[uf-discuss] RFC: sHTML Video Thumbnailing

[Benjamin Hawkes-Lewis]

Colin Barrett wrote:

> On May 28, 2007, at 1:49 AM, Benjamin Hawkes-Lewis wrote:

[snip]

>>  If a screen reader read (for example):
>>
>> quote Dorothy encounters the Lion end quote
>>
>> That would be rather strange, wouldn't it?
> 
> I dunno, I think that might be helpful. It's semantic information that 
> it's a portion of a larger document.

This is a tempting argument, but in theory and practice a problematic 
one. <q> and <blockquote> are not merely intended to be "portions of a 
larger document" but to be /quotations/:

http://www.w3.org/TR/html401/struct/text.html#h-9.2.2

I'm not sure that a thumbnail really is a quotation, although it's 
clearly conceptually close to one. When people talk of quotations from 
movies in everyday speech, they are talking about quotations from the 
dialogue not stills, e.g.:

http://www.imdb.com/title/tt0032138/quotes

I don't think "quote Dorothy encounters the Lion end quote" would be a 
human readable hint that it's a thumbnail from a video. Unless you 
happen to be a microformats guru. ;)

> Using a span tag as you suggest provides the UA with zero semantic information.

I'd prefer to provide zero information than potentially misleading 
information. I suppose one could prefix the alt with "still:" though:

alt="still: Dorothy encounters the Lion"

So then you might hear:

graphic still Dorothy encounters the Lion link The Wizard of Oz

A microformat parser could remove everything up to and including the 
first colon to get to the alternative text proper.

> I suspect having alt 
> tags that just link to a video which perhaps they don't want to watch is 
> annoying to people with screen readers -- although I think I would need 
> a bit more data about how screen readers work and how they're used to 
> really say anything else.

I believe your suspicion is wrong for three reasons:

1. Most screen reader users are not deaf as well as blind, and many 
screen readers still have some sight. So most screen reader users are as 
likely to enjoy watching videos online as you and I. Their big problem 
with sites like YouTube is that it's too hard to /find/ videos, not to 
watch them:

http://www.bbc.co.uk/blogs/access20/2007/05/access_20_interview_mark_prous.shtml

2. Most video formats, include Flash video and Quicktime, can include 
captioning which mainstream screen readers can push to a braille display 
for deafblind users. Videos /should/ include captioning and authors 
should make it clear when they do not: and ideally provide a transcript 
as an alternative. There are now captioning websites that either make it 
easy to add captions to online video or will even caption it for you for 
free on request:

http://www.bbc.co.uk/blogs/access20/2007/05/captioning_video_gets_easier.shtml

3. Even if a video is completely inaccessible for some reason, screen 
reader users may use the alternative text to understand why the author 
is referring to that video, to share the video or thumbnail with sighted 
friends or colleagues, or to demand an accessible alternative. Just 
hiding inaccessible materials would produce confusion.

On an even more basic technical level: leaving out alt would result in 
some screen readers reading the image src attribute. Leaving alt blank 
would result in some screen readers reading the href for any surrounding 
anchor link when alt provides the only text content for the link, e.g.:

<a href="http://www.example.com/video"><img alt=""></a>

By default, Window-Eyes would read something like:

link h t t p colon slash slash w w w dot example dot com slash video

Commonly, I think, the user will have reduced the verbosity to strip out 
most punctuation and only have to suffer:

link h t t p w w w example com video

Takeaway: include an alt for all images; and always include an alt with 
actual text if the img is the /sole/ content of the link.

--
Benjamin Hawkes-Lewis