RE: "at a Minimum" (resend)

Quote 
Minimal requirements don't dictate the implementation details of the
required functionality -- they merely define what base functionality
MUST be available to the user...
endQuote

One of my personal issues with the wording "at a minimum" is that it
lends itself to easily to giving implementation details.

As an example the new ATAG  
1.1 Ensure that the author can produce accessible content in the markup
language(s) or content type(s) supported by the tool. [Priority 1]
At minimum, provide a code editing view that will accept manually-coded
accessibility content and ensure that this content is preserved as per
Checkpoint 1.2.

What if a company wanted to product a product that hide the fact that it
was producing HTML under the covers, but provided all accessible content
through the UI in some method - ByDesign they wouldn't meet the minimum
but they still produced accessible content.

I'm running off (I'm being forced to spend time with my fiancé in
Hawaii), but Annuska can continue this discussion until I get back...
Heather Swayne
Microsoft

 -----Original Message-----
From: 	gregory j. rosmaita [mailto:oedipus@hicom.net] 
Sent:	Thursday, April 26, 2001 11:11 AM
To:	w3c-wai-au@w3.org
Subject:	Re: "at a Minimum" (resend)

aloha, all!

please ignore my previous post, which somehow escaped the bonds of my
system
autonomously, due to a virtual device driver freeze that caused all
input
via the keyboard to be input as if depressed in conjunction with the ALT
key, with the result that, when i typed an "s", the message was
unexpectedly
sent in an unfinished state, and i apparently wasn't able to "stop"
transmission, even though outlook express told me that i had been
successful
in stopping the send...

in any event, what follows is my completed post...

--- RESENT MESSAGE ---
From: "gregory j. rosmaita" <oedipus@hicom.net>
To: <w3c-wai-au@w3.org>
References:
<B7E0BEA478EBC24EBB2CF19F710760230170F0D2@red-msg-06.redmond.corp.micros
oft.
com>
Subject: Re: "at a Minimum"
Date: Thu, 26 Apr 2001 14:09:18 -0400
Organization: oedipal enterprises, (very) ltd.

aloha, heather!

while i am in philosophical agreement with some of the points you
outlined--in particular, the first bulleted point:

quote
Including text like "at a minimum could lower the bar, to allow product
groups to only do that minimum level of work.  As apposed to allowing
individual companies to define their own minimum, or standard, that they
want product groups to follow.
unquote

i am loathe to let my (and others') cynical opinion of human nature lead
me
to the ironclad conclusion that the "default" reaction to minimal
requirements will be to use them as a lazyman's loophole...

so, while i agree that there is a very real danger inherent in the
concept
of explicitly stating a minimal satisfactory implementation of a
checkpoint,
that isn't what is under discussion...  what the "at a minimum" clauses
are
intended to provide is a description of the absolute base functionality
required by the checkpoint, and realpolitik dictates that the dangers of
not
explictly outlining base functionality for a checkpoint far outweigh the
dangers of either providing a loophole for lazy developers by forcing
them,
at the very least,  to provide well-defined functionality...  minimal
requirements don't dictate the implementation details of the required
functionality -- they merely define what base functionality MUST be
available to the user...

moreover, when WAI guidelines (such as the User Agent Accessibility
Guidelines and the Authoring Tool Accessibility Guidelines) do not
contain
explicit requirements for minimally satisfying a checkpoint, the working
groups have
received requests from developers, claiming that the affiliated
Techniques
documents, due to their informative/non-normative status, cannot
sufficiently address what is minimally required to satisfy a particular
checkpoint...  so, while i share your concern about the potential
straitjacketing effect normative minimal requirements may have upon
developers, there is also a clear need to define normatively what
exactly
constitutes minimal satisfaction of the checkpoint in question, as well
as
the need for providing context for the checkpoint...

for example, what use would it be to simply state that a user agent MUST
offer a search facility, if the basic parameters for a search facility
aren't explicitly stated?  the ability to search is the absolute
minimum,
but the ability to search is quite an amorphous concept, so the base
functionality necessary for a UAAG conformance claim is spelled out
explictly in the User
Agent Guidelines [reference 1]

as for your last point, quote Will products have to implement "the
minimum"
even if they have "an advanced" solution? unquote, while the answer is,
"yes", the reply is not as onerous as one might assume... i know from my
experience as a member of the User Agent working group that what the UA
WG
intends by expressing minimal requirements for checkpoints is to express
the
functionality necessary to minimally satisfy the checkpoint, and NOT the
mechanism whereby the functionality is achieved -- that, as you
correctly
point out, is fodder for the Techniques document...  granted, in order
to
express the minimal functionality required by a checkpoint, it may be
necessary to more explicitly reference specific markup languages, but
where
the User Agent WG found it necessary to explicitly state base
functionality
with reference to a specific markup language or modality, the minimal
requirements are expressions of the minimal functionality required by
the
checkpoint, and not the "preferred", "ultimate", or "one-and-only" means
of
achieving that functionality...

i'm curious as to whether resistance to the idea of minimal requirements
has
been discussed with the product groups responsible for IE, as the MS
reps
who have played an integral part in developing and shaping UAAG did not
object to the insertion of minimal requirements into UAAG, and several
of
the minimum requirements that appear in the Last Call Cubed draft of
UAAG
were derived directly from comments and requests for clarification from
Microsoft reps/commentors...  is the resistance to minimal requirements
a
reflection of the broad net cast by ATAG, which--in the case of
Microsoft--would cover such diverse products as FrontPage, Word, and
Publisher, at least as regards their "Save to HTML/Web" functionalities?

gregory.

References
1. the "search" checkpoint is UAAG 9.6 in the 11 April 2001 Last Call
Cubed
draft: <http://www.w3.org/WAI/UA/UAAG10/guidelines.html#gl-navigation>

--- ORIGINAL MESSAGE ---
Date: Thu, 26 Apr 2001 08:20:10 -0700
From: "Heather Swayne" <hswayne@microsoft.com>
To: <w3c-wai-au@w3.org>
Subject: "at a Minimum"

With regard to the proposed changes for ATAG v2.  I have now talked with
several Product Groups here at MS, and the general feeling is that they
do not like the idea of including "at a minimum" within any of the
guidelines or sub text.

Some examples of their concerns:
* Including text like "at a minimum could lower the bar, to allow
product groups to only do that minimum level of work.  As apposed to
allowing individual companies to define their own minimum, or standard,
that they want product groups to follow.
* ATAG should not be telling product groups how to implement
guidelines.  The techniques document should be used to show examples of
how a range of products met a given guideline.
* "The minimum" for one product could be something totally
different than the WAIs suggestion as the minimum, does that mean it's
wrong?  Even if ATAG doesn't think so, others may.
* Will products have to implement "the minimum" even if they have
"an advanced" solution?

Heather Swayne
Microsoft

Received on Friday, 27 April 2001 22:02:30 UTC