>
and <TD> elements. Given this reasonably accurate assumption, the
code will walk the children of a <TABLE> node, and for each <TR>
node found will walk its children looking for occurrences of either
<TH> or <TR>, and add their text (contained in PCData nodes) to a
TStrings container, e.g. the Lines property of a TMemo control.
results as a list. The goal here is to retrieve a list of all comments, links,
meta tags, and images from the document, with the <TITLE> thrown
in for good measure. The code does this in several simple steps:
1) Parse the HTML.
2) Create a container for the list of matching nodes from the tree.
3) Call the GetTags method passing '*', and the container ('*'
indicates that all items in the tree should be returned).
4) Iterate through the container collecting matches and place their
contents in the StringList.
5) Destroy the container.
The important thing to understand here is that the TTagNodeList
class is just a list of pointers to nodes from the Tree. This is quite
beneficial in that once a desired node is located in the list, it may be
used as if it had been acquired by traversing the tree. For example,
when the case statement in Figure 6 encounters a TITLE element,
// Return TableNodes contents in a TStrings container.

procedure GetTable(TableNode: TTagNode; Lines : TStrings);
var
RowCtr,
DataCtr : Integer;
Node,
RowNode : TTagNode;
TempStr : string;
begin
Lines.Clear;
if CompareText(TableNode.Caption,'table') = 0 then begin
for RowCtr := 0 to TableNode.ChildCount - 1 do begin
RowNode := TableNode.Children[RowCtr];
if CompareText(RowNode.Caption,'tr') = 0 then begin
TempStr := '';
for DataCtr := 0 to RowNode.ChildCount - 1 do begin
Node := RowNode.Children[DataCtr];
if CompareText(Node.Caption,'td') = 0 then
TempStr := TempStr + Node.GetPCData + #9
else
if CompareText(Node.Caption,'th') = 0 then
TempStr := TempStr + Node.GetPCData + #9;
end;
TempStr := Trim(TempStr);
if TempStr <> '' then
Lines.Add(TempStr);
end;
end;
end;
end;
procedure TForm1.Parse(HTML: string; Lines: TStrings);

const
cKnownTags = '|title|img |a
|meta |!
';
cTITLE = 0;
cIMG = 1;
cA = 2;
cMETA = 3;
cComment = 4;
var
Index,
Counter : Integer;
TempStr : string;
Nodes : TTagNodeList;
begin
HTMLParser1.Parse(HTML);
Nodes := TTagNodeList.Create;
// Retrieve all nodes.
HTMLParser1.Tree.GetTags('*',Nodes);
for Counter := 0 to Nodes.Count - 1 do begin
TempStr := '|' + LowerCase(Nodes[Counter].Caption);
// Index of element name.
Index := Pos(TempStr,cKnownTags);
if Index > 0 then begin
Index := Index div 6;
case Index of
cTITLE :
Lines.Add('Title=' +
HTMLDecode(Nodes[Counter].GetPCData));
cIMG :
begin
TempStr := Nodes[Counter].Params.Values['src'];
Lines.Add(
Nodes[Counter].Params.Values['src']);
end;
cA :
begin
TempStr :=
Nodes[Counter].Params.Values['href'];
Lines.Add(TempStr + '=' +
HTMLDecode(Nodes[Counter].GetPCData));
end;
cMETA :
with Nodes[Counter].Params do
Lines.Add(Values['name'] + '=' +
Values['content']);
cComment :
Lines.Add('[Comment] ' +
HTMLDecode(Nodes[Counter].Text));
end; { case Index }
end; { if Index > 0 }
end; { for Counter := 0 to Nodes.Count - 1 }
Nodes.Free;
end;
Figure 5: Working with results as a hierarchy.
Figure 6: Working with results as a list.
As an illustration of just how difficult HTML processing can be, the

following caveats apply to the code provided in Figure 5 and would
have to be handled to provide a robust solution:
Subtables are not handled. That is, <TABLE> elements encountered within a <TD> element are ignored.
Row and column spanning is not handled.
<TBODY>, <THEAD> and a host of other table elements are
not considered (although admittedly they are rare).
Working with Results as a List
The procedure in Figure 6 demonstrates working with the parsed
On the Net
its contents are retrieved by making a call to the TITLE nodes
GetPCData method (which depends on the parsed tree structure
behaving as it appears to in Figure 2). Note that the PCData often
contains encoded items such as > and < (< and > respectively).
HTMLDecode is provided for handling most simple cases, but doesnt
handle all cases (notably non-US character encoding).
they contain images, not program content. Also, the channel appears
only in the first row of a set of programs that occur on that channel.
This example also demonstrates the use of the attributes from

an element. When an <A> element is encountered, the HREF
attribute is examined. If it exists, the <A> element is treated as
a link to some other resource. If the HREF attribute were not
specified, this might be an instance of <A> serving as an anchor
instead of a link. For more details on how the Params property
of the TTagNode behaves, see the Delphi help for the Names and
Values properties of the TStrings class.
Further Study
Searching the www.directv.com Program Guide
Applying the HTML parser to the original need turns out to be

another simple albeit involved exercise (see Listing One beginning on page 11). First, an understanding of the CGI scripts that
allow searching of the program guide is required. The search script is
rather crude and accepts only three parameters: timezone, category, and
search text. timezone is simply a number representing Eastern, Central,
Mountain, or Pacific. category allows the search to span all programs,
or to be narrowed to certain types of programs. The search text should
be all, or a portion of, the desired program name. The search is specific
to program names, and doesnt consider program descriptions.
The results of the search are returned as an HTML table including
channel, date, time, duration, and program name. The program name
is contained within a link to the program description, which will need
to be retrieved as well. This is slightly complicated by the fact that
the link provided is written using JavaScript, which we cannot simply
call. However, the URL produced by the JavaScript function is easy to
replicate, as it contains a program ID number that can be passed to
another CGI script that returns the desired description.
The next step is to parse the HTML and process the results into a
more useful format. TStringTable is provided as a simple container for
just this purpose. The TStringTable offers a non-visual equivalent to
TStringGrid with a few additional methods to make manipulating the
data a bit easier. Once the HTML table has been processed and placed
in the string table, a bit of house cleaning is required. For one, there
are rows at the end of the HTML table that need to be ignored, as
The contents can now be added to a ListView. Once that task is

complete, the descriptions can be fetched using the program IDs to
call the description CGI, and then added to the ListView.
While these examples are not terribly glamorous, the parser can
be applied to more meaningful problems. For instance, a friend of
mine has put together an extremely handy application using the
Pricewatch site (http://www.pricewatch.com) to monitor prices on
PC hardware. Pricewatch offers a current snapshot of pricing of a
particular piece of hardware from various vendors, usually sorted
from least expensive to most. However, it doesnt allow for viewing
of several different pieces of hardware at once, and it doesnt track
the history of the price changes for the hardware. So, the application provides a way to build a list of hardware to be tracked, and
then offers a simple trend analysis by gathering and saving off the
price information on a regular basis. This provides a useful picture
for the consumer of just how quickly the prices are moving downward on a particular item. If the pricing is in a steep downward
curve, waiting to purchase might be wise. If the curve is flat, the
time to purchase might be at hand.
The parser is used not only to retrieve the pricing data, but also to
help deal with one of the more significant issues facing those attempting to interface with sites they do not control: unexpected changes in
the target sites contents. In this case, a review of the <form> elements
from the main page is performed to ensure that the query mechanism
remains intact. As a further safeguard, the search results page is also
examined to verify a match against the expected HTML format. If
unexpected items are found in either case, processing cannot continue,
but at least the user can be warned of the situation.
In a more interesting demonstration of the parsers abilities, it has
been combined with a database to create a poor mans OODB
(object-oriented database). XML is used to wrap the data, and is
then stored in text fields in the database. When needed, the XML
is retrieved and the parser used to extract the data. Without going
in to detail, this is useful because the data stored in the database
can carry semantic information with it (the XML elements) that
provides information about the datas structure. In systems where
the data structure is dynamic, this provides a simple way to avoid
excessive database maintenance and further provides a clean, easily
understood mechanism for the exchange of data between various
applications and platforms. In the case mentioned here, a legacy
defect (bug) tracking system hosted on a Solaris platform was
wrapped in a Delphi-based UI.
Additional Demonstration Applications
To further demonstrate the power of the classes presented in this article,

two additional applications accompany this article. An HTML Engine
Demo application (see Figure 7) displays a great deal of information
about any selected URL, including meta tags, links, and images.
The Parser Test application parses any URL, or SGML/HTML/XML
document, and displays the results in a TreeView (see Figure 8). It can
also display links, selected tags, text, etc.
Room for Improvement

Figure 7: The HTML Engine Demo application.
This parser builds a reasonable basis for parsing of HTML and XML,
but offers significant room for further development.
On the Net
Further performance tuning. While some attention has been paid to this
area, no extreme efforts to speed things up were applied. Most notably,
Delphis string routines are not considered to be as fast as those provided
in some third-party string-handling collections (notably HyperString).
The files referenced in this article are available on the Delphi Informant
Magazine Complete Works CD in INFORM\00\MAY\DI200005RP.
Richard Phillips is a Development Manager for i2 Technologies working on their TradeMatrix technology integrating disparate datasources. Hes been using Delphi since it was
introduced and Borland Pascal since 1985. He can be reached at richardp@dallas.net.
Figure 8: The Parser Test application.
Incorporation of a DTD processor (DTDs can be parsed with

the existing parser, but no handling of the parsed contents is
provided). This would provide two main benefits: more thorough
parsing of elements based on a true understanding of their legal
contents, and no need for hard coding a representation of the
HTML DTD within the parser. Further, DTD-based XML parsing
would then be possible.
A DOM container model to complement the TTagNode model.
DOM represents a fairly well understood and commonly encountered model for representing the parsed contents of XML. While
it doesnt suit all needs, it does provide a useful, standard way to
communicate about the parsed elements.
XQL or other suitable extended model query mechanism. The
GetTags method is reasonably sufficient, but for more exhaustive queries against XML contents, a more advanced mechanism is desired.
For example, it would be extremely handy if GetTags could be passed
'order/partno' to indicate that were searching for all PARTNO
elements that are immediately below an ORDER element.
Resources and Alternatives

HTML 4.0 Specification http://www.w3.org/TR/REC-html40.
This is the single most useful resource to those seeking HTML
enlightenment. It is extremely detailed and well written.
HTML 4.0 Loose DTD http://www.w3.org/TR/REC-html40/loose.dtd.
A part of the HTML 4.0 specification, this offers the exact specification of just what the HTML rules are. It is upon this DTD (as
opposed to the strict DTD) that the parser is designed to operate.
XML: Creating Structures of Meaning, Visual Developer, Nov/Dec
1998, Vol. 9 No. 4. A quick survey of XML for the beginner. Syntax
and use are explored here with an eye to bringing the novice on
board.
Using Internet Explorers HTML Parser, Dr. Dobbs Journal,
#302, August 1999 (http://www.ddj.com/articles/1999/9908/
9908toc.htm). This article offers an examination of using the HTML
parser that is available within Microsofts Internet Explorer via
COM interface. The source for the article is in C++, but its not
difficult to follow.
XML from Delphi, Delphi Informant Magazine, July 1999, Vol. 5 No.
7. A beginners explanation of XML, and an excellent example of using
the XML parser included in Internet Explorer 4.0.
Begin Listing One Searching www.directv.com

