HSF Weekly Meeting #145, 20 September, 2018

Activity updates

Michel sent an email to FastJet developers about re-discussing their licensing choice.

Graeme wrote his Python intro course slides (technical solution of Jupyter + RISE was very successful). Available here and should be forked to HSF-training, once a few other people got a chance to check them.
INSIGHTS - this is EU funded. We should reach out to them as training contacts.
Henry Schreiner is teaching a Data Science course, the URLs are available.
Mark Neubauer also has a course on this topic.
Dario - we need a way to gather this material to curate it (there is a meeting with Sudhir Malik, Albert Puig, Dario Berzano soon).
- Liz, Pete, Graeme: HSF website would be ideal as everyone can contribute there with a known workflow.
- Dario - not sure that the GitHub PR is the best way: past and future events should be browsable and this is possible only if there’s a database under the hood. Managing this kind of information can be best accomplished with a Content Management System (CMS :) ). This requires some work, somebody willing and able to set it up (something like a Joomla web site). Food for thought for the Training WG.
IRIS-HEP contribution? Want to build the community around this is the key point. Having a curator would be really helpful, not to take on all the work, of course they can delegate to others as well.
Conclusion - there is a lot of useful things we can achieve now in the training domain.

Group met, September 19
- Very useful meeting, focused on 2 interesting Spack talks - one covered important recent developments, one on successful developments, using containers.
- Todd Gamblin (main Spack developer) joined us
  - He has been very happy with our collaboration and wants to prioritise HEP use cases.
  - Issue of Spack relicensing (from LGPL to MIT/Apache 2) was raised - we will chase our HEP contributors.
Next meeting October 3

Graeme contacted Chris Jones (FNAL) re. the OpenMP investigations we discussed last week.

PyHEP 2018 workshop presentations: all available on Zenodo, with a DOI attributed, see https://zenodo.org/communities/pyhep2018.
Can this be automatic from Indico? Could be a feature request, especially if there are more presentations to handle.
Eduardo gave a talk to LHCb about PyHEP that was well received.
- Could repeat that for other communities.
Has this been brought up in CMS/ATLAS? Not to our knowledge. Torre - could be useful at an ATLAS S&C week (but over to Davide!). Not about convincing people to use Python, but propagates information about the community and resources available. Analysis community are probably under-informed on all that’s available.
For the work that has been done, we need to link back better to the main HSF website.

Draft mandate document being worked on.
There will be many common items for each group in terms of planning and organisation.
Concrete goals can’t be set in a mandate (right?), but each group should go through a planning/doing/reporting cycle.
Pete - most important thing is to decide what the working groups should do. Organising, not driving. We ask for deliverables?
- e.g., inventory of active projects, overlaps and gaps.
- need to make sure that this is prominent.
Daniel - should be more explicit about common solutions and identifying areas that need more effort.
Pere - can facilitate the communication.
Eduardo, e.g., could encourage LHCb people to also try the Belle II analysis facilities.
Advocating new things to try seems a positive thing, not the same a deciding for people.
Liz - in the packaging we had to deliver a document that summarises the state of the art in an area, that was helpful to guide an informed decision.
Don’t want this to be a re-iteration/re-invention of the CWP process
- Graeme defended the short summary of areas of interest, as not everyone will go back and read a whole CWP chapter.
We do want to have a process that leads to concrete outcomes; that will require some discussion with potential convenors.
Action - work on document and try to converge on the most useful description; re-ordering things is fine and look for the most important points to stress.

This was announced at the beginning of the week (hsf-forum).
- Yet to go to generator community lists.
Indico ready and registration open.
Next coordinators’ meeting on Monday, firming up the agenda.

General Matters and Roadmap
- CSBS review comments
- No progress this week.
Publication status for Individual WG Papers
- Data and Software Preservation
  - Content added to HSF documents repo; good to go into arXiv.

HSF Logo in vector format - reminder of action on Benedikt.
Next HSF/WLCG workshop:
- Amber confirmed JLab availability for 18-22 March 2019.
- Last minute clash arose with OSG All Hands meeting (which was planned for University of Oklahoma).
  - Graeme suggested actually making a 3-way workshop.
  - This went down pretty well with all parties; planning will be quite tricky but there are also a lot of potential advantages to doing this.
  - Discussed with Amber and it seems that JLab does have the capacity for this (main auditorium seats 286, there are a number of smaller rooms for parallel sessions).
  - Michel will also join the discussions for HSF.
- Correction from last week - original BNL/Stony Brook offer was for $300 per person for 239 people, which was scaled up to $350 for 200 based on a misunderstanding of how the costs were calculated (fixed vs. per person).
The lesson learned:
- Please put tentative dates into the HSF calendar as soon as is reasonable, even if they change later on! OSG people are added as calendar editors.
Comment from Graeme: should the calendar be in “Communication” instead of “Activities”?
- Agreed.
Next week’s meeting: Graeme will be in the ATLAS I/O Review, can someone else chair? Will decide on coordination list, but we should try to have a meeting.