Adding initial commit of the ASWF Development Philosophy#1356
Adding initial commit of the ASWF Development Philosophy#1356carolalynn wants to merge 4 commits into
Conversation
Signed-off-by: Carol Payne <carol.alynn.payne@gmail.com>
| @@ -0,0 +1,43 @@ | |||
| # ASWF Development Philosophy | |||
There was a problem hiding this comment.
Curious why you called this "development philosophy" when actually it's narrowly focused on how projects should interact with "AI" coding assistants or tooling. Should we title this more straightforwardly?
There was a problem hiding this comment.
Hmm. I guess I don't view it as very narrowly focused - I actually tried to keep it pretty broad - obviously the doc is motivated by AI - but the philosophies in it are not really new, nor specific to AI other than tool attribution. However, I'm happy to name it something else - maybe "AI-Assistend Development Philosophy at the ASWF"?
There was a problem hiding this comment.
What I meant, by brief example:
"Write tests before code and aim for 100% coverage", "prefer imperative programming with a few OOP abstractions," and "organize work into 4 week sprints" are examples of software development philosophy.
"You're responsible for what you submit, be respectful of maintainer time, and put an 'assisted-by' line in commit messages" are just reminders of how to interact with the group courteously and are mainly in response to the proliferation of specific new tools (even though mostly the same courtesies were assumed all along).
|
|
||
| If you’ve read this far and are thinking “wait, I still don’t know what their policy on AI generated code is” \- that’s technically correct. That is not the purpose of this document. That policy should be decided on a project level, and may change frequently \- just as technology does. While we may add more to this document over time, the basic philosophy should serve as a guiding and testing framework for policy in individual projects and contributors. However, here’s a quick summary: | ||
|
|
||
| * Humans are responsible for code contributions \- and they must be able to explain the work. |
There was a problem hiding this comment.
Might want to also call out that they attest for the work, just to speak to the fact that if it comes out of an AI tool, the contributor is the one attesting they can contribute it.
There was a problem hiding this comment.
I'm not sure I understand what change you want here - you just want me to use the word attest? Sorry haha. We definitely cover this earlier in the doc, though I don't think I specifically use that word.
Signed-off-by: Carol Payne <carol.alynn.payne@gmail.com>
3 participants
Based on the google document reviewed by the TAC.