unit Unit1;
interface
uses
Windows, Messages, SysUtils, Classes, Controls, Forms,
Dialogs, ComCtrls, StdCtrls, ExtCtrls, GetURL, HTMLMisc,
HTMLParser, StringTable;
type
TForm1 = class(TForm)
WIGetURL1: TWIGetURL;
StatusBar1: TStatusBar;
Panel2: TPanel;
pbSearch: TButton;
ebSearchText: TEdit;
Panel3: TPanel;
ListView1: TListView;
HTMLParser1: THTMLParser;
procedure pbSearchClick(Sender: TObject);
procedure WIGetURL1Status(Sender: TObject;
Status: Integer; StatusInformation: Pointer;
StatusInformationLength: Integer);
private
procedure GetTable(TableNode: TTagNode;
Table : TStringTable);
function GetDescription(Node: TTagNode): string;
end;
var
Form1: TForm1;
implementation
{$R *.DFM}
const
cDirecTVSearchURL =
'http:// 206.17.88.15/cgi-bin/pgm_search.cgi/';
cDirecTVDescURL =
'http:// 206.17.88.15/cgi-bin/pgm_desc.cgi/';
// Return TableNode's contents in Table
procedure TForm1.GetTable(TableNode: TTagNode;
Table: TStringTable);
var
RowCtr,
DataCtr : Integer;
Node,
RowNode : TTagNode;
begin
Table.Clear;
if LowerCase(TableNode.Caption) = 'table' then begin
for RowCtr := 0 to TableNode.ChildCount - 1 do begin
RowNode := TableNode.Children[RowCtr];
if LowerCase(RowNode.Caption) = 'tr' then begin
Table.NewRow;
On the Net
for DataCtr := 0 to RowNode.ChildCount - 1 do begin
Node := RowNode.Children[DataCtr];
if LowerCase(Node.Caption) = 'td' then
Table.AddColumnObject(Node.GetPCData,Node)
else
if LowerCase(Node.Caption) = 'th' then
Table.AddHeader(Node.GetPCData);
end;
if Table.Row[Table.RowCount - 1].Count <= 0 then
Table.DeleteRow(Table.RowCount - 1);
end;
end;
end;
end;
function TForm1.GetDescription(Node : TTagNode) : string;
var
TempStr : string;
begin
Result := '';
if Node.ChildCount > 0 then
TempStr := Node.Children[0].Params.Values['href']
else
TempStr := '';
if TempStr <> '' then begin
// Parse out the description id
Delete(TempStr,1,Pos('(',TempStr));
Delete(TempStr,Pos(')',TempStr),Length(TempStr));
WIGetURL1.URL := cDirecTVDescURL + TempStr;
Screen.Cursor := crHourglass;
Application.ProcessMessages;
WIGetURL1.GetURL;
Screen.Cursor := crDefault;
StatusBar1.Panels.Items[0].Text := '';
if WIGetURL1.Status = wiSuccess then begin
TempStr := WIGetURL1.Text;
// Use brute force to scrape out program description.
if Pos('<BLOCKQUOTE>',TempStr) > 0 then begin
Delete(TempStr,1,Pos('<BLOCKQUOTE>',TempStr) + 11);
Delete(TempStr,Pos('</BLOCKQUOTE>',TempStr),
Length(TempStr));
end;
Result := TempStr;
end;
end
end;
procedure TForm1.pbSearchClick(Sender: TObject);
const
tzPacific = '0';
// Time zones.
tzMountain = '1';
tzCentral = '2';
tzEastern = '3';
cgMovies = '0';
// Categories.
cgSports = '1';
cgSpecials = '2';
cgSeries = '3';
cgNews = '4';
cgShopping = '5';
cgAllCategories = '-1';
var
Cols,
Rows : Integer;
NewItem : TListItem;
Node : TTagNode;
Nodes : TTagNodeList;
ResultTable : TStringTable;
TempStr : string;
begin
if ebSearchText.Text = '' then
Exit;
WIGetURL1.URL := cDirecTVSearchURL + tzCentral + '/' +
cgAllCategories + '/' + urlEncode(ebSearchText.Text);
Screen.Cursor := crHourglass;
Application.ProcessMessages;
WIGetURL1.GetURL;
Screen.Cursor := crDefault;
StatusBar1.Panels.Items[0].Text := '';
if WIGetURL1.Status = wiSuccess then
begin
if Pos('No program titles that match',
WIGetURL1.Text) > 0 then
ShowMessage('No matches found')
else
begin
// Attempt to parse HTML table we're looking for.
HTMLParser1.Parse(WIGetURL1.Text);
Nodes := TTagNodeList.Create;
HTMLParser1.Tree.GetTags('table',Nodes);
if Nodes.Count > 0 then
begin
ResultTable := TStringTable.Create;
GetTable(Nodes[0],ResultTable);
// Get rid of image tags at bottom of search
// response (the 2nd column has no contents).
with ResultTable do
for Rows := RowCount - 1 downto 0 do
if Cells[1,Rows] = '' then
DeleteRow(Rows);
// Ensure all cells are filled appropriately
// (in the HTML table, a RowSpan attribute
// allows the "Channel" to be displayed in
// one cell for several programs).
with ResultTable do
for Rows := 0 to RowCount - 1 do
for Cols := 0 to ColCount - 1 do
if Cells[Cols,Rows] = '' then
if Rows > 0 then
Cells[Cols,Rows] :=
Cells[Cols,Rows - 1];
// Add items to ListView (Program, Channel,
// Data, Time).
ListView1.Items.Clear;
for Rows := 0 to
ResultTable.RowCount - 1 do begin
NewItem := ListView1.Items.Add;
NewItem.Caption :=
ResultTable.Cells[4,Rows];
NewItem.SubItems.Add(
ResultTable.Cells[0,Rows]);
end;
// Retrieve program descriptions (program id
// contained in 4th column's node).
for Rows := 0 to
ResultTable.RowCount - 1 do begin
// It's rude to whack the server :-)
Sleep(500);
Node :=
TTagNode(ResultTable.Objects[4,Rows]);
TempStr := GetDescription(Node);
ListView1.Items[Rows].
SubItems.Add(TempStr);
end;
ResultTable.Free;
end // if Nodes.Count > 0 ...
else
ShowMessage(
'Error - Expected table...found none');
end; // else of Pos('No program titles that...
end // WIGetURL1.Status = wiSuccess...
else
ShowMessage('Unable to contact search server [' +
WIGetURL1.ErrorMessage + ']');
end;
end.
End Listing One
Delphi at Work
Office 2000 / Add-ins / COM / Delphi 5
By Ron Loewy
Delphi in the Office

Writing Office 2000 Add-ins in Delphi
dont have a personal productivity profiler on my machine, but if I had to guess, I

would say that I probably spend most of my time in front of a monitor using these
programs: Delphi for development, an Internet browser/e-mail for communication, and
Microsoft Office for productivity more specifically, Word for document writing, Excel for
spreadsheets, and the occasional PowerPoint presentation.
When Office 2000 was released, it immediately
caught my attention. It offered a unified Visual
Basic for Applications (VBA) environment for its
applications, Web documents, and even an assistant that I find useful rather than annoying (yes, I
am talking about the paper clip). The new feature
that really interests me, however, is the unified
COM-based add-in architecture. With this architecture, Im able to write enhancements and remedies (using Delphi) to the many things that irritate me about Office applications.
The problem with writing COM add-ins in Delphi
3 or 4 was that I had to create my own wrappers
around the published COM interfaces. Worse than
that was the prospect of having to use connection
points and event sinks to interface with events
published by OLE Servers. Because a useful add-in
needs to be notified of events in Office applications, Ive never looked forward to the task. As if
they were reading my mind, the good people at
Inprise released Delphi 5, which offers the great
/L+ switch to the type library import utility, giving
developers the ability to create object wrappers
around OLE Servers. I have no more excuses.
The Office 2000 Add-in Architecture
Office 2000 is available in many editions. One of

these is the Developer edition, and it includes the
ability to create add-ins in Visual Basic (VB). As
nice as VBA is in the latest Office release, VB hides
a lot of the stuff behind closed doors, and writing a
VB add-in doesnt teach you how COM is used in
the new add-in architecture.
On Microsofts Web site, I located Knowledge Base
article Q230689: Office 2000 COM Add-In Written in Visual C++ (http://support.microsoft.com/
download/support/mslfiles/COMADDIN.EXE).
This downloadable file describes the COM interfaces that take part in an add-in construction.
Inspecting the C++ code also reveals how we can

use Delphi to write Office 2000 add-ins.
An Office 2000 add-in is a COM automation object
that implements the IDTExtensibility2 interface.
This interface is remarkably simple, which explains
its use in many different applications, such as Word,
Excel, PowerPoint, and Outlook. The add-in must
implement all five functions defined in the interface:
1) OnConnection. This function is called when
the application connects to the add-in. The
add-in receives initialization information in
this call including the pointer to the applications object model entry point, the connection mode (e.g. was the add-in started
manually on application startup or via the
command line?), a pointer to the object that
represents the add-in in the applications object
model, and user-defined information.
2) OnDisconnection. This function is called when
the application disconnects from the add-in.
This is the place the add-in uses to clean the
resources it allocated, and remove the userinterface elements it added to the application.
3) OnStartupComplete. This function is called
only if the add-in was automatically started by
the application. When this function is called,
all the other add-ins have been loaded into
memory. If your add-in needs to communicate
with them, it can. I like to use this event to add
the user-interface elements to the application.
4) OnBeginShutdown. This function is called
when the application is getting ready to shut
down and will continue to disconnect from the
add-in. At this point, the add-in needs to stop
accepting user input.
5) OnAddInsUpdate. This function is called when
the list of registered add-ins is changed. My guess
is that if your add-in depends on another addin(s), this function might be of some interest to
you; otherwise, you will usually leave it empty.
Delphi at Work
// IDTExtensibility2 methods
procedure OnConnection(const Application: IDispatch;
ConnectMode: ext_ConnectMode; const AddInInst: IDispatch;
var custom: PSafeArray); safecall;
procedure OnDisconnection(RemoveMode: ext_DisconnectMode;
var custom: PSafeArray); safecall;
procedure OnAddInsUpdate(var custom: PSafeArray); safecall;
procedure OnStartupComplete(var custom: PSafeArray);
safecall;
procedure OnBeginShutdown(var custom: PSafeArray);
safecall;
Figure 1: The declarations of the IDTExtensibility2 methods.
Interfaces, Type Libraries, and Constants
To create the add-in, youll need to import some COM objects and
type libraries into Delphi. I used Delphi 5s TlibImp.exe (installed
in the /Bin sub-directory of the standard Delphi installation directory) to import everything. The new version of this utility supports
the new /L+ flag, which creates an OLE Server Delphi wrapper
around COM objects and automatically maps their properties and
events for easy Delphi use.
The IDTExtensibility2 interface that our add-in needs to implement
is declared in the file MSADDNDR.DLL, located in the \Program
Files\Common Files\Designer\ directory.
I used TLIBIMP /L+ \Program Files\Common
Files\Designer\MSADDNDR.DLL from the Imports
sub-directory of Delphis root directory. The result is the
file AddInDesignerObjects_TLB.pas (and
AddInDesignerObjects_TLB.dcr). We will need to use this file in
the uses clause of our project to gain access to the interface. For
some reason, TLIBIMP renamed the interface _IDTExtensibility2
(notice the underline prefix). Experience taught me to accept this
as a necessary evil. (I would suggest not fighting TLIBIMP for the
underlines that it loves to add at the start and end of interfaces,
properties, methods, or constants).
The next step is deciding which Office application (or applications)
we want to create the add-in for. This article will use Word 2000
as the sample. There are plenty of articles, books, and documentation resources about the different Office application object models.
Naturally, when you create an add-in for Outlook, Excel, or any
other Office application, youll need to access the object model for
the particular application and import its type libraries.
I imported Words type library from the file MSWORD9.OLB in
the \Program Files\Microsoft Office\Office directory. Similarly, if you
want to create an add-in for Excel, you will import EXCEL9.OLB;
for Access you need to import MSACC9.OLB; and for Outlook
MSOUTL9.OLB. TLIBIMP automatically imports the type library
of the shared Office components (MSO9.DLL). The result are the
files Office_TLB.pas and Word_TLB.pas.
Note that Borland supplied imports of the Office 97 files like
Word97.dcu. The unified add-ins architecture doesnt work in Office
97 applications, and we have to perform the new import to gain
access to the latest object models.
A Basic Add-in
A basic add-in implements IDTExtensibility2 and doesnt do anything

that actually interfaces with the host application. Its obviously a
useless add-in, but we must start with such an add-in before we write
our specific functionality.
Add-ins can be implemented as either in-process or out-of-process

COM servers. For this article, I created an in-process server (available
for download; see end of article for details). In Delphi, select File | New.
On the ActiveX tab, double-click the ActiveX Library icon. Save the file
in your development directory (I named it DIWordAddIn). Now, double-click on the Automation Object icon on the ActiveX tab. I named
the class AddIn and saved the implementation unit as AddInMain.pas.
In AddInMain.pas, I added AddinDesignerObjects_TLB, Word_TLB,
and Office_TLB to the uses statement. Then, I added IDTExtensibility2
to the class definition as one of the interfaces that the object implements. The class definition now reads:
type
TAddIn = class(TAutoObject, IAddIn, IDTExtensibility2)
...
To the protected section of the class definition, I added the

declarations of the IDTExtensibility2 methods (see Figure 1).
You can use Delphis CVC to complete the class definition, and
add the method implementation bodies in the unit source. To test the
add-in, add the following code to the OnConnection method:
ShowMessage('Connected to ' + WordApp.Name);
Add this code to the OnDisconnection method:

ShowMessage('Bye Bye');
The basic add-in is now ready to be compiled and registered

with Word.
Registering an Add-in with an Office Application
Like any other COM object, an add-in needs to be registered with

the system. By selecting Run | Register ActiveX Server, you can register
the object with the COM run-time manager. In addition to the
standard COM registration, you need to register the COM object
with the Office application for which it was created. To do so, you
need to create a new key in the registry. The keys name should be:
HKEY_CURRENT_USER\Software\Microsoft\Office\<AppName>\
Addins\<AddInProgID>
where <AppName> is the name of the application (Word in our case)

and <AddInProgID> is the name of the automation object. The
automation object in our case is called DIWordAddIn.AddIn (the
name of the ActiveX library and the name of the class).
We need to create several values under the following key:
HKEY_CURRENT_USER\Software\Microsoft\Office\Word\Addins\
DIWordAddIn.AddIn
A DWORD value called LoadBehavior determines how the add-in is

loaded and used in the application. In our case, we would set it to 3
a combination of Connected and BootLoad. A list of available value
combinations is shown in the table in Figure 2. The add-in is then
connected to the application, and starts when the application is opened.
Additional values can appear in the registry key, like a friendly name
that will be displayed in the applications COM manager dialog box,
and an indication of whether the add-in can be activated from the
command line.
Delphi at Work
The Office 2000 User Interface
A common set of objects that represents the user interface elements
is shared among all the Office applications. Menu bars, toolbars,
common controls (e.g. toolbar buttons and combo boxes), and even
the often-maligned Office assistant exist throughout the Office suite.
When you imported the Word type library earlier, the common Office
objects type library was also imported, and the Office_TLB.pas was
created. Unfortunately, when Delphi imports this type library, it doesnt
create OLE Server wrapper objects. We will have to manually create a
Delphi wrapper for the CommandBarButton object exposed by Office.
This object represents a simple menu item or toolbar button in Office.
Value Use
$0
$1
$2
$8
$16
Disconnected not loaded.

