Focus for a presentation at Phuse EU

[kieranjmartin]

We intend to present at Phuse EU. This issue exists to capture any discussion content for this

[SShankar-ssh]

Hi Kieran,

1. Higlight to what extent use of Git via Gitlab/Github can simplify the QC process ( as in you dont require a QC tracker with all changes made to a program. Commit history will do); code review is made more easier as we can see in the commit history immediatly the changes over time instead of comparing versions in for example NotePad++ or LSAF.

2. Highlight what is minimally required to pass audit. Is it sufficient to run final programs for a Dry Run or final analysis from main branch? Should we create tags and run programs in the tag? In LSAF one would like to see the final programs to be signed of. But storing programs for each milestone in main branch and running the programs from the main branch would be enough.

3. Highlight the minimum of features of GitHub thats beneficial for stat programming. So do we need all advanced features? Its great to have GitHub/GitLab so there is one platform where stat programmers, data engineers and data scientist can store their code. It would make it easier to collaborate cross-functionally. This point is also relevant to have an idea about the content of the training material when onboarding stat programmers on the Git journey.

KR,

Sonakshi

[trinath-eda]

Hey Martin,

1. Talk about the branching strategy like main for submission-ready code, dev for ongoing development, and qc for independent quality control.
2. Highlight on writing meaningful commit messages that explain the 'why' behind changes—not just the 'what.'
3. Highlight about Traceability of program (as each commit acts as a documented snapshot) and Reproducibility (like reproducing an analysis by checking out the relevant commit/tag/branch and rerunning the scripts).

[SShankar-ssh]

Additionally, I have used versioning in GitHub in combination with Domino Data Lab. It would be great if we could list platforms like Domino Data Lab and give a summary of user experience so companies who are looking into statistical software to facilitate the implementation of open source & git, can use this input in their decision.