Working Group Decision on ISSUE-66 - Image Analysis Heuristics

Original issue:

Poll results (including links to change proposals):

We have a change proposal that denies that there is a problem, and 
plenty of input that there is a problem with the current text.  We have 
partial refutations of a number of these problems, which we will 
grant... but only as far as they go.

For example, Matt writes that heuristics cannot exist.  Ian's (valid) 
reply is that there is no claim that it solves all problems, and that 
_any_ improvement is an improvement.

The last half of that statement would be worth considering, if there was 
_any_ evidence or rationale provided that the actual text in the spec is 
helpful.  Looking at the both the objections and the change proposals 
themselves, we are unable to find any such statements.

By contrast, there is widespread belief that the text is "not useful, 
confusing, and potentially harmful".  This statement, combined with 
Matt's specifics, we believe constitutes a strong objection.

The suggests that the right course of action is to remove the text. 
However, doing only that would not address Ian's objection, which we 
will characterize as "it would be a mistake to not provide helpful text" 
(the key difference between this phrasing and Ian's original is that he 
makes the further, and to date unsubstantiated, assertion that the 
controversial text that he authored is helpful text).

We believe that Larry Masinter, Martin Kliehm and Laura Carlson have 
suggested the correct way forward.  Larry stops short of saying that 
Ian's recommendation should not be included; instead Larry simply 
suggests that the document into which this recommendation be placed be 
one that has the potential to evolve independently.  Martin Kliehm and 
Laura Carlson both suggested UAAG in their objections as the proper 
place to provide this assistance.

Such an approach should satisfy all stated objections.

  - - -

Going forward, we should more carefully review change proposals and 
ensure that they actually provide rationale FOR the status quo, and not 
merely limit themselves to pointing out deficiencies in other proposals.

- Sam Ruby, on behalf of all three HTML WG co-chairs.

Received on Tuesday, 1 June 2010 21:52:38 UTC