Connected loaded.
BootLoad automatic application start.
DemandLoad load only on user request.
ConnectFirstTime load only once next application startup.
Figure 2: A list of value combinations for loading your add-in.
tion. We also need to define an interface pointer to the new toolbar

(CommandBar) well create, as well as two fields using the new
TButtonServer class to hold a toolbar button and menu item.
Like most of Microsofts applications, an Application object represents an

entry point to the object model. Office Application classes provide access
to the CommandBars property. This is a collection of CommandBar
objects: toolbars, floating toolbars, or menus. The Office object model
allows you to create or update existing CommandBars. Office_TLB.pas
contains the ICommandBar interface, which represents such an object.
In the private section of the add-in class, add:
A CommandBar object has a Controls collection, which is a collection of CommandBarControl items. A CommandBarControl is an
item that appears in a command bar; a simple CommandBarButton
is used to represent a simple toolbar button (or menu item), but
you can also create CommandBarCombo controls (combo boxes),
CommandBarPopup (drop-down menus), and CommandBarActiveX
(everything that wraps an ActiveX control).
In the OnConnection method, store the application pointer using

the following code:
If you check Office_TLB.pas, youll notice that, in addition to an

ICommandBarButton interface, an ICommandBarButtonEvents interface
is implemented. As you can imagine, if our add-in wants to add toolbar
buttons and menu items, well need to create new CommandBarButtons
and implement the Click event defined in the events interface.
Connecting to a connectable object is done via connection points and
sinking events (see Binh Lys two-part COM Callbacks series in the
June and July 1998 issues of Delphi Informant Magazine). Call me
lazy, but if I can avoid sinking events, I jump at the opportunity. Fortunately, inspecting the code that Delphi created for TWordApplication in
Word_TLB.pas can give us a simple template to create a Delphi wrapper
for every connectable object that does the event sinking automatically.
The file BtnSvr.pas is a module that contains such a manually-created
wrapper. In addition to properties of a CommandBarButton that I surfaced as standard Delphi properties, the InitServerData, InvokeEvent,
Connect, ConnectTo, and Disconnect methods were needed. If you inspect
the code, youll see that its virtually the same code that Delphi created for
TWordApplication simplified for the simple CommandBarButton events.
We must define the server data in the InitServerData method. You
can find the different interface GUIDs in Office_TLB.pas, define an
internal interface (Fintf ) to a CommandBarButton interface, and set the
InvokeEvent method to activate a Delphi event based on the DispID
defined in the events interface. Finally, the Connect, ConnectTo, and
Disconnect methods set Fintf to the desired interface and sink its events.
The Delphi Wrapper class defined in BtnSvr.pas is called
TButtonServer. Its derived from Delphis TOleServer and uses its
methods to perform event sinking, etc.
Connecting to the Application
Now we have a Delphi wrapper for a command bar button, and are
ready to interface with Word in our add-in. We need to declare a
TWordApplication field that will hold the reference to the Word applica15 May 2000 Delphi Informant Magazine
FWordApp : TWordApplication;
DICommandBar : CommandBar;
DIBtn : TButtonServer;
DIMenu : TButtonServer;
var
WA : Word_TLB._Application;
begin
FWordApp := TWordApplication.Create(nil);
WA := Application as Word_TLB._Application;
WordApp.ConnectTo(WA);
TWordApplication is defined by Delphi, and the ConnectTo method

is used to connect it to the interface passed by Word to our add-in.
Because TWordApplication maps the interface events to Delphi events,
we can now set event handlers using the standard Delphi syntax:
WordApp.OnEventX := EventXHandler;
For example, if we want to do something when the selection

in Word changes, we could write an event handler for the
OnWindowSelectionChange event.
Creating a New Toolbar, Button, and Menu Item in the Add-in

Before we add a new toolbar and button, we need to create an event
handler for the buttons OnClick event. For this sample, well create
a very simple event handler:
procedure TAddIn.TestClick(const Ctrl: OleVariant;
var CancelDefault: OleVariant);
begin
ShowMessage('Ouch, this hurts!');
CancelDefault := True;
end;
The CancelDefault parameter is used when you replace the functionality

of a pre-built menu item or toolbar button. Its unnecessary in this
specific case, because were creating a new button with our add-in.
Because our add-in is registered to start on application startup, we can
be sure the OnStartupComplete method will be called, and will use
it for the creation of the add-in user interface elements. Well define
BtnIntf as a CommandBarControl interface (the parent interface of the
CommandBarButton that well use for our button). Our first task is to
determine if our toolbar has already been created (see Figure 3).
Delphi at Work
DICommandBar := nil;
for i := 1 to WordApp.CommandBars.Count do
if (WordApp.CommandBars.Item[i].Name =
'Delphi Informant') then
DICommandBar := WordApp.CommandBars.Item[i];
// See if we already registered the command bar with Word.
if (not Assigned(DICommandBar)) then begin
DICommandBar := WordApp.CommandBars.Add(
'Delphi Informant',EmptyParam,EmptyParam,EmptyParam);
DICommandBar.Set_Protection(msoBarNoCustomize)
end;
Figure 3: Determining whether our toolbar has already been created.

procedure TAddIn.MenuClick(const Ctrl: OleVariant;
var CancelDefault: OleVariant);
var
Sel : Word_TLB.Selection;
Par : Word_TLB.Paragraph;
begin
Sel := WordApp.ActiveWindow.Selection;
if (Sel.Type_ in [wdSelectionNormal,
wdSelectionIP]) then begin
Par := Sel.Paragraphs.Item(1);
if (Par.Borders.OutsideLineStyle <
wdLineStyleInset) then
Par.Borders.OutsideLineStyle :=
1 + Par.Borders.OutsideLineStyle
else
Par.Borders.OutsideLineStyle := wdLineStyleNone;
end;
end;
Figure 4: Adding a new menu item by creating its OnClick event

handler.
ToolsBar := WordApp.CommandBars[Tools];
MenuIntf := ToolsBar.FindControl(EmptyParam, EmptyParam,
'DIMenu', EmptyParam, EmptyParam);
if (not Assigned(MenuIntf)) then
MenuIntf := ToolsBar.Controls.Add(msoControlButton,
EmptyParam, EmptyParam, EmptyParam, EmptyParam);
DIMenu := TButtonServer.Create(nil);
DIMenu.ConnectTo(MenuIntf as _CommandBarButton);
DIMenu.Caption := 'Delp&hi Menu';
DIMenu.ShortcutText := '';
DIMenu.Tag := 'DIMenu';
DIMenu.Visible := True;
DIMenu.OnClick := MenuClick;
Figure 5: Connecting a TButtonServer wrapper to the MenuClick

event handler.
Now we give our toolbar a unique name (Delphi Informant)

and check on startup if this toolbar already exists. If it does, we
set the DICommandBar field to it; otherwise, we use the Word
applications CommandBars propertys Add method to create a new
one. After the new toolbar is created, we set its protection to
msoBarNoCustomize, which prevents users from adding or removing buttons from our toolbar.
At this point, DICommandBar points to a valid toolbar. Were
now ready to get the interface pointer to our toolbar button
(CommandBarButton) via the Controls collection of this interface.
Because we set the protection of the toolbar to no customization,
we know that if there is a control on the toolbar, it will be our
button (well add just one button to the toolbar). If we cant find
any controls on the toolbar, well add a new one:
if (DICommandBar.Controls.Count > 0) then
BtnIntf := DICommandBar.Controls.Item[1]
else
BtnIntf := DICommandBar.Controls.Add(msoControlButton,
EmptyParam, EmptyParam, EmptyParam, EmptyParam);
Notice that the first item in a collection in Office is item #1, not
item #0 as we are used to from the Windows API or Delphis TList
and TStringList. At this point, BtnIntf has the desired toolbar
button interface, and we can create a TButtonServer wrapper
around it:
DIBtn := TButtonServer.Create(nil);
DIBtn.ConnectTo(BtnIntf as _CommandBarButton);
DIBtn.Caption := 'Delphi Test';
DIBtn.Style := msoButtonCaption;
DIBtn.Visible := True;
DIBtn.OnClick := TestClick;
Well use the ConnectTo method to connect to the toolbar buttons

events, and set the OnClick event to the TestClick event handler we
wrote earlier. Well finish by making sure that the toolbar is visible:
DICommandBar.Set_Visible(True);
TLIBIMP creates a read-only property for the Visible property of

the CommandBar interface, but it creates a Set_Visible method that
we use here. Youll see the same technique used in the property
implementation of TButtonServer. (My guess is that this is a bug in
TLIBIMP). Now well use a similar technique to add a new menu
item. First well create the OnClick event handler for the menu item
(see Figure 4). This event handler iterates the first paragraph of
the selected text between the available border styles. Here we take
advantage of the specific Word application capabilities.
In OnStartupComplete, we add the code in Figure 5 to obtain an
interface pointer to the Tools menu, search it for the existence of
our menu item, and, if it doesnt exist, add it. Well later create a
TButtonServer wrapper around it, and connect it to the MenuClick
event handler (see Figure 5).
Figure 6: Word 2000 displaying the toolbar created by our

add-in. Notice the first paragraph border created by the Tools |
Delphi menu option, which our add-in added to Word.
Notice the use of the FindControl method of the CommandBar

interface to look for a specific control based on a unique tag
assigned to it. If the control is found, it will be set to MenuIntf ;
otherwise MenuIntf wont be assigned, and well create the new
control (menu item). Figure 6 shows the toolbar that weve created
for Word with our add-in.
Delphi at Work
if (Assigned(DIBtn)) then begin
DIBtn.Free;
DIBtn := nil;
end;
if (Assigned(DIMenu)) then begin
DIMenu.Free;
DIMenu := nil;
end;
if (Assigned(DICommandBar)) then begin
// This is an interface, not an object!
DICommandBar.Delete;
DICommandBar := nil;
end;
Figure 7: Cleaning the user-interface elements with the

OnBeginShutdown method.
Cleaning Up
I used the OnBeginShutdown method to clean up the user interface

elements we created (see Figure 7). Notice that we free the
TButtonServer wrappers and the CommandBar interface.
Sharing Add-in Code among Office Applications
Because the add-in architecture is identical among Office applications,

you can use the same physical OLE Server DLL to serve multiple
applications. The trick is to determine which application is the one that
activated the add-in, and use the appropriate object model. The easiest
way to determine the application that activated the add-in is to assign
the Application IDispatch pointer thats passed to the OnConnection
method to an OleVariant variable, and use the Name property to
determine what application activated the add-in:
var
AppVar : OleVariant;
begin
AppVar := Application;
if (AppVar.Name = 'Outlook') then
begin
...
end
else if (AppVar.Name = 'Microsoft Word') then
begin
...
end else ...
Conclusion
Weve discussed the ease of creating add-ins for Office 2000 applications. The new COM-based add-in architecture makes it easy to share
add-in code among Office 2000 applications. The same COM object
can be used as an add-in for more than one application, and the
same user-interface objects can be shared among Office applications.
Also, if you create toolbars, menu items, or Office assistant Help
in one application, the same code will work in other applications.
For more information about Office development and the creation
of Office 2000 add-ins, consult the microsoft.public.officedev newsgroup hosted on msnews.microsoft.com.
The files referenced in this article are available on the Delphi
Informant Magazine Complete Works CD located in INFORM\00\
MAY\DI200005RL.
Ron Loewy is a software developer for HyperAct, Inc. He is the lead developer
of eAuthor Help, HyperActs HTML Help authoring tool. For more information
about HyperAct and eAuthor Help, contact HyperAct at (515) 987-2910 or visit
http://www.hyperact.com.
Dynamic Delphi
User Interface / Run-time Objects
By Ron Gray
Dynamic Forms
Creating Forms and Controls at Run Time
ne of Delphis strongest features is how easy it is to create user interfaces using

the VCL. Forms, components, and controls are created visually at design time, and
the VCL generates the code to create and display them at run time. These static forms
satisfy most UI requirements. When they dont, Delphi still makes it easy to create forms
and controls dynamically at run time based on run-time criteria. This opens exciting
possibilities for most applications.
Data-driven interfaces are the most common
reason for using dynamic forms and controls.
They can vary from simple mechanisms, such as
dynamic dialog boxes, to refined processes that
read menus, data entry forms, reports, and other
settings from information stored on the disk. In
other words, the data drives the application. As
long as the properties are known, creating forms
and controls is straightforward.
Creating Forms Dynamically
Use TForm to create most forms, such as a main

