Using the HTML5 doctype prematurely "considered harmful"
By the way, the "considered harmful" in the title is intentional even though it has been abused as a populistic phrase throughout the years: it seems fitting since no one less than Eric Meyer once wrote an essay on the subject of why "considered harmful" can be considered harmful by itself
Some people did notice Chris Wilson's remark (I for one did too) and took it for what it probably is: an afterthought that sofar cannot be considered a fact. Other people however are making a lot more of it and are actually making all kinds of assumptions or are even suggesting we should use the HTML5 doctype right now.
This may well undermine the future success of HTML5 itself. The reasons are two-fold:
First of all we should not assume that using the HTML5 doctype will trigger 'edge'-mode in IE8+. Actually, we should not assume anything about IE's future rendermode switching until either MS releases a specification that fully defines the behaviour (not likely) or until we can actually deduce it's exact behaviour by reverse-engineering IE8 when a (beta)version becomes available. Also the matter of a meta-directive actually overriding the same http-directive is something that in my opinion is a mistake and has been communicated falsely (although MS does have a history of ignoring correct http directives and doing something completely different based on some form of content-sniffing).
In fact, if we start using the HTML5 doctype right now Microsoft might be forced to lock that into IE8 rendermode after all when IE9 comes along that may or may not "break the IE-centered web" (it would be really helpfull though if MS would provide solid examples to their claims that version lock-in is really inevitable).
Secondly it just doesn't sound like a good idea to me to start tagging documents as 'HTML5' when the specification is just a first working draft with a lot of unresolved issues and which is very likely to change in some fundamental areas. Even though documents may be conforming to the present state of the specification (e.g. according to the highly experimental HTML5 validator) they may well be unconforming by the time HTML5 becomes a recommendation, or the HTML WG would be forced to take into consideration bad use of HTML5 besides the countless precedents set by bad use of HTML4 they already need to consider now. My worst nightmare would actually be when automated tools start inserting HTML5 DTD's far ahead of time...
Using the HTML5 DTD right now will in the end only re-inforce Microsoft's lock-in into the web, giving them reasons to totally ignore HTML5 by freezing these documents into some rendermode that in fact doesn't have HTML5 support at all. And maybe in the end that is exactly what Microsoft is aiming for. Maybe that would explain also the lack of Microsoft's involvement in the HTML5 WG (their promised review of the current draft is already 6 months overdue); maybe they just want to freeze HTML for once and for all in order to be able to shove some proprietary alternative forward... I would really not like to be right about that...
I personally think we should just ignore the IE8 render engine until microsoft is bold enough to make it the default without any lock in parameters
Whether or not future versions of HTML will be backwards compatible, is not clear yet. Forwards compatibility is the goal, but it may turn out to be quite impossible. In that case, we will get versioning thourgh doctypes, like <DOCTYPE html 6>.
Forwards compatibility would make life easy: one render engine would be sufficient to render all (standards compliant) web pages! That's why people try so hard to make HTML forwards compatible.
The big issue in here seems to be the impredictable behaviour of MSIE in the future and its refusal of being compatible with the standards. Damn that browser. Damn that company.
Comments are closed