One follow up thought I had on this came from today’s office hours where we were talking about using an additional parameter to boost early contributors, at least in terms CredSperiment dollar payout.
and later saw this
In a sense, the way this is described is a retrospective evaluation.
Partially thanks to the math teacher’s teaching, great math-related work was produced much later.
Since that later work is valued, it reflects positively on the teacher.
That situation feels very similar to this commit and perhaps the value of early contributors in general eluded to in the office hours. Possibly I’m looking at this problem wrong and such a retrospective boosting trait is what I’m looking for.
The best thing would be to get the actual data from the process of writing the book. E.g. if the book had been written using Git, we would have the commits.
I’m assuming it wasn’t written in Git, so we’ll need to create manual nodes to represent the creation process. With the manual plugin, we can still do a pretty good job with specific cred attribution.
If it were written in a text editor with distinct drafts, and the drafts went out to individuals for feedback, then we could make a node for each draft, and connect each draft to its reviewers. Each draft would also flow cred to the previous draft. Then we could also publish the drafts, and future cred-historians can tweak the weights on the individual drafts depending on how much new value each draft added.
However, the manual plugin doesn’t exist yet . So for now you could consider having adamhjk make an issue in the repo which is like “placeholder for thanking the SFOSC book reviewers” and contains references to them. Then you could put a really big weight on that initial commit (to flow cred to adamhjk) and a big weight on the reviewers (to flow cred to the reviewers). Later we would remove those weights once we had the manual plugin up and running.
The goal of the mana / boost mechanisms in the CredSperiment is to incentivize “curators” to do this work of improving the graph and finding underappreciated contributions. I have more to write on this subject.