- From: Henrik Dvergsdal <henrik.dvergsdal@hibo.no>
- Date: Tue, 17 Apr 2007 23:56:51 +0200
- To: public-html@w3.org
On Apr 17, 2007, at 19:48, Maciej Stachowiak wrote: > I don't know what you mean by formal in this case. If you are assuming > that a language has to define its grammar in a special symbolic syntax > rather than English to be well-defined, then I think you are just > wrong. OK. I respect your opinion on this. I'm just not ready to agree with you. Anyway that wasn't what I was trying to say. > Also, the spec changing for some time in response to implementation > feedback is neither unusual nor a bad thing for a web standard. I wasn't trying to say that either. I realize the post was probably badly phrased. Let me contribute a few loosely related reflections instead: Change in response to feedback is usual as you say. However, explicitly making a standard rely on external implementations in the what HTML5 does is unusual, at least in my experience. In practice both its extension and its intension will depend on a multiplicity of third party implementations. We are used to the former, but not the latter. This is what loosens its definition (a little), not the english prose as such. I don't think this is necessarily a bad thing though - it just challenges my conception of a standard, Seen over a large time span, HTML5 will assume an almost organic nature. It is a documentation framework as much as a specification, reflecting the reality of the web as much as making it. Hence HTML5 is not just a specification, it is more than that. I note that HTML5 represents a departure from HTML4 in more ways than just the addded language features and I suspect these changes may turn out to be at least as significant. We are adopting an entirely new way of dealing with this matter and the rules of the whole web-game are being changed. There will be new positions to take. The editor, for instance, will have quite a lot of power in this setup. With HTML 4.01 we could teach our students a static language and then explain the deviance in browsers. With HTML5 we must relate to a more complex and dynamic language, but we'll have less deviance to explain (hopefully). I think we will have to remake our teaching methods. The dynamic nature HTML5 is not likely to be good news for vendors with long development cycles and/or large customer bases. I think they will be forced to lag behind the smaller actors as the standard evolves. One thing I am concerned about is that we may be making things unnecessarily complicated. As I've noted earlier: apart from a few issues, HTML is currently quite interoperable. In practice most problems are related to CSS and the DOM. Sometimes I think we should just fix those issues in HTML 4.02 and live happily a few years more. Then we could add some new features and create HTML 5.00 etc. etc. But this would be in violation with our charter of course. -- Henrik
Received on Tuesday, 17 April 2007 21:57:24 UTC