 Database Journal: The Knowledge Center for Database Professionals
|
Time Out From Behaviorsby Nate Zelnick Let's Not Go Screwing Things Up, Now...
The DOM defines the standard interfaces between the document and the application and provides interfaces that allow that application to manipulate the document without having to know anything about it ahead of time. Defining this right the first time will save lots of headaches later. A good example of how not to do this is in HTML. The lack of a commonly supported DOM in Navigator and Internet Explorer is one of the reasons why it is so hard to write fully-functional HTML pages with fancy scripting tricks that work everywhere. Clearly, getting the XML DOM right before there's a ton of applications consuming XML is more important than getting a specification out the door fast. Otherwise, developers are going to have to spend a lot of time redoing their work. But that also means that lots of people have lots of opinions about how this should work. Reaching consensus--the essence of the standards process--takes time. Another controversial area is namespaces. One of the problems with XML derives from its strength: The ability to assign named structures to parts of a document. But since names can be assigned arbitrarily, how do you avoid collision between two tags that have the same name? The solution to this is to define a namespace for each element so they remain distinct. A draft specification for namespaces was released in April and seemed uncontroversial. But in the last few weeks it became clear that just about every bit of XML-based work depended on namespaces, so lots of people suddenly became interested in it very quickly. The latest draft--released in early August--is a work in progress that has been expedited. A finished proposal that will go up for vote is expected by the end of the month. Because so many other proposals depend on namespaces, it is absolutely critical that this be nailed the first time. Rewriting every subsequent specification later because an unforeseen problem with namespaces popped up is a nightmare nobody wants to contemplate. On top of these two core standards are a whole bunch of things that are needed before XML is ready for prime time. For instance, how XML is rendered for display is still largely unknown. There's a Note (W3C-speak for an interesting submission that hasn't moved to a standardization process) that describes how it might be done with CSS, but no standard. There's also a working group that's looking at XSL, but that will depend on namespaces. Then there's a whole hyperlinking infrastructure for XML documents called XLL (eXtensible Linking Language), which is also chock-full of dependencies and controversy. And then there's figuring out how to mix XML and HTML together in documents (called "Data Islands"). And on and on. When the Web first emerged from its academic roots, it seemed like anybody who had a good idea could just throw it out into the network and get people to adopt it. As the Web has matured and become an important vehicle for business and communications, the stakes are way too high to follow an ad hoc development style. While it's frustrating to watch the standards process bog down in complexity, it's also far better to wait and get it right than rush and fix it later. Just wait until the next version of HTML that is expected to be fully XML compliant. This change will mean that every IMG tag on the Web will be illegal. Sound like a fun transition?
Legal Notices, Licensing, Reprints, Permissions, Privacy Policy.
Whitepapers and eBooks
Webcasts
Contact the WebDeveloper.com® staff Last modified: Friday, 22-Aug-2008 13:46:48 EDT
|
Refresh Daily
|
and find truth, justice, and a clue or two.
