End user analysis code - no reuse potential, but how should they adopt best practice.
Add some paragraphs on scales and finesse of solutions.
Examples of reuable products - xrootd, frontier; COOL and CORAL are anti-patterns (!);
Is there data preservation group (working). Address issue of code maintainability.
Software deployment should also cover laptop deployment. (Containers?)
Homebrew/conda/spack? Is spack standard enough? (not used much outside supercomputers). Do we need a different tool? Maybe too much detail. Why have we not identified a common solution yet? CMake more in requirements.
CI - hosted vs. managed
Documentation - swings. More diversity. No need to host your own server any more.
Training - youtube channels, etc. Needs professional inputs. CERN IT project? University tuition.
Licences - Giulio
Validation - Graeme
Code preservation - Michel?
Deprecation - Graeme
Software distribution - Patricia
Move it to LaTeX? Not yet! Annecy discussion.
Next meeting w/b 16th, morning slot.