Re: Interrelationships between dimensions of variability

DOM, you make it sound so one-sided.  We have to deal with the fact that it
is a cost-vs-function tradeoff.  Bundling is efficient; leaving decisions to
be made later is somewhat costly and particularly it is risky.  But it
brings opportunity.  So we cannot justify microfine subdivision into aspects
on principle alone.  Each distinction we want to be captured has to be
justified with concrete value added.  This is the approach that I was
following at WWW2003 in articulating specific cases where it is important
that some things be handled differently between different user sessions with
a given web Ap.

The central model here in my mind is the sub-topic of "analysis of variance" in
multivariate statistics.  One over-samples the 'aspect' space by making many
different observations, and then analyzing the results to estimate how many
_independent_ variations there are actually going on.

Content construction policy is a dual to the description of demographics,
but in accessibility and adaptation we are confronted with the same issue of
trying to reduce the welter of controllable parameters into a smaller number
of functionally-allied aspects for the purposes of testing "is the raw
content flexible enough."

Just a few quick cites:

  From today's epistle (don't agglutenate class tokens):
  http://www.w3.org/mid/5.1.0.14.2.20030603092338.021b6400@pop.iamdigex.net

 From WWW2003 (what divisions should combine differently in different 
delivery contexts):

  Divergent adaptations - Flex Requirements
  http://www.w3.org/2003/Talks/0524-di-ag/slide6-0.html

New accessibility document from IMS:

  Welcome to IMS Global Learning Consortium: Specification: Accessibility
  http://www.imsglobal.org/accessibility/

Al

At 08:54 AM 2003-06-03, Dominique Hazaël-Massieux wrote:
>Hello,
>
>I'm in the process of evaluating the work needed to write the new
>chapter about the concepts used in the QA specification guidelines [1].
>One of the section of this chapter is supposed to be a generic approach
>on the relationships between dimensions of Variability [2].
>
>In my opinion, a good way to sensibilise the readers about this aspect
>would be to provide an informative table about typical problematic
>relationships between DoV. The goal would be to fill each cases of the
>table at http://www.w3.org/QA/WG/2003/06/dov-relations
>
>So, if you have any experience showing how such or such DoV
>relationships hurt interoperability, or some specific details on points
>that should be addressed when combining DoV X and DoV Y, please share
>them with us, and I'll try to keep the table up to date (and probably
>integrate it in the SpecGL later on).
>
>Thanks,
>
>Dom
>
>1. http://www.w3.org/QA/WG/#docs
>2. http://lists.w3.org/Archives/Public/www-qa-wg/2003Apr/0171.html
>--
>Dominique Hazaël-Massieux - http://www.w3.org/People/Dom/
>W3C/ERCIM
>mailto:dom@w3.org

Received on Tuesday, 3 June 2003 09:51:43 UTC