window, dialog box, or MDI child. The following
example creates an instance of the TForm component and calls the Show method:
procedure NewForm;
var
frmForm: TForm
begin
frmForm := TForm.Create(Application);
frmForm.Show;
end;
This example creates and displays a form, but

what will it look like? No properties are set,
Property
Possible Values
BorderIcons
BorderStyle
FormStyle
Height
Left
Position
biSystemMenu, biMinimize, biMaximize, biHelp

bsNone, bsSingle, bsSizeable, bsDialog, bsToolWindow, bsSizeToolWin
fsNormal, fsMDIChild, fsMDIForm, fsStayOnTop
<integer>
<integer>
poDesigned, poDefault, poDefaultPosOnly, poDefaultSizeOnly,
poScreenCenter, poDesktopCenter
<integer>
<integer>
WsNormal, wsMinimized, wsMaximized
Top
Width
WindowState
Figure 1: Properties of TForm that determine the type of window, its size, and position.
so the form is displayed using all the default

values. Its not clear what type of window it is,
how big it is, or where it appears on the screen.
These are basic properties (listed in Figure 1)
that should be set before displaying the form or
adding components to it.
Using the properties listed in Figure 1, the following example displays a modal, stay-on-top dialog
box in the center of the screen:
frmForm := TForm.Create(Application);
with frmForm do begin
BorderStyle := bsDialog;
FormStyle := fsStayOnTop;
Height := 185;
Width := 450;
Position := poScreenCenter;
end;
frmForm.ShowModal;
This code is much better than the previous example because it clearly provides pertinent information about the form, and doesnt rely on guesswork about default properties. Of course, there are
many other properties that can be set to further
customize the form. As a general rule, all the properties a developer would set at design time must be
set programmatically at run time. Properties that
are rarely changed from their defaults should be
set programmatically as well, to promote clarity
and avoid surprises.
Parent and Owner
The blank form created previously needs some

controls to give it some functionality. Before
discussing how to create controls dynamically,
however, its important to note the difference
between the Parent property declared in
Dynamic Delphi
TControl, and the Owner property declared in TComponent. The
Parent is always a windowed control that visually contains the control or form. The Owner is passed as a parameter in the constructor
and determines when the component is freed. The former property
is used for display purposes, the latter for memory management.
Both are important.
TControl.Parent
Many forms do not have a Parent. For example, forms that appear
directly on the Windows desktop (like the dialog box in the previous
example) have Parent set to nil. However, a form can be embedded
in any other visual control by setting the Parent to that control. For
example, the same form that can be shown as a stand-alone form on
the desktop can also be embedded in a tab sheet by simply changing
the Parent property to the tab sheet.
For controls, the Parent is usually the form on which its displayed.
But it could also be a panel, group box, or some control that is
designed to contain another. When creating a new control, always
assign a Parent property value for the new control.
TComponent.Owner
The Owner of a control or form determines when the memory used

by the control can be freed. Its passed to the constructor when
creating a new form or control. When a component is destroyed, all
of the components owned by it are also destroyed. Usually, the form
is the Owner of all the controls on it, so when the form is destroyed,
all the controls are destroyed with it. Likewise, most stand-alone
forms are owned by the Application object, so they can be freed
when the application is terminated.
Creating Controls Dynamically
Creating controls dynamically is not much different from creating

forms. Each new control must have an Owner (passed to the constructor and usually the form), and a Parent (usually the form, but can
also be a panel, tab sheet, or other container control). Set additional
properties to further customize the control.
The following example displays a button on the bottom right of
the form:
btnButton := TButton.Create(frmForm);
with btnButton do begin
Parent := frmForm;
Height := 25;
Width := 75;
Left := frmForm.ClientWidth - 85;
Top := frmForm.ClientHeight - 35;
Caption := 'Hey now!';
ModalResult := 1;
Show;
end;
There are a few properties of interest here. First, note that the buttons
position is relative to the form. The dimensions of the form can
change, but the button will still appear on the bottom right of the
form. When positioning controls relative to the form, use the client
area (ClientHeight and ClientWidth) rather than the actual form area
(Width and Height), which includes the forms caption and other
space that cant be used by the client. Panels and anchors are also
useful for positioning controls.
Second, if the form is shown modally, the ModalResult property can
be used to determine whether, and how, the button closes the form.
Third, the controls Parent is the form, but it could have been a previously
created panel or other containing control. For example, a panel could be
created dynamically and anchored to the bottom of the form. The panel
would then be assigned to the buttons Parent so it will resize along with
the panel. Finally, the Show method is only required if the buttons Parent
(the form in the previous example) is already visible. Otherwise, when the
form is shown, it will cause all child controls to be shown as well.
A Simple Example
Using the techniques previously described, a custom function

MessageDlgEx can be written to display a dialog box like Delphis
MessageDlg function, except that the button captions can be defined. The
function is shown in Listing One beginning on page 21. (The function,
and all other source discussed in this article, is available for download;
see the end of this article for details.) This allows new options to be
displayed rather than relying on the standard dialog box buttons, such as
OK, Yes, No, and so on. For example, the following code prompts for a
decision from the user:
nResult := MessageDlgEx('Update Available',
'New components are available for updating.',
mtWarning, ['Update Now', 'Remind Me Later',
'Tell Me More']);
Calling this function displays a modal dialog box like the one shown
in Figure 2.
The function performs some simple math to determine the correct
widths of the buttons and form, based on the widths of the button captions and message. Once the form and button widths are determined,
the function simply loops through the array and creates a button for
each item. The buttons ModalResult is its one-based position in the
array, as shown in Figure 3. The Return value is the ordinal position of
the clicked button. Code can be written to respond to each button.
Working with Messages
Creating controls at run time is easy, but dynamically responding to

their events is more difficult. Generally, actions and basic functionality
must be coded in the application, and events and controls are simply
Figure 2: The MessageDlgEx function creates the form and buttons dynamically.
for ii := 0 to High(AButtons) do begin
btnButton := TButton.Create(frmForm);
btnButton.Height := 25;
btnButton.Width := nButtonWidth;
btnButton.Left :=
frmForm.Width - ((ii + 1) * (nButtonWidth + 10));
btnButton.Top := frmForm.ClientHeight - 35;
btnButton.Caption := AButtons[ii];
btnButton.ModalResult := ii + 1;
btnButton.Parent := frmForm;
end;
Figure 3: Once the form and button widths are determined, the
function simply loops through the array and creates a push button
for each item.
Dynamic Delphi
linked to pre-defined actions. No matter how generic and dynamic
a form is, it still needs a save procedure. The MessageDlgEx function
mentioned previously dynamically responds to events by simply assigning each buttons ModalResult property to its ordinal position in the
array. How the event is then handled is outside the functions control.
With more complex data entry forms, buttons must be linked to
specific actions, such as the OK button that calls a save routine, and
the Cancel button that performs any rollback procedures. Assigning
a specific action to a control requires code that responds to the
controls events, or overrides the controls message handler to respond
to messages. The easiest way is to respond to the controls events.
Delphi converts most Windows messages sent to the control to events.
At design time, developers link the controls events to methods of the
form. For example, the OK button, named btnOK, responds to the
OnClick event by calling the btnOKClick method of the form. Event
handlers can be assigned programmatically. However, the compiler
expects a method of a class rather than a stand-alone procedure, so the
method itself cannot be created programmatically. In other words, the
OK buttons OnClick event cannot simply call a generic SaveChanges
function, but rather must call a method of the form. An easy way
around this is to visually create a blank template form that is used
dynamically at run time, and write generic event handlers that can be
used by the various controls that are created programmatically.
For example, the OnClick event of an OK button must call the
SaveChanges function. The forms code defines a generic OnClick
event for all buttons. Within the method, specific actions can be
programmed for each button. So, if the OK button generated the
event, the SaveChanges function is called:
procedure TForm1.ButtonClick(Sender: TObject);
begin
if TControl(Sender).Name = 'btnOK' then
SaveChanges
else
ShowMessage('You did not click the OK button.');
end;
When the OK button is created, simply assign the OnClick handler

as in the following code:
oButton := TButton.Create(Self);
with oButton do begin
Name := 'btnOK';
Height := 25;
Width := 110;
Left := 240;
Top := 16;
OnClick := ButtonClick;
Caption := 'OK';
Parent := Self;
end;
be linked to specific fields using the various data access components, so

the order of their creation and linking is important. A rule of thumb is
to simply program in the same sequence that would normally be done
visually. The basic steps to creating a simple data entry form are:
1) Create a TTable component and assign the DatabaseName and
TableName properties.
2) Create a TDataSource component and set the DataSet property
to the TTable object.
3) Create TDBEdit controls and set the DataSource property to the
TDataSource object, and DataField property to the field name.
4) Set TTable.Active to True.
5) Provide OK and Cancel buttons, or a navigation bar for posting
or canceling changes.
The code in Figure 4 can be called from a method of a blank form
to create data access components and controls on the form for data
entry. The example assumes there is an alias named DBDEMOS
pointing to Delphis sample database.
The code in Figure 4 creates the data entry screen shown in Figure 5.
var
tblTable: TTable;
dsDataSource: TDataSource;
dbNavigator: TDBNavigator;
oEdit: TDBEdit;
oLabel: TLabel;
begin
tblTable := TTable.Create(Self);
tblTable.DatabaseName := 'DBDEMOS';
tblTable.TableName := 'animals.dbf';
dsDataSource := TDataSource.Create(Self);
dsDataSource.Name := 'dsDataSource';
dsDataSource.DataSet := tblTable;
dbNavigator := TDBNavigator.Create(Self);
dbNavigator.Parent := Self;
dbNavigator.Left := 8;
dbNavigator.Top := 12;
dbNavigator.DataSource := dsDataSource;
oLabel := TLabel.Create(Self);
oLabel.Parent := Self;
oLabel.Left := 8;
oLabel.Top := 52;
oLabel.Caption := '&Name:';
oEdit := TDBEdit.Create(Self);
oEdit.Parent := Self;
oEdit.Left := 100;
oEdit.Top := 52;
oEdit.DataSource := dsDataSource;
oEdit.DataField := 'NAME';
tblTable.Active := True;
end;
Figure 4: Creating data access components and controls on a form.
Other generic event handlers can be defined in a template form, such as

ListBoxClick, ListBoxSelect, and others. By defining generic event handlers
at the form level, its easy to respond dynamically to most events.
Working with Data
One of the more compelling reasons for creating forms and controls
dynamically is the ability to generate data entry screens at run time
that are either designed by the user, or based on some other data-driven
criteria. Creating data access components and data controls is no different
than what has already been described. Of course, the data controls must
Figure 5: A dynamically created data entry form.
Dynamic Delphi
Of course, this is a simple data entry example, but the same techniques apply when creating more complex forms. For example, a selfconfiguring data entry form can simulate Delphis Form Wizard at
run time, automatically creating controls for each field in a selected
table. This requires looping through TDataSet.Fields, creating a control and associated label for each field in the data set (see Figure 6).
Note that the Caption of the associated TLabel control is set to the
field name. This is fine for small utility operations, but when using datadriven forms within an application, the applications framework should
provide a way to save and restore the DisplayName of each field. This
way, a more descriptive label is provided, rather than using the field
name.
Data validation is straightforward as well. Values can be validated
in the application before they are sent to the database server. For
field-level validation, use TField.EditMask to restrict data that can be
entered in the field, and TField.OnValidate to validate data before
the record is updated. To reject the current value of the field, raise
an exception in the event handler. For record-level validation, use
TDataSet.BeforePost, which is called just before posting the record.
Call Abort to cancel the Post operation.
Working with Other Components
Most components can be created dynamically with no problem. Menus,

grids, OLE containers, ActiveX controls, and many others can all be created programmatically at run time. This means that just about any data
entry form that can be created visually can also be created dynamically.
Of course, the Parent and Owner properties are still very important. And
each component has unique properties that must be set.
The Controls and Components Arrays
Managing controls created dynamically is somewhat different from

managing controls created at design time. When created at design time,
controls are usually given names that help define what they are. For
example, its obvious that a control named btnOK is the OK button.
The developer can easily write code that specifically references this
button. Dynamically created controls tend to be more generic and
should be. So its not always possible to refer to controls specifically by
name. Fortunately, components and controls are stored in arrays that
for ii := 0 to Table1.FieldCount-1 do begin
LabelCaption := Table1.Fields[ii].FieldName;
DataType := Table1.Fields[ii].DataType;
// ... create associated TLabel control using field name.
if DataType = ftString then
// Create a TDBEdit control.
if DataType = ftMemo then
// Create a TDBMemo control.
// ... and so on.
end;
Figure 6: Looping through TDataSet.Fields.

var
I: Integer;
Temp: TComponent;
begin
for I := ComponentCount - 1 downto 0 do begin
Temp := Components[I];
if not (Temp is TControl) then begin
RemoveComponent(Temp);
DataModule2.InsertComponent(Temp);
end;
end;
end;
Figure 7: Using the Components array.

can be accessed to help manage dynamically created controls.
The Controls Array
The TWinControl.Controls property returns an array of all controls

for which the windowed control is the Parent. When the Parent is
assigned to a control, Delphi automatically updates the Controls array
of the previous Parent and new Parent. This is a read-only property,
but it can be updated by simply changing the Parent of a control.
Using this array its quite easy to iterate through all the controls
without having to refer to each of them by name. For example,
the following code loops through the Controls array to disable all
controls:
for ii = 0 to ControlCount - 1 do
if Controls[ii] is TControl then
Controls[ii].Enabled := False;
The Components Array
TComponent.Components provides access to all components owned

by the component. The sample in Figure 7, taken from Delphis
Help file, moves any non-visual components on the form into a
separate data module.
Conclusion
The ability to create forms and controls dynamically opens exciting

possibilities for applications. Examples range from simple dialog
boxes, such as the MessageDlgEx function described in this article, to
elaborate data-driven menus and data entry forms.
Forms and controls have several key properties that must be set when
they are created dynamically. The Parent and Owner properties are
important for display and memory management purposes. Messages
can be trapped by defining generic event handlers to an empty form,
or by overriding the controls message handler. Finally, there are
techniques, properties, and methods that help manage controls.
MAY\DI200005RG.
Ron Gray is a software developer specializing in business database applications.

