- From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
- Date: Wed, 19 Sep 2012 12:00:20 +0200
- To: Silvia Pfeiffer <silviapfeiffer1@gmail.com>
- Cc: Steve Faulkner <faulkner.steve@gmail.com>, Maciej Stachowiak <mjs@apple.com>, John Foliot <john@foliot.ca>, Sam Ruby <rubys@intertwingly.net>, public-html-a11y@w3.org
I did not have time too look through it, but those I looked at either contained only a "#" or they contained (another) image file. With regard to the first (#) then I agree "misinformed" about the potential negative effect. With regard to image URLs inside @longdesc, then there are image light box solutions - libraries - that more or less consciously makes incorrect use of longdesc. (Today they would perhaps picked at @data-foo attribute instead - but that was not 'valid' then.) Of the few I scanned, no one contained text. However, there was an open source photo album CMS that, in an legacy version, inserted text into longdesc. That is to say: Some of the documented misuse seems to stem from CMSes and libraries that got it wrong. And I imaging that developers of CMSes and lightbox libraries would not have done such things if the negative effect could have been perceived by a browser that they themselves could use. = It is very important with implementation. Leif H S Silvia Pfeiffer, Wed, 19 Sep 2012 19:33:23 +1000: > Having had a look at this data makes me wonder ... since none of the > sites that you've crawled had used longdesc correctly, either your set > of sites is dubious or we are really staring at a totally mis-informed > Web developer community. > > Silvia. > > On Wed, Sep 19, 2012 at 7:53 AM, Steve Faulkner > <faulkner.steve@gmail.com> wrote: >> Hi all, >> >> As part of a survey of the top 10,000 web sites home pages carried out >> back in April [1] I grepped the instances of longdesc [2] >> this is what I found: >> >> 1938 matches in 86 files. >> >> You can review the data and draw your own conclusions. >> >> >> [1] >> http://www.paciellogroup.com/blog/2012/04/html-data-for-the-masses-data-dump/ >> [2] http://www.html5accessibility.com/HTML5data/dump/longdesc.html >> >> >> regards >> SteveF >> >> On 18 September 2012 22:34, Maciej Stachowiak <mjs@apple.com> wrote: >>> >>> On Sep 18, 2012, at 2:16 PM, John Foliot <john@foliot.ca> wrote: >>> >>>> Maciej Stachowiak wrote: >>>>> >>>>> In addition to issues with these specific suggestions, keep in mind >>>>> that a previously raised concern with longdesc is that the corpus of >>>>> available longdesc content in the wild appears to have very high level >>>>> of bad content. >>>> >>>> I encourage you or others to provide specific proof of that assertion. >>>> >>>> On one hand, we have professional content producers that are creating >>>> @longdesc content today (Pearson Publishing and the Government of >>>> Canada to >>>> name 2), who, if nothing else, are probably quite good at document >>>> management practices. >>>> (http://lists.w3.org/Archives/Public/public-html-a11y/2012Sep/0210.html) >>>> >>>> On the other hand, we have a 5-year old blog post from Mark Pilgrim >>>> (http://blog.whatwg.org/the-longdesc-lottery) that alludes to statistics >>>> that Ian Hickson accrued, but was unwilling to publicly share. >>>> >>>> Do you have any other "proof" of this assertion? Have you or anyone else >>>> "surveyed" the corpus recently to see if there have been any >>>> changes to this >>>> assertion over the past 5 years? (Note: I have not, but given that >>>> serious >>>> content publishers are now using this attribute routinely in their >>>> work, I >>>> can only surmise it has improved significantly - but feel free to dispute >>>> that claim with proof to the contrary.) >>> >>> I'm just mentioning the point of concern that was raised. I am not >>> interested in debating its validity. >>> >>> I will note that, if you want to persuade browser vendors to >>> implement something, claiming that the evidence provided is not >>> "proof" is unlikely to be a very compelling argument. Providing >>> actual evidence to the contrary may be more compelling. >>> >>> The rest of your message seems to be more about whether longdesc >>> should be "retained" in the spec in the sense of a conformance >>> requirement that has no engineering impact. I don't have any >>> substantive comments on that question. But I do note that it seems >>> to be back in the mode of "mandate something that browsers won't >>> implement". >>> >>> Regards, >>> Maciej >>> >>> >> >
Received on Wednesday, 19 September 2012 10:00:57 UTC