Jeni and Dan, etc,
I've been off the radar, so to speak, for a few weeks due to some ongoing
health issues. I will miss tomorrow's call too.
I have continued to work on my js implementation, which I still fully
intend to give to the group as a project deliverable. It works now for
simple csv-to-json transformations, but doesn't make turtle (yet). My
implementation is a stream transformation, so it is pretty fast and can
theoretically handle input streams of any size.
I'm a bit unclear on how simple "simple" needs to be. Does "simple" mean
no metadata at all? Or is there a "simple" metadata transformation I need
to implement?
I have experimented with taking metadata from the link headers, etc;
that's all pretty straight-forward. But merging multiple sources of
metadata is, of course, not straight-forward, and obviously not simple.
For now, my main concern is that my code fulfills the requirements of a
simple implementation.
I still need to document the project and clean up some rough bits. Then I
can register as an official testing endpoint for Gregg's CSVW Test
Harness. What's the deadline?
That's all for now.
Thanks,
Bill
--
Bill Ingram
Manager, Scholarly Communication and Repository Services
University Library
University of Illinois at Urbana-Champaign
(217) 333-4648
wingram2@illinois.edu