He has written numerous articles using different languages and is the author of
LookUp Manager, a collection of components for visually managing lookup codes
and abbreviations in applications. He can be reached at rgray@compuserve.com.
Begin Listing One The MessageDlgEx Function

function MessageDlgEx(const Caption, Msg: string;
AType: TMsgDlgType; AButtons: array of string): Word;
var
oForm: TForm;
oLabel: TLabel;
oButton: TButton;
nButtonWidth: Integer;
nAllButtonsWidth: Integer;
nCtrlHeight: Integer;
nMessageWidth: Integer;
ii: Integer;
begin
// Create the form.
Dynamic Delphi
oForm := TForm.Create(Application);
oForm.BorderStyle := bsDialog;
oForm.BorderIcons := oForm.BorderIcons - [biSystemMenu];
oForm.FormStyle := fsStayOnTop;
oForm.Height := 185;
oForm.Width := 450;
oForm.Position := poScreenCenter;
oForm.Caption := Caption;
// Loop through buttons to determine the longest caption.
nButtonWidth := 0;
for ii := 0 to High(AButtons) do
nButtonWidth := Max(nButtonWidth,
oForm.Canvas.TextWidth(AButtons[ii]));
// Add padding for the button's caption.
nButtonWidth := nButtonWidth + 10;
// Determine space required for all buttons.
nAllButtonsWidth := nButtonWidth * (High(AButtons) + 1);
// Each button has padding on each side.
nAllButtonsWidth := nAllButtonsWidth +
(10 * (High(AButtons) + 2));
// The form has to be at least as wide as the buttons.
if nAllButtonsWidth > oForm.Width then
oForm.Width := nAllButtonsWidth;
// Determine if the message can fit in the form's width,
// or if it must be word wrapped.
nCtrlHeight := oForm.Canvas.TextHeight('A') * 3;
nMessageWidth := oForm.Canvas.TextWidth(Msg);
// If the message can fit with just the width of the
// buttons, adjust the form width.
if nMessageWidth < nAllButtonsWidth then
oForm.Width := nAllButtonsWidth;
if nMessageWidth > oForm.ClientWidth then begin
// Determine how many lines are required.
nCtrlHeight := Trunc(nMessageWidth/oForm.ClientWidth);
// Add 3 more lines as padding.
nCtrlHeight := nCtrlHeight + 3;
// Convert to pixels.
nCtrlHeight :=
nCtrlHeight * oForm.Canvas.TextHeight('A');
end;
// Adjust the form's height accomodating the message,
// padding and the buttons.
oForm.Height :=
nCtrlHeight + (oForm.Canvas.TextHeight('A') * 4) + 22;
// Create the message control.
oLabel := TLabel.Create(oForm);
oLabel.AutoSize := False;
oLabel.Left := 10;
oLabel.Top := 10;
oLabel.Height := nCtrlHeight;
oLabel.Width := oForm.ClientWidth - 20;
oLabel.WordWrap := True;
oLabel.Caption := Msg;
oLabel.Parent := oForm;
// Create the pushbuttons.
for ii := 0 to High(AButtons) do begin
oButton := TButton.Create(oForm);
oButton.Height := 25;
oButton.Width := nButtonWidth;
oButton.Left :=
oForm.Width - ((ii + 1) * (nButtonWidth + 10));
oButton.Top := oForm.ClientHeight - 35;
oButton.Caption := AButtons[ii];
oButton.ModalResult := ii + 1;
oButton.Parent := oForm;
end;
Result := oForm.ShowModal;
end;
End Listing One
The API Calls
Active Directory / Windows NT / Windows 2000
By Simon Murrell
Active Directories
Using ADSI on Your Windows NT and Windows 2000 Systems
ctive Directory is the directory service used in Windows NT 4.0 and Windows 2000.
Its also the foundation of Windows 2000. To access the Active Directory Service,
you need to use the API that Microsoft provides, named ADSI (Active Directory Service
Interfaces). ADSI is a set of COM interfaces used to access the different directory services.
Programmers can currently access four network directory structures using the providers
supplied in ADSI: WinNT (Microsoft SAM database), LDAP (Lightweight Directory Access
Protocol), NDS (NetWare Directory Service), and the NWCOMPAT (Novell NetWare 3.x).
ADSI now makes the Windows NT administrators job easier. ADSI allows the administrator to
perform common tasks, such as the addition of
new users, managing printers, security settings, and
controlling the NT domain. Because ADSI uses
COM interfaces, well-known languages, such as
Visual Basic, Visual C++, C++Builder, Delphi, or
any other COM-enabled language, can use ADSI
to write new client software. Well-known ISVs
have used ADSI to make their software applications directory-enabled.
Active Directory runs on either Windows NT
4.0 or Windows 2000. Client applications can
run on Windows 95, 98, NT 4.0, and 2000.
To use ADSI, you need to install the ADSI
COM interfaces. You can download the ADSI
2.5 SDK from the Microsoft ADSI Web site at
http://www.microsoft.com/adsi. The SDK contains documentation, online help, and samples.
The only problem is that the samples and documentation are aimed toward Microsoft products,
namely Visual Basic and Visual C++. This is one
of the main reasons Im writing this article: so
there are examples for the Delphi community.
The documentation is also mainly aimed at the
WinNT provider not the more common
LDAP provider used in Microsoft Exchange and
Microsoft Site Server.
The Demo Application
The sample application I wrote, shown in Figure

1, demonstrates how to use some of the functionality provided by the WinNT provider (this
application is available for download; see end
of article for details). The sample application
is used to connect to a domain within your
organization. Once connected to the domain,
the sample application will list the NT users
and groups found on the PDC, along with the
computers participating in the domain. Ill also
demonstrate how to view services from the NT
computers found within the domain, along with
adding and removing users and viewing users
found within the NT groups.
Using ADSI to Control Windows NT/2000
Figure 1: Sample application.

The WinNT provider allows the developer to

control and manipulate the Windows NT SAM
database. You can use WinNT to gain access to
resources, such as users, groups, computers, file
shares, printer jobs, printer queues, and services.
To use ADSI from a Delphi application, you need
to import the Active Directory type library. To do
this, select the Project | Import Type Library option
from the menu, select ActiveDs (Version 1.0), and
click OK (see Figure 2). Delphi will create the
type library declaration .pas file, and import it
into your project.
The API Calls

// Procedure is used to bind to the ADSI WinNT directory.
procedure TMainFrm.actOpenWinNTExecute(Sender: TObject);
var
UnknownObject: IUnknown;
DomainPath: WideString;
Domain: IADsContainer;
begin
// Assign domain path.
DomainPath := 'WinNT://' + ADSIDomainName.Text;
// If login details are used then.
if cbUseLogin.Checked then
// Use login details and create domain object.
OleCheck(AdsOpenObject(PWideChar(DomainPath),
PWideChar(ADSIUsername.Text),
PWideChar(ADSIPassword.Text), 0, IID_IADsContainer,
UnknownObject));
else
// Create domain object.
OleCheck(ADsGetObject(PWideChar(DomainPath),
IID_IADsContainer, UnknownObject));
// Assign domain object.
Domain := UnknownObject as IADsContainer;
// Get Lists from Domain.
GetDomainInformation(Domain);
end;
Figure 2: Importing the Active Directory type library.
Binding to the WinNT Directory Service
Connecting to the WinNT directory service is simply a matter of

finding the domain controller, and then binding to the required
object. To bind to the WinNT provider in Delphi, you can use the
ADsGetObject or ADsOpenObject function.
The ADsGetObject function is used to bind to an ADSI object by
specifying the objects path, along with the interface identifier and
object. The declaration of this function is:
function ADsGetObject(lpszPathName: PWideChar;
const riid: TIID; out obj): HResult; stdcall;
external 'activeds.dll';
The first parameter is the path name used to bind to the object
in the underlying directory service. The second parameter is the
interface identifier for a specified interface on this object; and the
third parameter is the indirect pointer to the requested interface.
By default, this function uses secure authentication, i.e. the function uses the security context of the current user.
The second function, ADsOpenObject, is used to bind to an ADSI
object using an alternate security context by stipulating a username
and password of the required user. The declaration of this function is:
function ADsOpenObject(lpszPathName: PWideChar;
lpszUserName: PWideChar; lpszPassword: PWideChar;
dwReserved: LongInt; const riid: TIID; out obj): HResult;
stdcall; external 'activeds.dll';
The first parameter is the path name used to bind to the object in
the underlying directory service. The second and third parameters
are the username and password of the user whose security credentials you want to use. The fourth parameter is a reserved provider
flag, which stipulates the authentication method you want to bind
with. The fifth parameter is the interface identifier for a specified
interface on this object. Finally, the sixth parameter is the indirect
pointer to the requested interface.
Figure 3: Binding to the ADSI object using the default security

credentials or alternate credentials.
Thus, the first function defaults to the logged users credentials,

and the second function allows the developer to specify security
credentials to bind to the ADSI object. Figure 3 shows the procedure
used to bind to the ADSI object, by using either the default security
credentials or alternate credentials.
Lets step through the procedure to see exactly whats going on.
First, three variables are declared. The first is the interface variable,
which the binding functions are going to return from the specified
object path:
Second is a WideString variable, which is used to generate an object

path in the binding functions:
DomainPath: WideString;
Third is an IADsContainer variable, which will be assigned to the

returned interface variable:
Domain: IADsContainer;
The IADsContainer variable is going to be used to retrieve all

the users, groups, and computers from the ADSI object specified.
Alternatively, you can declare this variable as IADsDomain, but I
want to enumerate the children objects found within the domain.
(Read the SDK to find out the exact differences between the
IADsContainer and the IADsDomain.)
The following statement assigns the path for the object we want to
retrieve. If your domain name is PRISMA, for example, to retrieve
the domain ADSI object you would need to assign the path as
WinNT://PRISMA:
// Assign domain path.
DomainPath := 'WinNT://' + ADSIDomainName.Text;
The API Calls

The next section of code uses an alternate security credential, or the
default logged on security credential:
// Assign temporary object.

ADsObj := IUnknown(ADsTempObj) as IADs;
// If login details are used then.

if cbUseLogin.Checked then
// Use login details and create domain object.
OleCheck(AdsOpenObject(PWideChar(DomainPath),
PWideChar(ADSIUsername.Text),
PWideChar(ADSIPassword.Text), 0, IID_IADsContainer,
UnknownObject));
else
// Create domain object.
OleCheck(ADsGetObject(PWideChar(DomainPath),
IID_IADsContainer, UnknownObject));
Once the ADSI object has been assigned, we check the child objects
class. According to the type of class, we then pass the ADSI to the
respective procedure to add the properties from the ADSI object to
the respective ListView component on the form:
Then we assign the indirect pointer to the IADsContainer, so we can

query the domain for children objects:
// Assign domain object.
Domain := UnknownObject as IADsContainer;
Finally, we pass the IADsContainer to the procedure that retrieves the

child objects within the domain:
// Get Lists from Domain.
GetDomainInformation(Domain);
Searching the Domains IADsContainer
To search through the domain for child objects, we use the

GetDomainInformation procedure. We pass the IADsContainer that
we retrieved from the binding procedure. The source code to search
the IADsContainer is shown in Figure 4.
The following is a detailed explanation of the source code in Figure
4. We declare the following variable so we can enumerate the child
objects with the container. Well use this variable to search through
the children variables in the container:
Enum: IEnumVariant;
The following is a temporary variable used to store the children

retrieved from the container object:
ADsTempObj: OLEVariant;
The following is the interface variable of the children in the container object:
// If object is a user object then.

