<fb6fbf560712191625y202eea74g9cb9a44aef43da51@mail.gmail.com>
Current votes: None.
Shannon asked a series of questions, and ended with > Anything short of a formal document on the WHATWG > can't ... Ian, I think it would be worthwhile to formally make an explicit statement that the standard is intended to be primarily *descriptive* rather than *prescriptive*. I would suggest the Introduction of the document itself, or possibly even the abstract. Shannon, Your questions were great -- for a prescriptive standard. The problem is that the w3c (let alone the whatwg) doesn't actually have any enforcement power whatsoever. We can all agree that people (and BigCorp) shouldn't do Bad Things. Sometimes they do anyhow, and a standard can't prevent this. (Even laws can't actually prevent it, though they make it less common.) The biggest goal of HTML5 (and CSS2.1, for that matter) is simply to document what actually happens, ugly or not. The new features are (generally) being added only after agreement that the browsers will all add them -- and after some work to ensure that they will do so in a compatible manner. A descriptive standard won't clean up the existing web, but it at least reduces the harm. Browsers still won't be able to take out workarounds that were originally written to account for bugs in netscape 2, but they will be able to stop writing workarounds for the incompatible workarounds that someone else wrote for the same bug. Perhaps more importantly, page authors should be able to rely on the spec. In the past, some features have been vaporware, which caused people to lose faith in the spec and start relying on (often broken) examples, which meant that the browsers were under even more pressure to support the workarounds. Tying this back to video codecs, it would be great if we could tell authors "provide at least format X, if the browser can't support that, it probably doesn't do video at all." It would be bad if we told them that and it wasn't true -- regardless of why it wasn't true. It would also be bad (though less so) if format X were a poor technical choice. Pick a random web page with images, and look at the alt attributes -- that is the best we could hope for if format X is viewed only as a worst-case fallback. -jJ