Welcome Guest. | Log In| Register | Membership Benefits

News In Review

June 28, 1999

Print this story
Print this story
XML Makes Object Models More Useful

continued...page 2 of 5

UML's Drawbacks
An even bigger limitation of the UML approach is the UML model itself. UML does not define a standard file format, meaning that each UML tool vendor stores the representation of its UML model in a proprietary format. While the vendors typically give you a way to get at their proprietary format, doing so requires a learning curve, and such a solution is not transportable to a model constructed with a different vendor's tool.

The result is that the UML model is generally limited to what the vendor provides out of the box, which is usually some form of code generation. Code will be generated once, perhaps being used to jump-start the development.

At this point, having extracted all possible value from the tool, the model is rarely kept up to date as development progresses. At best, it will be pulled off the shelf once in a while to show to a new team member with the caveat that "a lot has changed since then."

Enter XML. Those of you who have experience with or a feel for XML may be scratching your heads at this point. After all, what in the world does XML have to do with systems architecture? Traditional uses of XML have centered on XML being the next generation of HTML, and indeed this is where XML has made its initial mark.

HTML has done a remarkable job of taking the Web from infancy to adolescence, but as the Web matures, HTML's deficiencies are becoming more glaring. HTML does a good job presenting data but a rather poor job describing that data. A document marked up using HTML can present a list of book titles and authors in myriad styles, font sizes, and colors; but no matter how pretty it looks, there's nothing in that document that can be used to tell if a book is authored by Ben Franklin or about Ben Franklin.

XML, on the other hand, marks up a document with user-defined tags that describe the document's content and leaves presentation to be taken care of separately.

Consider the following HTML fragment:

Ben Franklin-Shocking Facts about Electricity

All this markup tells me is that when rendering this information in a Web browser, display the first part in bold and the second part in italics.

An XML representation of the same information would be marked up in such a way that anyone receiving that XML document would know that a book containing title and author components is being described, as in: <BOOK> <AUTHOR>Ben Franklin</AUTHOR> <TITLE>Shocking Facts about Electricity</TITLE> </BOOK> The way this information is presented is up to the recipient. Separating data from presentation in this manner is highly useful, as will be shown a bit later.

XML is also making its mark in the world of data standards. Pre-XML standards are being updated to take advantage of XML--witness the efforts being made around bringing electronic data interchange and XML together. In terms of new standards, one would be hard-pressed to find an industry data standard under development that's not based on XML. A case in point is Microsoft's recently announced BizTalk initiative. BizTalk is a new cross-platform electronic-commerce framework that makes it easy for businesses to integrate applications and conduct business over the Internet with trading partners and cus- tomers. The BizTalk framework is based on new XML schemas and industry standards that enable integration across industries and between business systems, regardless of platform, operating system, or underlying technology.

continued...page 3, 4, 5
return to page 1



Back to This Week's Issue

Send Us Your Feedback

Top of the Page