Comments from Laura Promberger from SFT ( I had invited her, but she was off site today)

 

Comments on Enric's CHEP talk

Slide 7: pre-configured stacks - if people want container images they should add it to the cvmfs container wishlist and get it from --> /cvmfs/unpacked.cern.ch Slide 8: - as all the experiments are moving to tokens you will need to have token support if you want users to use their experiment data

Slide 9: - i know for environment with containers, interactive sites are using prewarmed containers so that they are quickly available for users --> maybe the same is possible for dask workers?

 The presentation is missing how to get start wiht the analysis facility. links to tutorial? links to the portal? etc. should be on every presentation you are giving!

 

How to get new users?

 # Lower entry barrier - have 3-5 jupyter notebook of the most common usages (physics analysis, RDataFrame vs Coffea... ) the user can clone and get started, and continue modifying to put his solution on top - have jupyter notebook that shows how to scale it out to lxbatch, grid, whatever..

# Make SWAN more known - Provide hands-on tutorials ( i think that is the most effective thing )

-->>> more or less you need to find out how to contact the right target audience and then make it super easy to join/use 

 

SWAN galleries:

- Accessible at: https://swan-gallery.web.cern.ch

- To add a repository, PR to: https://github.com/swan-cern/gallery