if AdsObj.Class_ = 'User' then
AddUserToList(ADsObj);
// If object is a group object then.
if AdsObj.Class_ = 'Group' then
AddGroupToList(ADsObj);
// If object is a computer object then.
if AdsObj.Class_ = 'Computer' then
AddComputerToList(ADsObj);
Once youve assigned all the values from the child objects, your
application should be filled with information, as in Figure 5.
// Procedure retrieves the domain information.
procedure TMainFrm.GetDomainInformation(
Domain: IADsContainer);
var
Enum: IEnumVariant;
ADsTempObj: OLEVariant;
ADsObj: IADs;
Value: LongWord;
begin
// Empty User, Group, and Computer lists.
UserListView.Items.Clear;
GroupListView.Items.Clear;
ComputerListView.Items.Clear;
// Assign enumerator object.
Enum := (Domain._NewEnum) as IEnumVariant;
// Search through enumerator object.
while (Enum.Next(1, ADsTempObj, Value) = S_OK) do begin
// If object is a user object then.
if AdsObj.Class_ = 'User' then
AddUserToList(ADsObj);
// If object is a group object then.
if AdsObj.Class_ = 'Group' then
AddGroupToList(ADsObj);
// If object is a computer object then.
if AdsObj.Class_ = 'Computer' then
AddComputerToList(ADsObj);
end;
end;
Figure 4: Source code to search the IADsContainer.
ADsObj: IADs;
We now assign the enumerator object of the container to the Enum

variable, so we can begin searching the container:
Enum := (Domain._NewEnum) as IEnumVariant;
Once weve assigned the variable, we begin searching through the

enumerator variable, assigning each child object to the temporary
OLEVariant object:
This OLEVariant object reference is then assigned to the ADSI object:

Figure 5: Your application should be filled with information.
The API Calls

Creating and Removing Users from Computers
With the WinNT provider, one can create and remove users from any
computer within any domain. To create a user within the domain on
a specified computer, you need to bind to the computer to which you
want to add the user. Once youve bound to the ADSI container object
of the required computer, you need to call the Create method. The Create
method of the container object takes two arguments. One is the type of
ADSI object you want to create; the other is the name describing the
new ADSI object. The Create method then returns a reference to the new
ADSI object created. Figure 6 is the source code that creates the user.
As usual, first the variables are declared. The first is the computer
container variable to bind to. From this container, well add the new
user to the domain:
ComputerObj: IADsContainer;
The following is a temporary interface variable, which is used to store

the newly created ADSI object:
TempUserObj: IUnknown;
The following is the IADsUser variable, which were going to assign

from the newly created ADSI user object. It will enable the application to represent and manage the end-user account on the domain:
UserObj: IADsUser;
var
TempUserObj: IUnknown;
UserObj: IADsUser;
PDCName: WideString;
NewUserName: WideString;
AdsPath: WideString;
begin
// Retrieve information from user.
PDCName := InputBox('Create New User',
'Please type in the name of the Domain's PDC : ', '');
NewUserName := InputBox('Create New User',
'Please type in the user name : ', '');
// Assign AdsPath.
AdsPath := 'WinNT:// + PDCName + ',computer';
// Create computer object.
OleCheck(AdsGetObject(PWideChar(AdsPath),
IID_IADsContainer, ComputerObj));
// Create new user.
TempUserObj := ComputerObj.Create('user', NewUserName);
UserObj := TempUserObj as IADsUser;
// Set information back to directory.
UserObj.SetInfo;
// Refresh list.
actOpenWinNT.Execute;
Figure 6: Creating a new user.

This WideString variable will be used to store the name of the PDC
(Primary Domain Controller) computer of the domain. You can specify
any other computer name, and the source code will add the user to the
computer, but the source code will only return the users from the domain:
We then call the Create method from the ADSI container object,
which will create the new user and return the new user object back to
the temporary ADSI object. We then assign the temporary interface
object to the ADSI user object:
// Create new user.

TempUserObj := ComputerObj.Create('user', NewUserName);
UserObj := TempUserObj as IADsUser;
The following WideString variable is used to store the name of the

new user one wishes to add:
We then save the user information back to the WinNT directory:
NewUserName: WideString;
The following WideString variable is going to store the ADSI

object path of the computer were going to bind to, e.g. WinNT://
PDCCOMP,computer:
In the following, we call the InputBox method to retrieve the names of

the PDC computer and the new user we want to add:
NewUserName := InputBox('Create New User',
'Please type in the user name : ', '');
Here, we assign the ADSI object path of the computer that we want
to bind to:
// Assign AdsPath.
AdsPath := 'WinNT://' + PDCName + ',computer';
We then create the new IADsContainer object from the specified

ADSI object path:
// Set information back to directory.

UserObj.SetInfo;
Then we refresh the list to demonstrate the new user addition:

// Refresh list.
Removing a user from a computer is a similar task, except we dont need

to create any user objects, and the user will call the IADsContainers Delete
method (instead of the Create method), as shown in Figure 7.
The Delete method has two parameters: one is the type of ADSI
object to delete; the second specifies the name of the user to delete:
// Create new user.
ComputerObj.Delete('user', UserName);
Viewing Users in Groups
With the WinNT provider, you can also control the NT groups by
adding and removing users from the groups, and performing other
group maintenance tasks. In the example application, Ive decided to
enumerate the group and view the users found within the groups in
the specified domain. The code in Figure 8 binds to an IADsGroup
The API Calls

object, and then uses an IADsMember object that is an enumeration
object used to view and control the users within the IADsGroup. The
code demonstrates listing the users found within different groups.
Members is an interface variable for managing the list of users or

ADSI objects with the NT group:
Members: IADsMembers;
The first variable is IADsGroup, which will be used to represent and

manage the NT group information from the directory. Well use the
Members property of the object, which is of type IADsMembers:
The following WideString variable is going to store the ADSI

object path of the group were binding to, e.g. WinNT://
PDCCOMP/Domain Administrators:
GroupObj: IADsGroup;
var
UserName: WideString;
begin
UserName := InputBox('Create New User',
'Please type in the user name to delete : ', '');
if MessageDlg('Are you sure you want to delete user : ' +
UserName + ' ?', mtConfirmation,
[mbYes, mbNo], 0) = mrYes then
begin
// Assign AdsPath.
AdsPath := 'WinNT:// + PDCName + ',computer';
// Create new user.
ComputerObj.Delete('user', UserName);
// Refresh list.
end;
Figure 7: Removing a user from a computer.

var
GroupObj: IADsGroup;
Members: IADsMembers;
Enum: IEnumVariant;
TempUserObj: OLEVariant;
UserObj: IADsUser;
TempListObj: TListItem;
Value: LongWord;
begin
// Clear List.
GroupListView.Items.Clear;
// Assign AdsPath.
AdsPath := 'WinNT:// + MainFrm.ADSIDomainName.Text +
'/' + GroupName;
// Create group object.
OLECheck(AdsGetObject(PWideChar(AdsPath), IID_IADsGroup,
GroupObj));
// Assign members.
Members := GroupObj.Members;
Enum := (Members._NewEnum) as IEnumVariant;
while (Enum.Next(1, TempUserObj, Value) = S_OK) do
try
UserObj := IUnknown(TempUserObj) as IADsUser;
// Create new list item.
TempListObj := GroupListView.Items.Add;
// Assign property.
TempListObj.Caption := UserObj.Name;
except
on E:Exception do
end;
Figure 8: Listing the users found within different groups.

We declare the following variable so we can enumerate the child

objects found within the IADsMembers container. Well use it to
search through the child objects in the container:
Enum: IEnumVariant;
This variable is a temporary interface variable used to store each

ADSI object found in the Members variable list:
TempUserObj: OLEVariant;
Well use the IADsUser variable to retrieve the end-user information

for child objects found in the NT group. It will enable the application
to represent and manage the end-user account on the domain:
UserObj: IADsUser;
We assign the ADSI object path of the group we want to bind to:
// Assign AdsPath.
AdsPath := 'WinNT://' + MainFrm.ADSIDomainName.Text +
'/' + GroupName;
We then create the new IADsGroup object from the specified ADSI
object path:
// Create group object.
OLECheck(AdsGetObject(PWideChar(AdsPath),
IID_IADsGroup, GroupObj));
Next, we retrieve the NT groups member list, and assign it to the

enumerator variable:
// Assign members.
Members := GroupObj.Members;
Enum := (Members._NewEnum) as IEnumVariant;
Then we search through the list of NT users, and assign each member
to the temporary interface variable:
while (Enum.Next(1, TempUserObj, Value) = S_OK) do
Next, we assign the temporary interface variable to the user object. Then
we create the list item and assign the users username to the list item:
UserObj := IUnknown(TempUserObj) as IADsUser;
// Create new list item.
TempListObj := GroupListView.Items.Add;
// Assign property.
TempListObj.Caption := UserObj.Name;
The API Calls

Once the search through the user list is complete, the result should
look similar to Figure 9.
of returning an IADsService ADSI object, you must return an

IADsServiceOperations ADSI object.
Controlling NT Services
For example:
The WinNT provider can also control NT services, NT servers, and

workstations. The code in Figure 10 is similar to code used to retrieve
the domain information. Instead of searching for users, groups, or
computers, however, the
code searches for NT services and assigns the services to the respective service list according to the
computer selected in the
computer list.
To retrieve information
from NT services, we
need to bind the ADSI
object representing the
service. ADSI provides the
IADsService object, which
allows us to maintain the
information about the NT
service running on its host
computer. The code in
Figure 11 binds to the
respective NT service and
displays the name and display name of the service.
// Create Computer Object.

IID_IADsServiceOperations, Result));
Next, bind to the IADsServiceOperations ADSI object using the

GetServiceObj function, which returns IADsServiceOperations:
// Assign service object.
ServiceObj := GetServiceObj;
If the service is stopped, then call the IADsServiceOperations Start

method, which starts the NT service:
AdsPath := 'WinNT://' + ComputerName + '/' + ServiceName;
var
Computer: IADsContainer;
ComputerPath: WideString;
Enum: IEnumVariant;
AdsTempObj: OLEVariant;
AdsObj: IADs;
Value: LongWord;
begin
if Item.Caption = '' then
Exit;
// Assign computer path.
ComputerPath := 'WinNT:// + ADSIDomainName.Text +
'/' + Item.Caption;
OleCheck(ADsGetObject(PWideChar(ComputerPath),
IID_IADsComputer, UnknownObject));
// Assign computer object.
Computer := UnknownObject as IADsContainer;
// Remove items from list.
ServiceListView.Items.Clear;
Enum := (Computer._NewEnum) as IEnumVariant;
// If object is a service object then.
if AdsObj.Class_ = 'Service' then
AddServiceToList(ADsObj);
end;
Then we create the new IADsService object from the specified ADSI
object path:
Figure 10: This code searches for NT services and assigns the
services to the respective service list according to the computer
selected in the computer list.
Figure 9: Listing users from a group.
First, an IADsService variable is declared to represent the NT service

information from the directory:
ServiceObj: IADsService;
This WideString variable is going to store the ADSI object

path of the NT service were going to bind to, e.g. WinNT://
PDCCOMP/Messenger:
We assign the ADSI object path of the service we want to bind to:
// Assign Service Object.

OLECheck(ADsGetObject(PWideChar(AdsPath),
IID_IADsService, ServiceObj));
We then display the properties retrieved from the respective NT

service:
lblServiceName.Caption := 'Service Name : ' + ServiceName;
lblDisplayName.Caption := 'Service Display Name : ' +
ServiceObj.Get_DisplayName;
To start and stop the NT service, you need to use the

IADsServiceOperations ADSI object. You will bind to the NT
service the same way as in the previous code, except that instead
var
ServiceObj: IADsService;
begin
// Assign AdsPath.
AdsPath := 'WinNT://' + ComputerName + '/' + ServiceName;
// Assign Service Object.
OLECheck(ADsGetObject(PWideChar(AdsPath),
IID_IADsService, ServiceObj));
// Assign labels.
lblServiceName.Caption :=
'Service Name : ' + ServiceName;
lblDisplayName.Caption := 'Service Display Name : ' +
ServiceObj.Get_DisplayName;
Figure 11: This code binds to the respective NT service and

displays its name and display name.
The API Calls
Figure 12: Controlling NT services.

