Anne van Kesteren

Standards for standards

As requested, I read An Open Letter to WaSP. The proposal is something like this: “The one way of following standards, standardized by the WaSP.” Because there is a lot of debate going on about non-trivial issues some people are scared the main message from the web standards community is not visible for those who not know. (Or at least, something close to that.) For example, I use “web standards” and omit some optional tags. Some people are compelled to say something about that anytime they can and whenever I give them the possibility to do so. Leaving out optional tags, however, is not what it is about and therefore my weblog is not always useful for newcomers. (Actually, it never is, but consider some other weblog, like that from Faruk Ateş. I am glad by the way that he joined the no-www crowd. Eventually he will use pure ― optional tags omitted ― HTML as well, it just takes some time. Oops, I did it again.)

Anyway, standards for standards. The Web Standards Project starts some activity with smart people figuring out which set of standards to promote. Knowing the WaSP, this will be XHTML 1.0 (despite all its issues), CSS 2.1 and perhaps the UTF-8 serialization of Unicode. Perhaps some more standards, who knows. Perhaps even HTML 5 if they take long enough. However, that might not make sense. The problem is that specifications are hard to read. The problem is also that specifications are full of bugs. Another problem is that the XHTML and all are not really standards, but I think we can safely forget about that one.

What I think that would be a good idea, at least for HTML, is for some organization to make an overview of all elements and attributes. Whereby with all I mean all correct and appropriate.) Describe when and how they should be used and then explain how you can use these to create documents. Either HTML or XHTML. In the end it will be parsed as HTML anyway due to the text/html MIME type so that is not really relevant (although people should be aware of the fact). Something similar could be done for CSS where it is even a bit easier as it has only one serialization (well, only one that usable). If this is not what is meant with “implementation approach” please update the relevant posts or post something new. I shall close the comments on this one to prevent the aforementioned flame wars in my general direction.