This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Created attachment 1364 [details] Demoing the proposed changes to the spec text. PROPOSAL: 1) In the current title of the specification: (“HTML5 Image Description Extension”) Add the word “independent”, like so: ”HTML5 Independent Image Descriptions Extension” 2) In the body of the specification: Delete the 5 occurences of the word ”long” 3) In the body of the specification: Add the word “independent” in front of each occurence of the word “description” To see the suggested changes in practise, I have edited the current editors draft and added in the proposed changes with the help of <ins>, <del> and <mark>.. BACKGROUND: In his first comment in Chromium bug-224285, Vlad Alexander indicates that the wording “long description” instills in users/authors that the difference between @alt and @longdesc is one of “short description” versus “long description”. On the background that alt text really is ‘alternative text’/‘text alternative’, he suggests that the *real* dichotomy should be seen as “replacement text” versus simply “description”. PROBLEM: I strongly agree with Vlad in how he contrasts @longdesc to @alt. However, to simply replace “long description” with simply “description”, quote: [1] ]] easily sends the, IMHO confusing message, that unless one opens the longdesc URL, one hasn’t received a description of the image. My proposal is ”independent description”: “Open an independent description of this image”. Justification: 1. “independent description” contrasts with “replacement text and thus sends the message that “if the contextually adapted replacement text/@alt text wasn’t enough, then please check the independent description, where the current text flow hasn’t dictated the shape of the content“. 2. “independent description” vaguely hints that the menu would open (and @longdesc would contain) a URL (since URLs as independent resources is an association that I believe many share). Thus, “independent description” would instill in both users and authors adequate expectations about what @longdesc should contain and about what happens if one opens it. [[ [1] https://code.google.com/p/chromium/issues/detail?id=224285#c10
The proposed edits can now be seen at the following URL: http://lists.w3.org/Archives/Public/www-archive/2013May/att-0046/longdesc-independent.html
(In reply to comment #0) > PROPOSAL: > > 1) In the current title of the specification: > (“HTML5 Image Description Extension”) > Add the word “independent”, like so: > ”HTML5 Independent Image Descriptions Extension” > > 2) In the body of the specification: > Delete the 5 occurences of the word ”long” > > 3) In the body of the specification: > Add the word “independent” in front of each occurence > of the word “description” One more detail to the proposal: 4) In the Abstract section: Replace the word "extended", “to link extended descriptions” with the word ”independent” “to link independent descriptions”
(In reply to comment #2) > One more detail to the proposal: > > 4) In the Abstract section: > Replace the word "extended", > “to link extended descriptions” > with the word ”independent” > “to link independent descriptions” Justification: The wording “extended description” seems to be intimately linked to the "long(er) description” way of describing what the longdesc attribute is. This bug suggests to replace any variant of “long description” with “independent description”, based on the insights from Vlad that was cited above.
I propose to reject this change. 1. The Use Cases and Requirements notes the same contrast Vlad does, in terms I believe are similar. The rest of the document relies on HTML accurately describing what an alt does, which I think is a proper separation of concerns. 2. It is not clear in what sense the description is independent of the image it describes, so I do not believe that it improves clarity. I seriously don't think there is much value in messing with the title, and if there was it would be to change it to use the words "longdesc attribute" The attached version of the document doesn't help convince me. And less so because Search and replace without more careful checking was not a good enough tool to convey your meaning. (By the way if you are going to do that, please *at least* make it clear in the edited version what it represents - leaving stuff for search engines to find and thus for people to read with essentially no context should be done carefully. I believe that's one of the reasons the copyright license of the document does not permit it).
(In reply to comment #4) > 1. The Use Cases and Requirements notes the same contrast Vlad does, in > terms I believe are similar. The rest of the document relies on HTML > accurately describing what an alt does, which I think is a proper separation > of concerns. (The fact that ‘independent’ (IMO) fitted in so nice does (IMO) show how fitting it is. The adding of 'indpedent' fits in. It expresses the same thinking. *However*, adding in the word "independetn" brings out the thinking more - it makes it easier to get an idea about what longdesc is.) > 2. It is not clear in what sense the description is independent of the image > it describes, Strawman: The description is of course not meant to be "indepentn of the image". The dichotomy is, as told: "replacement text" vs "description text". Ans so, the description is meant to be independent of the *context* of the image. Because that is Vlad’s point: the alt text cannot be written without regard to the context where the image occurs, whereas description can be more - or less - independent of that context. That the description is is supposed to be useful independently of the context, is also reflected in the use case section, which e.g. speak about reuse of description. (The use case section was one which I was particulary happy wiht the addition of ”independent”.) >so I do not believe that it improves clarity. Understood as "independent of the image”, then I agree. I think it does improves clarify if one understands it correctly, see above. (Of course, it is possible that the spec should try to explain in what way the description si independent.) > I seriously don't think there is much value in messing with the title, and > if there was it would be to change it to use the words "longdesc attribute" Why not change the title to something that includes 'longdesc attribute'? I cold perhaps live with that. This bug is a reaction to how the spec *currently* stands. Please note: While I do get the sense that *you* consider this an all-or-nothing kind of bug (where you seem to be on the 'nothing' side of it), I don’t consider this bug that way. I seriously and deeply thing that ‘independent description’ (in the sense 'standalone description’ or "more context free description") adds value to this spec. And I would be very happy if you viewed it the same way, but I have no problem swallowing if you would want to go through the spec more thoroughly and see where it fits in (or more important: where it doesn’t fit it.) Hint: The template for how HTML Working Group editors are supposed to response to bugs offers the option "partially accepted.". > The attached version of the document doesn't help convince me. And less so > because Search and replace without more careful checking was not a good > enough tool to convey your meaning. I was surprisingly satisfied with the result ... ... but it is not true that I *only* did a find/replace. E.g. I changed all occurences of “a description” to “an independent description”. In addition to that, there are two bugs which this bug block - I filed those bugs as a result of reading the find/replace result. I also replaced ”long” and ”extended” with “indepdendent” (chicken or egg: technically I added in ”independent” first, and read the result, and found that I had to remove ”long” - which reflected in my edit - later I found I had to remove ”extended” too.) > (By the way if you are going to do that, please *at least* make it clear in > the edited version what it represents - leaving stuff for search engines to > find and thus for people to read with essentially no context should be done > carefully. I believe that's one of the reasons the copyright license of the > document does not permit it). Even if www-archive@ probably is a little under the radar(?), that’s a fair point. Charles, before I eventually do anything more, are on the same page with regard to what I have meant? (namely "description independent of the context").
We're probably on the same page with regards to the purpose and nature of the description, but it is unclear that the changes are important or even especially helpful. I will however run over the document today before publishing the editor's draft, with a view to this bug and what would make things clearer.