Open Source to require documentation for the first time for AI?

Hello @samj and @shujisado ,

Although not required initially, what I’ve been observed is that poorly described and documented projects tend to be abandoned unless a strong volunteering effort (or enterprise sponsorship) is done to correct this.

CHAOSS metrics have established two measures:

Several OS governance projects also address this:

So, although not a requirement of Open Source on its own, it has became a staple of well managed projects.

In the particular case of Ai systems, without a good description on why a particular model and dataset was chosen you are left with very little reasons to adopt it.


As for the other matters, I’ll refer back to another discussion: The ZOOOM approach on Asset Openess - #3 by gvlx