// Start service.
if ServiceObj.Get_Status = 1 then
ServiceObj.Start;
To stop the service, use the Stop method of the IADsServiceOperations

ADSI object. First, as before, bind to the IADsServiceOperations ADSI
object by using the GetServiceObj function found, which returns
IADsServiceOperations:
// Assign service object.
ServiceObj := GetServiceObj;
If the service is running, call the IADsServiceOperations Stop method,

to stop the NT service.
The dialog box in Figure 12 demonstrates the NT service functionality mentioned here:
// Start service.
if ServiceObj.Get_Status = 1 then
ServiceObj.Start;
Conclusion
In this article, we explored the ways in which the ADSI

directory services can be put to use. With the demonstration
application included in this article, we were able to see functionality in the WinNT provider that makes life easier for developers
and administrators.
ADSI is the way of the future for Windows 2000. It has done for
directories what ADO did for databases. Its up to you to
find ways of tapping into this resource and making it work to
your benefit.
MAY\DI200005SM.
Simon Murrell is currently the Senior Developer at Galdon Data, a well-known

Microsoft Solutions Provider. He has used languages such as Delphi, VB, C++,
Java, and JavaScript in his professional career. Simon can be reached at
simonm@galdon.co.za.
Greater Delphi
VisiBroker / CORBA / IDL2PAS
By Eric Whipple
VisiBroker 3.3 for Delphi

Keeping Up with a CORBA World
know what youre thinking: Does the assembly method of development ever actually
work, or is it just something I told the vice president to make him feel better about
the value of distributed systems? The assembly method, of course, refers to the creation
of software by gluing together pre-made pieces of functionality to create a total
development solution. By taking advantage of the flexibility and power of interface-based
development, the CORBA (Common Object Request Broker Architecture) standard has
brought us one step closer to this development dream.
The question so many Delphi developers struggle
with is: Can Delphi hack it in a CORBA world?
The answer is, Yes. VisiBroker 3.3 for Delphi
firmly establishes Delphi as a major tool for creating powerful, Windows-based CORBA clients.
written with other development tools. Instead,

you were forced to manually code the clients
communication with the ORB. And that can
take you further into the guts of CORBA than
you want to go.
In the Beginning ...
A Trip to the Bank
Before the arrival of VisiBroker 3.3, Delphi

developers were already using Delphi to create
CORBA clients and servers. Using the CORBA
wizard to create a new CORBA object, and then
using the Type Library editor to add a few methods and properties to a new interface, is, in fact,
pretty easy. But in todays development community, a good CORBA client must be able to
play well with others.
The biggest barrier to creating a CORBA solution with Delphi is that theres no simple way to
create a Delphi client for a CORBA server written in Java or C++. If both the client and server
are written in Delphi, you can simply add the
xxx_TLB.pas file from the server object to the
client project, and not worry about the details of
CORBA communication. But what if the server
wasnt written in Delphi? One of the fundamental advantages of CORBA is that it is languageindependent. Until the release of VisiBroker 3.3,
however, Delphi did not provide a way to create
a CORBA client based on existing interfaces,
To see just how useful Delphi can be, lets look at

a real-world example of VisiBroker. A major problem facing larger companies today is how to deal
with corporate inertia. Strangely, Websters doesnt
define corporate inertia, but if it did, it would
probably look something like this:
cor.po.rate in.er.tia \kor-p-rt in-r-sh\ n - A
principle by which a corporation in a certain technological state will tend to remain in that state unless
acted upon by an external force. The magnitude of
the inertia is directly proportional to the mass of the
corporation that it acts upon.
As an example, imagine that youre the IT manager for a regional group of banks owned by
First American Bank of Chicago. First Americans main offices are located in Chicago, of
course, but youre managing one of its recently
acquired subsidiaries in the southeast. Your local
branches are equipped with the latest in ATM
hardware, which supports Windows-based applications. First American requires you to com-
Greater Delphi
municate with their CORBA servers, but has not developed latestgeneration ATM client software because many of its older subsidiaries cant purchase or run the applications. Its your responsibility
to create a quick and easy client application to connect to the First
American server.
In the past, First Americans corporate inertia would have
caused a serious problem for your Delphi developers. Because
Delphi had no way to create client code based on existing interface declarations, you wouldve been forced to develop
your Delphi CORBA clients from scratch by hand. This
includes the creation of marshalling and stub code for every
implementation object, which is tedious to say the least. Even if
your developers are intimately familiar with Delphi and CORBA,
the cost of writing the application (in developer hours) alone
could prohibit you from realistically attempting it (not to
mention debugging it).
A Better Way to CORBA

VisiBroker for Delphi allows you to automate the creation of
CORBA clients for existing CORBA servers by creating client-side
stubs based on interfaces and types defined in the servers IDL
file(s). IDL (Interface Definition Language) is a generic language
for specifying CORBA-compliant interfaces and complex data
structures. It can be generated by almost any development tool that
allows the creation of CORBA objects, e.g. Delphi. Because most
IDEs can generate IDL, Delphi can be used to create clients for
any CORBA object, no matter what language its written in. In our
bank example, VisiBroker for Delphi relieves you of the burden of
First Americans corporate inertia.
VisiBroker for Delphi uses an IDL2PAS compiler to create clientside Pascal CORBA code from any IDL file. In other words,
IDL2PAS creates client-side stubs (based on interface and object
definitions found in the IDL file) that allow a client application
to communicate with any CORBA object
that supports the listed interface. The key
is that no other information (aside from
the IDL file) is required from the server.
Inside VisiBroker 3.3
Its important to note that VisiBroker for

Delphi is not a Delphi ORB (object
request broker); that is, its not written
in Delphi. It uses the same ORB found
in VisiBroker for C++. A special DLL,
orbpas33.dll, provides a wrapper around the
C++ library, orb_br.dll. The main reason
for this approach is that the C++ ORB is
already one of the leading CORBA ORBs
out there, and it avoids the cost and timing
issues involved in writing a Delphi ORB
from scratch. (More information on this
topic can be found in the VisiBroker for
Delphi documentation.)
Figure 1: The BankServer.Idl file.
Figure 2: The BankServer_i.pas file after some condensing.

Probably the first thing youll notice

about VisiBroker for Delphi is that its
not installed as part of the Delphi
environment. In fact, its run from a
DOS window. The IDL2PAS command
can be used with a variety of flags to
tweak its behavior. After installing the
product, youll find two new files in
your bin directory: IDL2PAS.bat and
IDL2PAS.jar. The batch file is simply
used to run the jar file on whatever
IDL file you supply. As you might
have guessed, you need a Java Virtual
Machine (JVM) to run IDL2PAS. Most
Windows machines have at least one
JVM installed, so this shouldnt be a
problem.
Although the concept of IDL2PAS is
fairly simple, the importance of this
tool cannot be overstressed. Now if a
pre-packaged Java CORBA server exists,
all you need to write a customized client
are the IDL files to the CORBA objects
that you need to talk to. Because IDL is
Greater Delphi
a generic language, it doesnt matter what language the server was
written in. Lets see how this new tool affects our bank example.
We start by executing the IDL2PAS command on BankServer.Idl by
opening a DOS window and typing the following:
IDL2PAS BankServer.Idl
The BankServer.Idl file is shown in Figure 1. IDL2PAS generates

two new Pascal files: BankServer_i.pas and BankServer_c.pas. The
BankServer_i.pas file has a specific purpose. Its job is to declare (in
Pascal) all of the interfaces and any types defined in the IDL file.
Although the file appears large, its likely to contain more comments
than code. Each line from the IDL file is identified with a comment
in the resulting Pascal interface definition.
After some condensing, the BankServer_i.pas file looks like the
code shown in Figure 2. Notice that our IAccountManager has
magically become two interfaces. Not only is there a Pascal declaration for our initial interface, there is also a factory interface, named
AccountManagerFactory.
Again, I know what youre thinking: Why isnt it named
IAccountManagerFactory? The notion that all interfaces should begin
with the letter I is really more of a COM standard. Because
Delphi supports both COM and CORBA, it always puts the I
prefix on the servers it creates, whether
COM or CORBA. VisiBroker, of course,
is a CORBA implementation and therefore does not use the same convention.
of unnecessary COM structures, such as CoClasses. You will notice

some COM-like items, such as references to the QueryInterface
method, but according to documentation, this is only a thin layer:
... the TCorbaObject class definition has a COM-style with a
queryinterface, addref and release methods and a GUID. In reality, however, it only uses COM for automated garbage collection of the stub
objects. Under the covers the TCorbaObject implementation substitutes
CORBA-based calls in place of the operating system COM calls.
The xxx_c.pas file typically contains two types of classes: the helper
class and the stub class. The helper class (found in the xxx_c.pas
file) allows a client to make calls with non-primitive parameter types
(structs, arrays, and other non-standard types defined by the developer)
and allows for the typecasting of returned objects via the Narrow
method. In addition, the helper class contains two versions of the
all-important Bind method. This is what allows a client to create a
connection to a CORBA server. More information on the helper class
can be found in the VisiBroker for Delphi documentation. The stub
class is used to marshall method calls to the ORB. Youll see very similar
method calls for each member of your interface as the stub class creates,
formats, and invokes requests and responses. Its unlikely that youll
need to directly edit the contents of this class.
VisiBroker for Delphi also contains a program named IDL2IR.
The IDL2IR command populates an interface repository with
The AccountManagerFactory interface is

very simple. The job of any factory
object is to generate instances of another
object (in this case, IAccountManager).
Its declaration is shown in Figure 2. It
contains a single method named
CreateInstance, which returns a reference
to an IAccountManager object.
Looking again at Figure 1, we see that
the IDL file contains an enum declaration
named AccountType. This enum type will
be turned into a Pascal type declaration.
However, when Delphi creates IDL, it
always puts the enumerations at the
bottom of the file. If there are method declarations that pass parameters of that type,
the IDL2PAS command will fail. This is
because at the time the compiler tried to
resolve the parameter type, the type had
not yet been declared. The simple solution
is to open the IDL file and move the enum
declaration to the top.
The second file generated by IDL2PAS is
the BankServer_c.pas file. This file holds
all of the client-side stub code for communicating with CORBA objects on the
server. This file is similar in nature to the
implementation section of the traditional
xxx_TLB.pas file. A big difference between
the two, however, is the virtual absence
Figure 3: Bind to a factory object and request an IAccountManager object.
Figure 4: Then, we can begin making calls to the servers IAccountManager interface.
Greater Delphi
objects and other constructs contained in an IDL file. An interface
repository contains information about the ORB and any objects
with which it is currently communicating. This can be very useful
for dynamic binding.
Building a Client
The downloadable code for this article (see end of article for details)
contains a CORBA server and a traditional client, as well as an
IdlClient application. Lets take a look at how to use the xxx_i.pas
and xxx_c.pas files. The client application consists of a simple
form thats used to call methods on the server. After running
IDL2PAS on BankServer.Idl, the resulting BankServer_i.pas and
BankServer_c.pas files are used by the main form of the client application (notice that we need nothing from the server once we have
the IDL file). Once weve called CorbaInitialize, all that remains is
to bind to a factory object and request an IAccountManager object
(see Figure 3). Once thats done, we can begin making calls to the
servers IAccountManager interface (see Figure 4).
Comparing the two client programs, we see there isnt much difference in terms of time or complexity. The difference, of course, is
that if the server had not been written in Delphi, the IdlClient
would still work as-is, while the standard client would have
required a good bit of work.
The Dreaded Known Issues Clause
The first version of any tool always includes some known issues.
It would be misleading to say that VisiBroker 3.3 for Delphi is
any different. Two of the most notable issues include the fact that
VisiBroker for Delphi only helps on the client side; it cannot

create server objects based on IDL files. Because of this, the
technique of CORBA callbacks is not supported. According to the
documentation, long-term goals include a more complete mapping of the C++ ORB, as well as support for CORBA callbacks
and server-side code.
Conclusion
VisiBroker 3.3 for Delphi significantly strengthens Delphis

expanding capability to create flexible and global solutions in what
is fast becoming a CORBA world. Its ability to integrate with prepackaged CORBA components in a language-independent environment will allow Delphi to help set the standards in a young
and powerful market.
MAY\DI200005EW.
Eric Whipple is a Delphi trainer and mentor for Pillar Technology Group, Inc.
of Detroit, a full-service consulting, training, and mentoring firm specializing in
project management and in the analysis, design, and development of distributed,
enterprise systems (http://www.knowledgeable.com). Eric is a Delphi 4-certified
developer and trainer, and can be reached at ewhipple@knowledgeable.com or
by phone at (317) 915-9031.
Best Practices
Directions / Commentary
Some Comments on Commenting
o you comment your code? Commenting is one of those topics upon which you can hear conflicting opinions from
various programming experts pseudo, self-proclaimed, or legitimate. Some say you should always comment
your code. Others say, basically, that commenting is for ninnies and newbies; its unnecessary for real programmers,
because you should be able to read the code and determine what its doing.
In the classic book Code Complete by Steve McConnell (Microsoft
Press, 1993), Socrates is quoted as saying, I think that people who
refuse to write comments 1) think their code is clearer than it could
possibly be; 2) think that other programmers are far more interested
in their code than they really are; 3) think other programmers are
smarter than they really are; 4) are lazy; or 5) are afraid someone else
might figure out how their code works.
After maintaining many other peoples code on several different
projects, I agree with Socrates and am firmly entrenched in the
commenting-is-a-good-thing camp. But not just any kind of commenting. Many reluctant commenters end up producing worthless
comments, thus completing a self-fulfilling-prophecy.
Those who object to commenting often say that: 1) they dont have
time for commenting; or 2) all you need to do is read the code to
see what its doing.
Let me respond to these objections in turn. First, in regard to having
no time for commenting (Im too busy doing the real work of
coding): If you dont have time for commenting your code at the
time you write it, when will you have time for it? Using PDL
(programming design language; see Code Complete for details and
examples) to delineate what a section of code should do, and adjusting and/or enhancing it immediately following coding, when the
logic (and the reason for it) is still fresh in your mind, is a coding
best practice. As in analysis and design, this work done up front
will, in the long run, save you time.
As to comments being unnecessary because the code explains itself:
The problem here is a failure to understand the true nature and
purpose of commenting. Its true that low-level comments (at least, if
the code has been written as cleanly and simply as possible) shouldnt
be necessary. What comments should convey is a high-level view of
the code. In laymans terms: what does it do? why? when is it called
and by whom? and how does it fit in with the overall body of code?
Viewed in this light, commenting can be likened to news reporting:
Be sure to convey who, what, why, where, when, and how. Done
right, commenting is an integral part of the coders responsibility (to
the coder, maintenance programmers who will have to work on the
code, and the coders employer or client).
Not all comments are good comments, however. For example, this
type of comment:
// Loop through the doohickeys.
for i := 0 to Doohickeys.Count-1 do
sl.Add(Doohickey[i]);
is superfluous. Its obvious from looking at the code that its looping
through the doohickeys. Any programmer can see that.
Here, on the other hand, is a good example of commenting. The
reader is provided with information thats neither transparent nor
trivial:
function AddLeading0s(ADate: string): string;
begin
// If "m/", make it "mm/".
if (Pos('/',ADate)=2) then begin
ADate := '0'+ADate;
Result := ADate;
end;
// If now "mm/d/", make it "mm/dd/".
if (ADate[5] ='/') then
Result := Copy(ADate,1,3)+'0'+Copy(ADate,4,4);
end;
To make things crystal clear, include in the comments examples of

