Re: Impact of not setting baseline and writing SC as functional outcomes

These are good proposals that should be integrated into the guidelines
irrespective of the outcome of the "baseline" discussion.

I support most emphatically the claim that the success criteria should
be written in generic terms, independently of variations in UA and
other technology-dependent details.

I would also argue that the very term "baseline" is misleading; a
baseline in typesetting is the line above which a character is
printed; yet the main use of the concept of a baseline in these
discussions is to require content to be compatible with user agents
that meet the baseline and preferably some of those which fall below
it to various degrees.

Perhaps the concept of a profile is preferable here: the content has a
profile, comprising the formats, protocols and API's that it is
written to support, some of which it requires in order to be
operative. A user agent also has a profile, or a set of parameters
defining the supported and active formats, protocols and API's, in
addition to its accessibility-related characteristics. Where the two
profiles match, at least in so far as the required items in the
content's profile are included in the user agent's profile, the encoded
content will be successfully decoded and acted upon.

Any sample profiles (or baselines if you still prefer that term)
provided by this working group must be non-normative, and should be
offered either in the guide to the guidelines document, or in a
separate document. It is reasonable to limit the scope of techniques
to the sample profiles provided. Also, as pointed out earlier in this
discussion, there should be a limitation imposed on the range of
acceptable content profiles by the guidelines themselves, based on the
existence of implementations that support, or largely support, UAAG
Level A. That ground has been adequately covered and summarized by
Wendy, and thus need not be recapitulated here.

Received on Friday, 1 April 2005 05:16:25 UTC