[esip-semanticweb] debrief/lessons on ToolMatch Mixer

[Lynnes, Christopher S. (GSFC-6102)]

I think the ToolMatch Mixer was pretty successful.  We successfully engaged some members of the community with little prior direct knowledge of the Semantic Web.  And we also surfaced a number of important issues that would be helpful to address early on.

Here are some lessons:
(1) TitanPad does not work so well for iPad users.  Is HackPad any better?

(2) Nearly everybody wants to get to some shortcuts for linking data and tools, most of them based on the data format. This is not as trivial as it appears at first glance. 

(3) In the absence of format-based (or other) shortcuts, they will script it. Which is not necessarily bad, but can lead to statements that are…unverified.  Or factually untrue in some cases.

(4) Because of #3, it would be useful to have the ability to accept and mediate contradictory statements, which implies allowing negative statements (tool A does NOT visualize collection B), and/or to verify statements (not always possible).

(5) W.r.t. #2, will will STILL need a substantial number of individual, well-curated tool-works-with-data statements to ground-truth the inferencing.

(6) Folks also had a desire to carve out their own namespaces for identifiers, rather than using the globalchange.gov namespace.  I'm kind of OK with that, but what do the rest of you think?

(7) Dydra:  meh. Good for non-technical folks like me to get something up in a hurry, but that's as far as it goes.

(8) So…should we shoot to put the ontology in the ESIP Bioportal as soon as it is ready?  Or what?  It would be good to have a stable home for the repository as we build out the population in the next few months...
--
Dr. Christopher Lynnes, NASA/GSFC, Code 610.2, 301-614-5185