Loading…
PIDapalooza 2018 has ended

Welcome to PIDapalooza 2018...where anything goes...as long as it goes on forever.   


View analytic
Tuesday, January 23 • 11:30am - 12:00pm
Making groups a first-class abstraction of PIDs

Sign up or log in to save this to your schedule and see who's attending!

There's a lot of discussion about PIDs for groups right now, and different discussions include different ideas of groups. Some talk about groups of people, and some talk about groups of objects more generally.  We probably don't want to set up lots of new types of PIDs for each kind of group, but what else can we do?

In this session, we'll talk about making groups a first-class abstraction of PIDs. PIDs now have a single abstraction: that of a naming authority and a local namespace. Groups can be supported in PIDs today by using the relationType, but this is buried, not a high-level goal. If we had the concept of groups in PIDs, this would be useful in organization IDs (an organization is really a group, perhaps with a hierarchy), software IDs (software projects are groups of software releases), and data IDs (data collections and data sets are groups of data or data streams, and data streams are also groups of data).

Can we change the underlying assumptions about PIDs at this point?  Should we?  Let's talk, and figure it out.

(slides from this session are available at https://doi.org/10.6084/m9.figshare.5812881.v2 

Speakers
avatar for Daniel S. Katz

Daniel S. Katz

Assistant Dir. for Scientific Software & Applications, NCSA; Research Assoc. Prof., CS, ECE, iSchool, University of Illinois Urbana-Champaign



Tuesday January 23, 2018 11:30am - 12:00pm
Stage 1