Push Button Paradise

Dear Web Hypertext Application Technology Working Group,

You might be a little surprised to hear this from me, but I think that there exists a good opportunity, in fact a vital need, for a coordinated group to work on community standards outside of the purview of the W3C. Here, I offer some comparisons and offer some concrete suggestions.

I've been involved in EXSLT to some degree. Here we have a loosely-coordinated, widely-respected group that has set an excellent example of how to go about such an endeavor. I don't know of any overlap between the XSL Working Group and EXSLT membership during the early days, but nevertheless the EXSLT work provided a great compliment to XSLT 1.0 and XPath 1.0, became widely implemented, and provided valuable input for later versions of XPath, XSLT, and even XQuery.

The key to their success has been their ability to unify rather than splinter the community. Hard-core XSLT fans, even Working Group members, still appreciate EXSLT and use it as a tool that makes XSLT even better. People less thrilled with XSLT too have found that it makes many kinds of transformation tasks more bearable. Of course, developers that hate XSLT still hate XSLT+EXSLT, but not everyone can be pleased. So as a fuzzy, non-concrete suggestion, try to be more like EXSLT. But specifically,

Concrete suggestion: Make uniting the community a specific written objective of the overall WHATWG work.

Even in naming, the E in EXSLT stands for Extensions, so their positioning has always been clear: 'we are something that complements XSLT'. Imagine the difference if their effort had been named, say, "XSLT2". Naming matters.

Concrete Suggestion: Don't use the confusing name "HTML5", unless you get W3C approval first. Pick a name that highlights a complementary role.

Actually, for now it seems like the W3C will not do anything that will be called HTML 5.0, having quite thoroughly switched over to XHTML for continued development (still, I stand by the above concrete suggestion). I understand that corporate WHATWG members are quite sore with this decision, not to mention the continued proliferation of namespaces, which is indeed an ugly thing. The namespace problem in particular just keeps getting more and more convoluted, and is heading for a steady-state where every element has a long, unique, and nonrememberable URI-string attached. But stop and think about what's going on here. Take Web Forms 2.0, for example. It is being pulled in many directions:

1. It's a manifesto against namespaces (yay)

2. It's a manifesto for restarting development of HTML 4

3. It's a manifesto for strict backwards-compatibility

4. It's a demonstration of willingness to work outside of W3C

That's a lot of baggage for one little spec to carry. It's not surprising that it's missing other important things, like goals of community unification (above) or transitional compatibility (below).

Concrete Suggestion: The namespace propaganda is critically important, but out of place in Web Forms 2.0. Drop it. Find another channel for that. I'll help.

Concrete Suggestion: I feel your pain about HTML 4 development, but that should be a bonus feature, not a core requirement--it's doing too much damage to your forwards-compatibility story. Read on for more about that.

XSLT 2.0 is pretty far along, and the majority of things that sparked the need for EXSLT in the first place are now a core part of the language. Folks who learned EXSLT, say how to work with the node-set extension, are in better shape to take advantage of new goodies in XSLT 2.0. When plowing the fields of EXSLT, the managers weren't looking behind, they were looking ahead. In Web Forms 2.0, the diagram showing XForms getting translated into an intermediate format is a great start, but it still has the feeling of an afterthought.

I noticed that Hixie's presentation at http://www.hixie.ch/advocacy/whatwg-presentation/ doesn't seem to be working very well in either Firefox 1.0 of Tiger Safari, or at all in Opera beta 8. I'm not saying this as criticism, as his presentation was great at XTech. It's just that when I look at things like FormFaces, and see how well these guys do at making XForms markup work in current browsers, including Firefox, Safari, and Opera, I feel like I'm not getting the whole point of making a whole separate language that's "backward compatible", but apparently still quite hard to get working in current browsers.

I was among the first people to download and attempt to build the freshly open sourced Netscape browser. I was there, struggling to figure out the legacy code. When they finally announced that a better foundation was needed, I understood. But again, the transition story was weak, and it caused noticeable problems.

I'll say that XForms doesn't have a great transition story at this point--more work is needed. It's probably ironic, but a group outside the W3C is probably in the best position now to build this bridge.

Concrete Suggestion: Make forwards-comp ability to XHTML 2.0, including XForms, an explicit written goal of WHATWG work, specifically Web Forms 2.0.

I've noticed lots of neat things happening in the world of SVG lately. Opera supports it on mobile devices. My nightly Firefox build supports it natively, and even Safari has a page indicating desire for SVG. I find all this fantastic, even if SVG is utterly non-backwards-compatible with existing web content.

Concrete Suggestion: Make forwards-compatibility to SVG an explicit written goal of WHATWG work, specifically Web Apps 2.0.

One final suggestion. I believe that during the critical formative phase, the EXSLT group consisted entirely of people affiliated with non-W3C member companies, though they still kept up communication. WHATWG, on the other hand, is principally made up of W3C member companies, which causes poor perception and sundry whispered questions about motives, moral authority, etc. I think the questions are unfounded, but understandable.

Concrete Suggestion: Individuals in WHATWG that work for W3C member companies need to join the appropriate Working Groups.

If you follow these suggestions, you might find some changes are needed to your specifications. But you'll also find something remarkable as one-by-one critics change into supporters.

I wish you continued success in your endeavors. Keep putting out great products. If you have more questions about this letter, feel free to ask anytime.

Oh yeah, and I'm still available for consulting. :)

Sincerely,

Micah Dubinko

posted at: 22:56 | under: 2005-06 | 4 comment(s)