Proliferation of Custom DTDs

[Mhtml]

I am currently involved in the development of a large, I guess the term Web 2.0 is accurate despite the stigma of Zeldman, project.

It has become necessary, well not exactly necessary but desirable from the developers perspective (mine), to invoke the use of a customized DTD in order to lessen the overhead of the Javascript involved and in order to keep my pages semantically (I am really starting to hate this word and it's underivatives) correct through the use of custom element attributes (basically I'm lazy).

Now, it is of my opinion that the rise in these sorts of applications (this particular project is designed to mimic a traditional network distributed, binary, application) that we will see custom DTDs popping up all over the place and our push for web standards will not extend entirely into this domain and as such we could see a decline in the adoption of complete standards compliance! Also on note of Zeldman's mention of "tag soup" I disagree entirely, in fact (take this you "puritans") I propose that CSS and Xhtml techniques documented by (your beloved) author's like him contribue greatly to "tag soup" (ala; <div class="clear">...).

For those of you who've not read the article to which the references pertain it is available here: http://www.alistapart.com/articles/web3point0

To summarize: What are your oppinions on this behavior, and [sic] it's expansionist (context semantic comparison) possibility?

[drhowarddrfine]

we will see custom DTDs popping up all over the placeWe already do. If pressed, I can link you to a page that lists thousands of custom DTDs for every market place you can think of.
our push for web standards will not extend entirely into this domainNot true. Because the creation of those custom DTDs is based on a standard for creating them.

I read that article some time ago but don't recall what it was about but I'll read it again later.

[Mhtml]

We already do. If pressed, I can link you to a page that lists thousands of custom DTDs for every market place you can think of.Well, within the domain of Web 2.0 is sort of what I was refering to but sure, if you can provide a resource I'd definitely like to read it! This thread is sort of just about raising awareness more than anything. I like to throw some decent discussion into the mix every now and again, beyond that of mere code problems.

Not true. Because the creation of those custom DTDs is based on a standard for creating them.I don't think I clarified myself enough contextually, I was refering to not just custom DTDs but also the standards of accessability.

[felgall]

The X in XML stands for eXtensible. The whole point of the standard is to provide a standard way to extend the Document Type Definition to include whatever is relevant to any particular namespace.

[Mhtml]

I don't think I clarified myself enough contextually, I was refering to not just custom DTDs but also the standards of accessability. I know that, I covered that in my original post re: validation.

I'm not sure people are getting what I'm trying to say: Given disregard to some standards (official or ethical) in this new generation are we likely to see that applications are pushing the boundaries of these specifications and bleeding into new areas in order to create increasingly dynamic functionality?

For example, this project: http://qooxdoo.sf.net ... My project borrows some elements (design wise, not code) from that project. The shift from flat markup to dynamic scripting DOM.

I'm trying to talk about the bigger picture here, not focus on one particular instance as of those examples I cited in my original post.

[drhowarddrfine]

Disregarding standards now is allowed because browsers are required to make a best effort at rendering bad or improper code. In xhtml/xml, there is no such thing. You either follow the rules or fail. If you forget an end tag, the whole page fails. So not following the standard just won't work.

[jkd]

I say go with XHTML and toss in all the custom attributes you want. Gecko, Safari, and Opera can all handle real XML, and so that isn't a problem at all. Just use XML namespaces. As for IE, it certainly doesn't know how to handle XHTML, but it does consistently accept custom attributes just fine (custom tags are another thing... if you want to use those there is a whole set of hoops you need to jump through for IE).

As for "legacy browsers", they parse out custom attributes very consistently, so no sweat really. It's just important to keep everything well-formed.

But yeah, embrace XHTML and extend it if needed. Compare what you're doing to WHATWG's HTML5 and see if you can use anything from that specification so at least you aren't reinventing the wheel. But don't ever let this "must-be-valid" nonsense get in your way - you can break the DTD very safely (utilizing custom attributes namely - which may be why there was the proliferation of marginleft/margintop attributes on the body tag in the 4th generation), as long as you keep everything well-formed in markup.