- From: Mike Brown <mike@skew.org>
- Date: Sun, 17 Jun 2007 22:18:58 -0600 (MDT)
- To: ian@hixie.ch
- CC: public-html@w3.org
I'm surprised that after all the work put into other aspects of HTML 5, someone who claims to appreciate clear, concise, formal technical specifications would ask why the draft's cavalier attitude of "sometimes when we say X we mean Y, except when we mean Z" matters. As an implementer, application developer, data architect, and writer, it has been my experience that carelessness with terminology in a spec such as this often leads to varying interpretations & expectations and disparate implementations. I've pored over dozens of specs in excruciating detail [1] to implement and/or document them, and have encountered numerous instances of "close enough" terminology having irritating, progress-impeding consequences. Little details someone thought inconsequential if left open to interpretation in 1993 became entrenched, irreversible idiosynchrasies of the technology by 1998. XPath, XSLT, and DOM edge-case references I thought were mere nuance in 2000 became critical to how I went about coding an implementation in 2004. Sure, maybe the examples I gave for the HTML 5 draft will indeed prove to be inconsequential, but I thought you'd appreciate the attention to detail, even if you don't agree about the potential for confusion. I've found that even the hobby of curating a Wikipedia article like the one on HTML (the history section of which I've just expanded again) has been arduous because of ambiguities in the status of the specs -- the "HTML is dead, long live XHTML: the latest HTML version" evangelists had to be beaten down with copious citations pointing out that HTML 3.2, 4.0, and 4.01 are all still in force and retaining much traction. I'm definitely not looking forward to the fight over the possible interpretations and ramifications of the HTML 5 draft's intent to "replace" / be "the latest version of" HTML 4.x, XHTML 1.x, and DOM Level 2 HTML, all of which have a far greater scope than the HTML 5 draft, not to mention can't really be superseded in quite the sweeping way that's implied. Besides, the comments from a "fresh pair of eyes" looking at your spec for the first time should be welcomed. Concerns, even trivial ones, should be addressed in the spirit of "the fact that someone barely got into reading the spec and got hung up on that issue indicates the topic may need to be better explained, even if the suggested fix wasn't much of an improvement." As more people learn about, review, and send their feedback about the draft to the working group, you are going to have to deal with this kind of thing more and more often. Mike [1] ...though not ECMAScript, hence my assumption that "[[Get]]" was an accidental artifact of editing.
Received on Monday, 18 June 2007 04:19:08 UTC