- From: John Foliot - WATS.ca <foliot@wats.ca>
- Date: Tue, 24 Feb 2009 17:23:34 -0800
- To: "'Jonas Sicking'" <jonas@sicking.cc>
- Cc: "'Rob Sayre'" <rsayre@mozilla.com>, "'HTML WG'" <public-html@w3.org>, "'W3C WAI-XTECH'" <wai-xtech@w3.org>
Jonas Sicking wrote: > > Isn't the question at hand here: would saying MUST rather than SHOULD > result in more sites being accessible? > It is my contention that mandating some form of fallback sets the stage to ensure that something is provided by the content authors. Whether or not that *something* will fit the bill every time will always be a coin toss*, but as it stands right now, we're getting zero *somethings*, so even if half the time it is wrong, the chances have increased by 100% that possibly it will be right at least half the time - and certainly way better than what we have as examples today. (*this is where education and best practices documentation is critical. However, in this area, we've come a long way with trail-blazers such as A List Apart, WaSP, and Opera's Web Curriculum to name but 3 highly known and respected avenues) Since <canvas> is also relatively new, we have the added bonus of teaching people how to use it correctly the first time 'round - part of the problem with HTML 4 and earlier is that there were no real learning resources around, most folks learned by copy and paste, and/or adopting bad practices without understanding that they were bad, or the consequences thereof (I further assert this is why @longdesc gets such short shift today). But now, a whole new class of web developer has emerged, ones who have been trained and understand the importance of standards and standards implementation. So I also want to tap into that vein, and mandating something to make <canvas> work means that these standards aware developers will learn what they must learn to implement <canvas> properly - and that includes ensuring it is accessible. Now, If anyone out there has a better idea, I am quite open to hearing and discussing it - HTML5 is supposed to be a collaborative and consensus based spec after all. But to simply leave the spec as is, 'suggesting' that fallback content be provided, is tantamount to admitting that it will rarely if ever be provided - what exactly is the incentive for doing so? So without turning up the heat, there is no need to worry about the requirement. JF
Received on Wednesday, 25 February 2009 01:24:14 UTC