input and output. For example:
// Example: 991231 is received and converted to 12/31/99.
function ConvertFromYYMMDDToMMDDYY(
ADateToConvert: string): string;
begin
// DateToConvert should be length of 6 (YYMMDD).
Assert(Length(DateToConvert)=6,
SDatePassedOfUnexpectedLength);
Result := Format('%s/%s/%s',[Copy(ADateToConvert, 3, 2),
Copy(ADateToConvert, 5, 2),
Copy(ADateToConvert, 1, 2)];
end;
Adding a history of modifications can also be beneficial. If a block of

code breaks, consulting the revision history can help you determine
what was changed and when. Its also a good idea to sometimes
simply comment out replaced code, rather than deleting it, especially if a simpler version of the code was replaced by a more complicated (although perhaps more efficient) version.
Besides commenting procedures and functions, it is also helpful
to add high-level comments to units (see Figure 1). In this way,
a maintenance programmer (the original programmer, or someone
else) can quickly determine the general purpose and functionality
of the unit. Unit header comments can also follow the reportorial
style of answering (where appropriate) who, what, why, when,
where, and how: Who wrote it (so they know to whom they
should direct questions); what is the purpose of the unit; why
any unusual practices were followed, and/or what factors influenced the decision-making process regarding the architecture and
Best Practices
{ ************************************************************** }
{
}
{ BSaleObj (unit)
}
{ No variable/instance name. Aggregation used in dm (this
}
{ object is encapsulated in that one as a private field).
}
{
}
{ This is the custom report object for the "Billed Sales
}
{ Report by Sales Rep" and "Billed Sales Credit Report by
}
{ Sales Rep" report. These are two quite dissimilar reports, }
{ but happen to be accessed from the same interface.
}
{
}
{ See NBFCW017
}
{
}
{ "BILLED SALES REPORT BY SALES REP" portion:
}
{ "Order Number" is extracted from Arledger.OrderNum and
}
{
ARLedger.VendorCode
}
{
}
{
"Bill-To" is extracted from Customer.Company based on
}
{
ARLedger.CustNum
}
{
}
{
"Margin" is calculated - (CustOrder - InvoicedCost)
}
{
}
{
"Profit %" is calculated }
{
(((CustOrder-Cost) / CustOrder)*100)
}
{
}
{
. . . additional notes left out for brevity
}
{
}
{ Coded by Rupert "Ruprecht" Pupkin III
}
{
}
{ Revision History:
}
{
05/27/1999
Began coding:
}
{
09/28/1999. It has been decided that the "Cost X Factor" }
{
and "Cost" columns are not needed. Rather than code them }
{
out, it will be faster to simply suppress them from
}
{
printing. A new column "Sell Price"*, will take the place }
{
(in the string list) of "Cost".
}
{
* Cost * Quantity.
}
{
}
{ ************************************************************ }
Figure 1: An example of a unit comment.
implementation; what other units use this one; how the various
methods are used, etc.
You can make it easier on yourself to add unit and method comment
headers if you create templates for them. You can do this by following
these steps:
1)
2)
3)
4)
5)
6)
Select Tools | Editor Options.
Select the Code Insight tab.

Select the Add button in the Code Templates section.
Provide a Shortcut Name and a Description (such as Uhdr and
Unit Header).
Select the OK button.
Type in your boilerplate placeholders, for example:
{
Name of unit
Purpose of unit
Anything unusual
Coded by:
Revision history:
Began coding
}
7)
8)
Select the OK button.

Repeat the process, this time creating a template for
method headers.
You can now insert these templates in the code editor by

pressing CJ and selecting the desired template (I guess
the J stands for Just when you thought everything in Delphi
made sense...).
Albert Einstein said that if you cant explain something to
an eight year old, you dont really understand it yourself.
Commenting your code lets you prove that you truly
understand your own code.
Clay Shannon
Clay Shannon is a Delphi developer for eMake Corporation, located in Post
Falls, Idaho. Having visited 49 states (all but Hawaii) and lived in seven, he
and his family have settled in northern Idaho, near beautiful Coeur d Alene
Lake. He has been working (almost) exclusively with Delphi since its release,
and is the author of the book Developers Guide to Delphi Troubleshooting
(Wordware, 1999). You can reach Clay at clayshannon@usa.net.
The Future of Computing: Preparing for Delphi for Linux (cont. from page 36)
writers I enjoy reading as much as Swan, so I recommend this
book highly.
The last book is probably the most unusual. Maximum Linux Security: A Hackers Guide to Protecting Your Linux Server and Workstation
(SAMS, 1999) lists no author. Whats going on here? Is the content
so controversial that the author must disguise his or her identity? Perhaps. Its well known that some of the most successful security experts
have had experience with hacking into seemingly secure networks.
This work is particularly well researched, with a wealth of specific
examples of security setups that failed.
The opening section provides basic information on Linux security,
including the role of the system administrator. The second section
concentrates on the Linux user and deals with password and code
(virus) issues. The third section explains techniques that apply to
networks, while the final and longest section covers the Internet.
Because Linux is particularly popular in running network servers,
this book may be the most useful to Delphi developers looking

forward to Delphi for Linux. Its informative and entertaining; I
recommend it highly.
Many developers have contributed to the growth of Linux as one of the
most highly regarded operating systems. The Internet has been crucial to
this development. Next month well conclude this three-part series with a
look at the Linux Internet sites as we prepare for Delphi for Linux.
Alan C. Moore, Ph.D.
Alan Moore is a Professor of Music at Kentucky State University, specializing in music composition and music theory. He has been developing
education-related applications with the Borland languages for more than
10 years. He has published a number of articles in various technical
journals. Using Delphi, he specializes in writing custom components and
implementing multimedia capabilities in applications, particularly sound
and music. You can reach Alan on the Internet at acmdoc@aol.com.
File | New
Directions / Commentary
The Future of Computing: Preparing for Delphi for Linux
ast month, we explored some of the issues related to the eagerly anticipated Delphi for Linux. This month, well
continue that discussion and take a look at five books that might be helpful in preparing for this momentous
event. First, Id like to share a message I received from a reader regarding my January 2000 column on Delphi
and Visual Basic.
[In your article] I think you left out one of the most important
points for using Delphi: it will be ported to Linux. Microsoft (MS)
will never do this. Therefore as the world evolves away from MSs
buggy software there will be no MS competition. Once Inprise ports
[Delphi to Linux] there will be Corel [WordPerfect] Suite on the
desktop with Paradox as a desktop database and Delphi as your
enterprise solution. Where will MS be? Who knows and who cares.
Some people say MS will never loose the desktop. Once consultants
realize they can pocket 50% of the money that MS charges for their
OS and still leave the client with a surplus it wont take long for the
massive defection.
Interestingly, I received this the day after I submitted my previous
column. Will these predictions come to pass? Who can say? However,
it is clear that Linux will provide increasing competition for Microsoft Windows in the coming years. Windows and Linux do have one
interesting thing in common: Most of the high-end programming
for each is written in C. So in this column, I will discuss both the
general books introducing Linux and books about programming for
Linux using C.
In writing last months column, I found one book particularly
helpful in developing a general understanding of Linux: Caldera
OpenLinux 2.3 Unleashed by David Skoll (SAMS, 1999). If youre
looking for an excellent introduction to Linux, this is worth
examining. The first part covers all the basic information about
Linux, from the historical information I included in my previous
column, to working with the Linux shell, to Linux text editors. It
also includes an introduction to the X Windows system, Linuxs
graphical user interface. Many developers will be interested in
using Linux on a server. For them, the second section dealing with
System Administration will be of particular interest. This work
covers all the basic tasks, including working with file systems,
printing, setting up a TCP/IP network, and configuring a system,
among many others.
However, Linux Programming Unleashed by Kurt Wall, et al.
(SAMS, 1999) is written specifically for developers and deals with
advanced programming topics such as input/output, memory
management, and communication. Not surprisingly, the emphasis
is on C/C++; however, there are sections on Java and other languages. The first section of the book provides directions for set36 May 2000 Delphi Informant Magazine
ting up a complete development environment, including using

the Make tool to build large applications, writing software that
can be automatically configured, establishing a version-control
system, using programming macros (with Emacs), and so on.
Although there is no CD-ROM, there are some code examples
and many useful links to Linux programming sites. (The remaining books include CD-ROMS.)
If you want to get up to speed quickly using C for Linux, SAMS
Teach Yourself C for Linux Programming in 21 Days by Erik de
Castro Lopo, et al. (SAMS, 1999) is ideal. It assumes little or no
knowledge of programming and starts from the basics. In the first
week you learn about Linux and C, including the elements of a
C program. There are separate chapters (days) teaching program
control, syntax elements, and structure, such as expressions and
statements. In the second week you explore more advanced aspects
of the language, including pointers, strings, structures, scope, and
more advanced program control techniques (including loops). The
last week introduces practical examples, such as using libraries,
working with memory and disks, and exploring several advanced
topics. This book is particularly appropriate for the novice. The
next book is better for intermediate-level programmers.
Tom Swan is probably one of todays best known computer programming writers, having written over thirty popular works. Tom Swans
GNU C++ for Linux (QUE, 1999) is an excellent introduction,
providing detailed coverage of the essential topics, from setting up
your system to working with the X Windowing environment. As
with the previous title, this book begins with the basics. However,
it provides more detailed information, accompanied by a plethora
of useful tips.
After an introductory section, Swan provides the basics of C/C++
programming. Section three explores the details of object-oriented
programming. Among the useful topics here are techniques for
handling exceptions. Sections four and five introduce advanced
C++ techniques and C++ class libraries, respectively. Section six
introduces the X Windowing environment. Section seven includes
five appendices that provide additional information. There are few
The Future of Computing: Preparing for Delphi for Linux
continued on page 35

May 2000, Volume 6, Number 5: On The Cover

Uploaded by

Copyright:

Available Formats

May 2000, Volume 6, Number 5: On The Cover

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

May 2000, Volume 6, Number 5: On The Cover

Uploaded by

Copyright:

Available Formats

May 2000, Volume 6, Number 5

Cover Art By: Arthur Dugoni

Parsing the Web Richard Phillips

Delphi in the Office Ron Loewy

The API CAlls

Active Directories Simon Murrell

FraserSoft Announces GenHelp

companys component Help

import Pascal and C++ header

efficiency for end users.

reporting tool (works with Crystal

SkyLine Tools Imaging Releases Doc-to-Net 6.0

tools offered for documents are

The Imaging Source Ships TX Text Control 7

processing software, in reusable

trol is a 32-bit programming

Blinkinc Announces DeltaPatch 1.2

Blinkinc announced DeltaPatch 1.2, a multi-platform

more frequently and in less time.

or directories to their new versions and create a patch file

SkyLine Tools Imaging

image, vertical and horizontal

and vertical credit messages can

SkyLine Tools Announces ImageLib Combo@TheEdge 5.0

3 May 2000 Delphi Informant Magazine

SkyLine Tools Imaging

New Wave Software Offers SPI 2.5

tion. SPI-protected products can

and built-in e-commerce. The

Quest Announces Schema Manager 3.0 and Data Manager 3.0

tions, and support for SQL

Quest Software, Inc.

Delphi 5 Update Pack Available

5 May 2000 Delphi Informant Magazine

Informant Communications Group Launches ComputerBookstore.com

cent off the suggested retail

Corel and Inprise Merger to Create Linux Powerhouse

businesses will have a presence in

operating system grew out of its

HTML / XML / SGML / Parsing / Internet

Parsing the Web

SGML directives, and

Elements are the facet of the SGML content with

where TITLE is considered to be the name of the

In HTML, we immediately note that there are

The <! is a markup declaration open delimiter, and indicates that an

indicates a comment; --> terminates a comment.

<!ELEMENT P - O (%inline;)* >

The remaining SGML directives are denoted by the beginning

As mentioned before, XML is also derived from SGML. While

Any items encountered in the content that are not contained in

Figure 1: Sample HTML.

Any content located outside of start/end tags, comments, or

Consider the sample HTML shown in Figure 1. The parser would

Figure 3: TTagNode node types.

Figure 4: Using GetTags.

The GetTags method of a TTagNode will return a list of all children

Figure 2: Hierarchical representation of

TTagNodeList has two methods that offer another approach