System Requirements

This thread actually started at the 21 June face2face but it is nearly 
identical to Charles' previously titled "baseline capabilities" endeavor in 
the archives at
http://lists.w3.org/Archives/Public/w3c-wai-gl/2000OctDec/0147.html

His proposal "A technology is considered to be sufficiently implemented 
when it is
implemented in at least two free products available that meet the following
conditions:"

seems to me to be not quite the right direction to approach this matter 
from. We might provide a "matrix of assumptions" from which content 
providers select items to which their offerings conform, perhaps rooted in 
EARL assertions so that there is no cast-in-stone list of "sufficiently 
implemented technologies".

Just as the presence of text tacitly assumes literacy, so one who posts 
streaming media has presumed an appropriate media decoder as a prerequisite 
for accessing the presented material. A video teaching one to use a video 
machine is often pointless!

Perhaps our "urge to control" gets the better of us when we deplore that 
providers might move towards exclusion when the facts of "the market" might 
very well encourage inclusive strategies. Providers need to be able to say 
"in order to access this material you must have a means of parsing 
(something-or-other)ML. You can't direct authors to use, e.g. CSS without 
them able to say "only works with browsers that acknowledge CSS" - same 
with SVG et al.

Just as we had a blind spot for the universality of text ("text is not 
sacred") we now are overlooking all the fundamental "system requirements" 
that we have silently assumed. We might as well provide a mechanism for 
explicitly expressing those requirements. An example is that some 
state-wide policy I heard of demands that all access to the Web within that 
government's purview be through a certain version of NetScape, even though 
doing so "disenfranchises" many users to much material. Same with many 
intranet applications - the users of the system *must* have a particular 
setup. The WWW is just such a closed system.

It will be convenient to have some RDF-like assertion that one can refer to 
and make claims within which will enable a user to know if there's any 
point in logging on to that source. Our conformance claims mechanism can 
include objective patterns of such matrix subsets. Just as the download 
information for software via FTP/HTTP includes the size of the target file, 
it must include the (ta-da) *SYSTEM REQUIREMENTS" as a conformance matter.

If there is any practical/realistic alternative to this practice (which 
already exists in unspoken form), let's find it. If not, let us embrace it, 
however unwillingly.

--
Love.
                 ACCESSIBILITY IS RIGHT - NOT PRIVILEGE

Received on Thursday, 21 June 2001 09:13:16 UTC