Challenge Cases

[Kevin Thiele]

Welcome back!

I actually posted this message a week ago, but my email address has changed
slightly, and the list server rejected me. So here it is again.

At the Sydney TDWG meeting we agreed that we would continue with the SDD
discussion on this list, but try to keep a tighter focus. The last active
period (about 12 months ago) was an important brainstorm session, but didn't
seem to be very effective at actually getting us to the goal of a workable
standard.

We decided this time to try working through some challenge cases - real or
made-up instances of descriptive data that need to be accommodated in the
standard. By agreeing first on a +/- complete set of challenges, then
working through the challenge cases in order from simple to difficult, we
should be able to reasonably bound our problem while keeping an overview of
the territory while actually sinking out teeth into the nitty-gritty.

Attached is my first attempt at a set of challenge cases, presented in +/-
this form to the Sydney meeting. The first challenge case has an exemplar,
the others have not yet. This document as it's worked up will be placed on
the TDWG web site for working reference. As the standard evolves, this will
also be put up on the site, possibly with progressive status indicators for
parts of the standard (e.g. working, proposed, normative - may we one day
get to normative)

I suggest that we should first add to or modify the list of challenges.
Propose a challenge (with exemplar) to add to the list. Once we are happy
with an approximate list of challenges (keeping in mind that others will
become clear as we proceed, so there's no need to agonise over this step),
we'll start with challenge 1. We'll throw up the challenge, give a week or
so for contributors to propose data structures that can meet the challenge,
then compare and discuss alternate solutions.

Gregor I think will shortly be posting a summary of the meeting discussions.
We agreed, I believe, that the goal is to provide a standard that can
adequately address the descriptive data requirements (ie be a superset) of
all existing programs (e.g. Lucid, DELTA, DELTA Access, Biolink) but not be
limited to existing programs. It should be able to function as an
interchange standard, but should not be limited to that. Bob Morris agreed
to provide shortly a discussion on interchange vs interoperability standards
and conflicts that may arise in trying to allow for both goals. We agreed
that XML will be the basis for the standard.

May the force be with us!

Cheers - k