HSF Weekly Meeting #118, November 30, 2017
Present/Contributors: Dario Menasce, Michel Jouvin, Graeme Stewart, Eduardo Rodrigues, Pere Mato, Guilherme Amadio, Liz Sexton-Kennedy, John Harvey, Frank Gaede, Xiaomei Zhang, Daniel Elvira, David Lange
News, general matters
-
HSF/WLCG Workshop
- Preparation of the second email with the straw man agenda
ongoing, now uploaded as a draft to indico
(progress hindered by several key people being CWP ghost
writers).
- The large room is not available on Thursday: may impact our plan
to have a final plenary session that morning.
- Intention is to have a second announcement before the end of the year,
with registration open if possible.
CWP
-
General Matters and Editorial Board
- Draft 2 released as planned on Nov. 17. All sections present, but
Physics Generators is in fact a draft 1. Commenting period
extended to Tuesday morning CET (but no other extension
planned!).
- Last EB meeting on
Nov. 21: focused on impact assessment. See the minutes for
details. Summary of the summary:
- adding impact statements was
felt not appropriate for this document which is not a funding
proposal. It will be role of the funding proposals (e.g. S2I2)
and strategic projects (e.g.: WLCG one) to establish the
priorities out of the roadmap, relevant in the particular
context.
- Draft 2 has been read by many people (exact number difficult to
say) and commented by several tens. Generally the feedback is
good/very good. Most comments are minor rewording. Physics
Generator and ML sections are the one where the largest work
is needed (see below).
- It was found easier to accept the suggestions
progressively when there were not controversial, so draft 2
has been continuously evolving for the better. It will make the
final step for the final version easier and gave an
opportunity to get further comments on the modifications.
- Still work to be done on the conclusion to make it sharper, in
particular regarding the need to reduce the number of
solutions in the community for a specific problem.
- Signatories:
good progress, 130 signatories today (1/4 mentioned a CERN
affiliation). Difficult to assess the diversity. Will probably
need to involve the EB to ensure that there is no obvious
people missing. Torre has not mentioned his role: omission or
Atlas constraint?
- Liz reached out to the IF community.
- When do we close the ability to add your name?
- Deadline was 13 December for the 1st “final” version, but we
will be happy to accept signatures after that.
- Several presentations about CWP: Andrea Valassi at LHCb
Computing Workshop, Benedikt at a German meeting, Michel at
LCG France workshop.
-
Simulation
- CWP section: final
- arXiv: needs a final walkthrough (~1 day), to
reflect changes made to the CWP. Need to finalise author list.
-
Machine Learning
- CWP section: comments received, in particular from LHCb, that
some existing work was not enough mentioned. Meeting this
afternoon to converge.
- We will copy the new text into Draft 2 ASAP.
-
Software Development
- CWP Section: final
- arXiv: only references left to do, pending for weeks :-(
-
Software Trigger and Event Reconstruction
- CWP section: a few comments to process
-
Data Organisation, Management and Access
- CWP section: still waiting DUNE figures about raw data
throughput. Liz reported that there is still a lot debate about the data
rates internally and the physics program, so it just may not
be possible to give this in time.
-
Data and Software Preservation
-
Data Analysis and Interpretation
- CWP section: a few comments to process, including on the R&D
action definitions
-
Workflow and Resource Management
- Did not make it for the global CWP. Content verified and nothing
has been found missing (already covered in other sections).
-
Visualization
-
Event/Data Processing Frameworks
- CWP section: a few minor comments to deal with.
-
Careers, Staffing and Training
- CWP section: a significant improvement done after release of
draft 2, based on (very good) comments/suggestions by Dan Katz.
Already merged in draft 2. Almost final, except a comment
from K. Pedro saying the “section is totally inadequate” with
reasons that are in fact covered in the section… Remains a
difficult section as HSF has no authority to mandate some of
the required changes: it is not like R&D action that we can
decide to do or not.
-
Facilities and Distributed Computing
-
Conditions Access
-
Generators
- CWP section: a lot of comments received, generally good but with
some important additions/changes. Discussion with experts done
in a shadow document and finally converged.
- We will copy the new text into Draft 2 ASAP.
-
Security
-
Submission of WG papers to ArXiv
- EB agreed that it is delegated to WG convenors: an ArXiv tag
defined for each document.
Sheet
created listing the HSF report tags and tracking the
submission status.
Activity updates
Licensing
- HepMC: Graeme met with Jorgen and we discussed relicensing under
LGPL. Jorgen is now consulting with Matt.
- Belle II Licensing TF met last week: Michel exposed the LHC
experiment plans to move to GPL/LGPL v3 (including the HepMC
license change). Belle II may follow the same direction, question
remains open whether the copyrights can be transferred to a
limited number of party. Requires more discussions. Would be
interested by a common meeting with other experiments to share
problems/approaches regarding SW licensing.
- LHCb had a discussion last week on (C) (to CERN) and licence (GPLv3).
Some issues with Dutch institutes having trouble relinquishing
copyright were raised.
Training
- Graeme had some follow up discussions with the LHCb and ALICE
StarterKit teams. He proposed them contributing to the training
session at the HSF workshop in Naples.
Packaging
- Group met yesterday:
https://indico.cern.ch/event/681894/
- Further progress on use cases (almost converged). Good
presentations from ATLAS on their build, deployment and
runtime use cases.
- Next meeting in 2 weeks time, will cover CVMFS + GraphDriver and the
Portage build of the test stack.
AOB
- GSOC 2018
- We believe the process starts at the beginning of the year.
Starting discussions now to find people to coordinate.
- PyHEP
- Converged on 6 topics. In general 2 presentations each topic + a
lot of discussion time.
- ATLAS Documentation Workshop
- ATLAS are holding a workshop on documentation at their next S&C
week (11 Dec).
- “The documentation toolset (both narrative and code doc) has
been growing rapidly in the open source world. We want the
workshop to help identify promising directions and tools
to pursue, within the context of preserving adequate
stability, continuity and preservation of investment in
the existing documentation.”
- We reached out to some contacts in Belle II, Linear Collider and
SuperNemo, but we would still like to hear from anyone one
else with a good story.
- LSST might be interesting to look at (Google “LSST the Docs”).
- Next week is CMS week. Friday plenary where Liz will present the
CWP.