ISSUE-30, a poll replies comment

	Two generalizations about the poll replies:

(1) The conformance supporters concentrated their replies on the 
technical and practical features longdesc has. Whereas the conformance 
objectors focused on a negative spiral that longdesc is inside.

(2) One group believe @longdesc is useful and improvable/repairable. 
Another group think it is unrepairable (because it was broken from the 
start or because it has landed in a negative spiral).

	Unrepairable/Repairable:

My own poll reply said that @longdesc is repairable.  This view is 
based on the (infamous) longdesc lottery article, which concluded that 
close to 100% of longdesc usage is incorrect, in a machine checkable 
way. While that statement was meant to be the final bullet to longdesc, 
it also means that, since Validator.nu performs URL validity checking 
for all conforming elements, then, in practical terms and if the 
conformance requirements are strict, the 'obsolete' option would return 
an only marginally higher amount of validation errors than the option 
of making longdesc fully 'conforming' would do.

Unlike Tantek, I don't believe the 'longdesc' name is any more 
misleading than 'src' is. Even if a better name could have been 
possible, the fundamental problem is little use (especially compared 
with 'src' and 'href'). Since HTML4 validators did not perform any URL 
validity checking whatsoever, authors did not get the necessary 
resistance when they tried to validate @longdesc with "creative" 
content.
-- 
leif halvard silli

Received on Sunday, 4 July 2010 00:06:22 UTC