Everyone,
It
was great to meet you at SOS20!
I
promised to follow up by email to keep our conversation
going, and to help nucleate a community around workflow
management issues common in scientific computing and
commensurate with our shared interest in HPC. We
should consider what we want this discussion and
community to be about. I propose that in addition
to normal collaboration (coding, research, etc.) we help
each other with general questions, share experience,
help communicate domain science needs, and to engage
others in the workflow community. It would be especially
good to engage members of project teams for other
workflow engines such as Pegasus, Fireworks, Galaxy,
Taverna, etc.
One
of the things we wanted to figure out was whether or not
we should pursue our discussion as part of the Science
Working Group, which is managed by the Eclipse
Foundation, or if we should start something on our
own. My personal recommendation is to do this as part of
the Science Working Group because we already have a
mailing list, community forum, and other infrastructure.
Please write back and let me know how you feel.
I
also have five action items from the end of our BOF in
addition to sending out this email. These are only what
I heard, so please feel free to add to the list. They
include:
1.)
Figuring out if there is an opportunity for Aiida to
work with other workflow engines.
2.)
Learning more about Robert Clay's Scientific Data
Management Proposal.
3.)
Planning or attending workshops to help build community.
(Robert is working on a couple already).
4.)
Figuring out how to connect the Systems and Workflow
communities.
5.)
Start a Google Doc for discussing Provenance models.
(I'll get to it.)
Looking
forward to working with you,
Jay