- From: Ben Boyle <benjamins.boyle@gmail.com>
- Date: Tue, 9 Sep 2008 23:27:22 +1000
- To: "John Foliot" <foliot@wats.ca>
- Cc: "James Craig" <jcraig@apple.com>, "Lachlan Hunt" <lachlan.hunt@lachy.id.au>, wai-xtech@w3.org, public-html@w3.org
(offlist) On Tue, Sep 9, 2008 at 4:34 AM, John Foliot <foliot@wats.ca> wrote: > The > relevancy or irrelevancy of providing information to end users is not for > authors to decide: you, as a sighted user can see that patch (and in fact > the whole photo collage) and study it or ignore it as you choose - a > non-sighted user does not have that ability, and thus they are being > short-changed. I find the 'not for authors to decide' remark odd. Maybe it's not a decision about relevance, but it is certainly in the hands of authors if (and if so, how) they provide content. Got me thinking that maybe trying to get authors to do more of this is not the smart approach. Maybe we need those in the community with the skills and time to describe images that exist on the web... maybe a browser plugin that people can share to annotate images. Hmm... surely these already exist? Wow, look at them all [1]. Bit like the blind photographers that get friends to describe their photos, why not get the community to describe all the other images on the web? Think of all the descriptions we could write instead of posting to this mailing list! Hmm... got to cycle the modem. "Broadband" in Australia, dodgy. My approach to authoring, I prefer content to be available to everyone. Yes I understand that this means images need alternatives. But I also apply this thinking to @longdesc and @summary (and hidden skip links) content. I don't want that text hidden by most browsers. If I want to summarise information in a table, then I want that summary available to everyone. I wish everyone could see the images (I expect another letter from the Fred Hollows Foundation any day now). Generally I can't do too much about that, but when I take the time to describe an image in detail, that description is for everyone. And I am really looking forward to using figure/legend to do this. I know longdesc, as argued for on this list, has a different purpose. It's about explicitly describing the image, describing it honestly, exactly what you see, with no inference or omissions. When I think about what it means to do this, I remember times I have listened to my Aunt describe photos to my (blind) Uncle. All the little details, calling out the bits of interest, drawing on similarities and contrasts between other photos or real life ... "oh here's a photo of Ben, oh he is making a funny face. he has a shirt with some sort of cartoon robot on it..." Depending on the photo, on the context, that description might be short or much longer ... You can put some of this in figure/legend (or just generally "near" the image, is a pretty good approach imo). You can put a little more in @alt. Something short and snappy that's key (about the image) but not in the figure/legend. That's usually enough I feel. If I want more, I just put more in the figure/legend. And yes, if it reached the point where it was describing things that were so obvious (in the picture) as to be silly, perhaps that description should be hidden and @longdesc would do that effectively. I contend that even then, it might be helpful to present that description visually, so it could benefit others. People with comprehension difficulties. People with dodgy broadband that just died and left them with a page full of missing images. It has value beyond screen readers! It's not like it is difficult for other people to skim over, it won't ruin their experience. And it needn't ruin the aesthetics of the page either. Design it into the page already. But that's one approach. One choice. I fully support authors choosing to write and hide long descriptions. But I have to ask myself, and this group, why do I need @longdesc when I have these css techniques? [2] example: figure legend { position: absolute; top: -10000px; ... } example: figure legend .longdesc { position: absolute; top: -10000px; ... } (assuming some of "figure legend" is visible, and the extra "longdesc" bit is hidden). This technique is already in use, already recommended, and works in a lot more situations than longdesc (i.e. not confined to images). It has the advantage of not having to create a separate page (reduced overhead for authors) and doesn't send screen readers "round the back" (figuratively speaking), it reduces http traffic. Unless there is a burning concern that where css is disabled and screen readers are not in use, long descriptions must be hidden. That would be a need for something directly implemented in HTML (like @longdesc), but I think that's fishing for a need that doesn't really exist. At the very least, I don't need to be able to author content that way. You can't explicitly associate the description with the image in HTML4, where we only have @alt and @longdesc, but with figure/legend on the table for HTML5, I do believe this problem be resolved. Except that figure/legend is not intended to be greater than one paragraph. I can work with that. (Have I missed anything important?) [1] http://www.makeuseof.com/tag/web-annotation-tools-research-annotate-collaborate/ [2] http://webaim.org/techniques/css/invisiblecontent/ cheers Ben
Received on Tuesday, 9 September 2008 15:00:32 UTC