Present: John Unsworth, Catherine Plaisant, Martin Mueller, Stan Ruecker, Amit Kumar, Matt Kirschenbaum

Agenda:

Jira tickets: signing up is not enough--you have to add yourself as a watcher. AMit will create a fake ticket with Stan and Catherine as watchers, to test.

Staffing of interface development: Alejandro is signed up for summer, and he's working weekly with Matt (possibly with Andrew skyped in).

Post-hackfest: Stan will ask hackfesters to update Jira tickets. Amit will add JMU to watch.

Data cell discussion, Feb. 22nd: Analytics of text display and visualization post on the wiki followed on that discussion. Is the discussion of mid-level xml structure feature creep? Is the MySQL datastore the right place to treat this problem? What about SEASR/RDF? What to do, when to do it, how to do it?

Big ideas: units for display aren't necessarily the same as units for analysis; the latter may need to be more flexible. Also, you may want to choose chunks (for either) that are not xml containers. Not all these chunks would necessarily be indexed--but (says Amit) when you go to containers determined at run-time, things get slow. The advantage of RDF is that an RDF datastore can store lots of relationships, deriving them at run-time. This could augment the MySQL datastore.

Martin: if we can use the RDF datastore to complement NySQL, that would be great. How do we avoid duplication? Can we proceed in parallel?

Amit: this is where workflow becomes important, so that the same data is being fed to two datastores, with something like a unique ID that binds the two together. Steve Ramsay needs to finish the workflow process, perhaps with help from Amit, and then John Norstad could come down to meet with Amit and NCSA people (Bernie in particular), to get a working end-to-end example.

Matt has set up ManyEyes call 2 pm Eastern: anyone is welcome: we will be talking about the hands-on route to collaboration, using their APIs. Amit and Stan should be talking about implementation. The proxy is ready to export data on demand; do we embed their applet? Do we push data to their applet? Catherine and/or Matt will send an example to Martin W.

Martin Mueller's brief on the Mellon meeting, end of February... Loretta did a presentation; Bamboo (a UK project seeking funding); the most useful part was a discussion of how to get the projects to collaborate, perhaps by meetings that would nudge participants to collaborate across projects, perhaps by making PIs from one project advisors on another. A lot of discussion was on infrastructural projects (catalog systems, administrative systems). SEASR and MONK seem slightly anomalous in the group of projects that Chris oversees; for most of these, Google seems to be taking up the space that they want to occupy. Fedora was also a subject of discussion (Lagoze, Payette were there): the issue arose of whether connecting everything with everything is a good idea. Scholarly tools are at the margin of this discussion, though that doesn't mean they are not of interest. SEASR seemed more central; the connection with Fluid is going to be worth pursuing.

Sara is working with Meandre, limited use (no stopword removal, no other filters). Proxy calls exist: can we get an update from Matt, Alejandro, Andrew. Stan will coordinate a call this week, with Catherine, Amit, Matt, Alejandro, Andrew, to review progress on this.

Catherine: Anthonny's reimbursement for featurelens? Out of the MONK budget? Did Suzi get this to UMD? If not, can we direct to interface development?