alt and authoring practices

Looking over the lengthy discussions on the alt issue, it seems to me  
there's some fundamental misunderstandings going on: especially  
regarding the omission examples. Since alt is properly used as an  
alternative when the doc.ument is not available (leaving longdesc and  
other mechanisms for other textual content), then the often sited  
Flickr example doesn't really apply. I'm more familiar with iPhoto  
(and its integration with .mac), so let me use that as an example.

iPhoto extracts sometimes hundreds of photos from my camera in a bulk  
operation. With a few clicks I can publish those photographs to the  
web as a web gallery. These photographs are presented as a web gallery  
along my other web galleries. A visitor to the web gallery sees the  

1) the initial web gallery page with a 'key' photograph for each  
gallery (like a photo album). This key photo constitutes a navigable  
link to view the thumbnails of all of the photos in the gallery. It is  
a photo that should therefore have appropriate value for its alt  
attribute. Something along the lines of "view entire x gallery", where  
x is the name I've given to the gallery in iPhoto.

2) next on the gallery page, each photo in the gallery is represented  
by a thumbnail. By clicking on the thumbnail a higher resolution and  
larger photo is loaded. Again, these thumbnails require alt text.  
Something like "view fullsize image" or view image y fullsize" where y  
is the name of the photo. The initial name of the photo is  
automatically generated by iPhoto or my camera and usually in the form  
IMG_####.jpg. This means the alt attribute would look like this:  
‘alt='view image "IMG_1234" fullsize”. If I changed the name  
of the photo to be more title-like it might instead be:  ‘ alt='view  
image "Joe at the Beech" fullsize' ’.

3) Finally the image is displayed full size with a series of controls  
for downloading and viewing the next image. Here the image "Joe at the  
Beech" would properly have the attribute ‘alt='return to contact sheet  
of the entire gallery '’. Here the alt is easily generated by the  
software authoring tool. The controls on the page are iconic and  
therefore also require alt text: and again it is alt text  
automatically derived from the title of the photos  (or simply  
"download photo", "next photo", and "previous photo" to avoid adding  
the titles of the relevant three photos).

In this iPhoto scenario it is very rudimentary for iPhoto to add the  
appropriate alt text to every one of these photos. It doesn't matter  
that I'm adding hundreds or even thousands of photos to my html  
document with just a few clicks in a GUI authoring tool. The user of  
iPhoto need not know anything about HTML4, HTML5, or have ever heard  
of the W3C.

Where things might matter is with the optional longdesc attribute, or  
with the image format supported description and title metadata. Here,  
iPhoto allows me to add descriptions to my photos and give them  
appropriate titles. However, that information remains in the image  
supported metadata (inside the jpeg file) and is not included as a  
document fragment with a longdesc attribute reference pointing to the  
document fragment. Here, it may be cumbersome for user (author) to  
provide descriptions of every photo of hundreds uploaded to the web.  
However, it is even more disappointing for an author who has already  
included such descriptions to have their authoring tool leave them out  
of the generated content. Again, this has nothing to do with the @alt  
attribute however. The @longdesc attribute is optional in HTML4 (and  
clearly some mechanism to explicitly associate embedded media with  
semantically rich descriptions should be included in HTML5)

If Flikr has something different that I'm missing, then I think  
someone needs to elaborate how Flikr works that would necessitate some  
different approach and an optional alt.

Take care,

Received on Saturday, 3 May 2008 13:14:02 UTC