HSF Weekly Meeting #129, 8 March, 2018
Present/Contributors: Graeme Stewart, Dario Menasce, Torre Weanus, Charles Leggett, Paolo Calafiura, Pere Mato, Daniel Elvira, Liz Sexton-Kennedy, David Crooks, Eduardo Rodrigues, Maria Girone, Pete Elmer, David Lange, Oli Gutsche, Helge Meinhard, Giulio Eullise, Simone Campana, Mark Neubauer, Martin Ritter
HSF/WLCG Workshop
-
General points
- 196 people registered.
- Rooms for each of the sessions have now been put into Indico.
Note that plenary sessions are all in the hotel; during the
parallel sessions we use the main hotel room plus 2 rooms
in the conference centre. The two venues are 200m apart.
- We have coffee available in both locations when necessary. All
the lunches are at the conference centre.
- Vidyo Rooms are created (one for each location). Graeme is the
owner. If everyone knows the moderation PIN is this enough to
manage these?
-
Outcomes
- First list of general questions that convenors could consider
when thinking about their sessions. Aimed at identifying what
projects we can realise after Naples; and for those we cannot,
how do we enable them?
- What projects were identified that we can advance from the
CWP Roadmap as a community?
- How will these be followed up on?
- Are the timescales for the project well defined?
- Is the project sufficiently covered for effort right now?
Do we need to ask for further involvement (other
experiments or groups)?
- Does the project have resource needs that are not yet
covered? (Infrastructure and funding.)
- What projects were identified that we can’t advance from
the CWP Roadmap as a community?
- Are these critical items for the next decade?
- Can a case for funding these items be developed?
- Are there links with other (non-HEP) communities…
- That are established or that would be desirable?
- How will the projects developed improve our community’s use
of developer effort between experiments? Can we enhance
that?
- What’s the best way of organising ourselves (meetings,
workshops, mailing lists) to work effectively in the next
few years?
- Post CWP R&D: How does HEP-TrkX connect with ACTS (for example)?
How do these projects work together? This is an important
question.
- R&D that will inform the ATLAS and CMS computing TDRs is a very important milestone.
- HSF could curate a map of projects, make sure they are well
known.
- Good for projects to have information pages.
- Knowledge Base is still there, but doesn’t seem to be well
maintained or curated at the moment (updated curation is
crucial!).
- Motivation is a bit back to front - it’s the things you
don’t know which are the ones you care about.
- Could we take a Depsy
like approach about finding cross-links between
software projects?
- CMS have worked on a list of projects they are involved in,
some EU funded.
- Have regular project presentations - take inspiration from the
much loved Concurrency Forum.
- Want to address what the meaning of an HSF project actually is -
what value can HSF bring to projects that come under the
umbrella.
- Source of advice and expertise.
- Means of raising awareness.
- Do we help with funding applications? Yet to be proven, but
surely does no harm.
- Discussion on this foreseen in Software Development session.
-
Monday
- Opening Plenary
- Not much change - Taylor Childers will contribute to the US
funding situation talk
- Technology Watch
- Call for input broadcast to the community.
- HEP Use Cases
- Need to go through the list to find speakers. Liz/Daniel
will help follow up with Erica.
-
Tuesday
- Data Management/Lakes
- The data management sessions will have a small adjustment
(One talk of 5’, highlights for XDC). Otherwise the agenda
is confirmed.
- Frameworks and Infrastructure
- Iterating on question/themes for session speakers.
- Training (Parallel)
- Speakers invited, all have accepted: they were provided a
template to help them delivering homogeneous talks from
their individual perspective. This will facilitate the
round table: needs and proposal will be based on an equal
footing. First outcome should be a proposal on how to
organize the “Training” activity by joining forces among
existing (and future) initiatives.
- DPHEP (Parallel)
- Good agenda - not clear if all speakers are found.
- Workload Management (Parallel)
- For the workload management session all speakers are
effectively confirmed.
- Analysis Facilities and Use Cases
-
Wednesday
- Programming for Concurrency and Co-Processors
- Visualisation (Parallel)
- Performance and Cost Modeling (Parallel)
- Discussed yesterday, seems well planned.
- Software Development (Parallel)
- Asked Sandro about static analysers. Will discuss about the
profilers soon.
- Simulation (Parallel)
- Talks in good dhape. John can’t be physically in Naples (need a second person to
help in-vivo).
- Security (Parallel)
- Discussion on OpSec section on Monday this week, slide
preparation underway.
- Technical Discussion on Frameworks (Parallel)
- Some problems organising this session. Now envisage more of
a follow on discussion based on CWP and continuation of
the Frameworks and Infrastructure session.
-
Thursday
- Closing Plenary
- Need to chase up SKA talk.
CWP
-
General Matters and Roadmap
- CERN Courier article got a few improvements after Ed Board
feedback. Will appear in April edition.
- Will circulate the proof to the startup team.
- Symmetry magazine article status we are not sure about - to be
followed up.
-
Publication strategy for Individual WG Papers
-
Simulation
- Ready to go to arXiv now, with a finalised author list.
Merged into hsf-documents repo.
-
Machine Learning
-
Software Trigger and Event Reconstruction
- Done. Needs merged into hsf-documents repo.
-
Data Organisation, Management and Access
- Current contents considered in very good shape
- Plan is to convert to LaTeX in the coming days, check the
consistency with the global roadmap next week and call for
authors as soon as the content is final
-
Data and Software Preservation
-
Data Analysis and Interpretation
- Added abstract (was missing) and completed editing pass.
Will make pull request with the latest changes and then
get sign-off from authors in preparation for submission.
-
Visualization
-
Event/Data Processing Frameworks
- A version went to the github document repo. Still needs more
work.
-
Careers, Staffing and Training
- Want to incorporate thoughts from Naples and revise, given
the wider community involved there.
-
Facilities and Distributed Computing
- Waiting until after the WLCG strategy document is done.
-
Conditions Access
-
Generators
- Should discuss this at Naples and see how to move forward.
PyHEP Workshop
Activity updates
Licensing
- HepMC3 authors agreed to assign copyright to CERN and apply an LGPL
license. CERN copyright is recognised as using the lab as a
favourable copyright holder for the community. Authors
acknowledgement file makes sure contributors get appropriate
credit (“git log” tells all, of course!).
- Could HepMC3 become an HSF project? Would need to improve a few
technical points, but in principle it’s a good candidate.
(N.B. HSF is not a legal entity, can’t hold copyright.)
- Discussion on copyright - authors should decide who the appropriate
entity is to hold this (associated labs are good). Aim for simplicity is our
advice - one copyright holder is best.
Packaging
- Meeting
yesterday.
Converged on use cases document and had an interesting discussion
on Spack’s missing/awkward features.
- Next meeting 21
March. Look at
the use of Nix in LHCb.
AOB
- Agreed that getting DOIs for technical notes is a good thing to do.
Just need to give people practical advice on this (Eduardo).
- US conceptualisation on Research Software Sustainability Institute
starting. Workshop in Berkeley to kick off next month. See
http://urssi.us/. Mark Neubauer will
attend.