SourceCred

Cred vs. Credibility

Is SourceCred looking to measure all-time cred (i.e. an objective measurement of all contributors/contributions), or the current credibility of active members? It seems that implementing time-based cred (the most commonly asked for feature), would be more measuring the latter.

Cred, as it’s currently implemented as the weights that slosh around the pagerank graph, is conferring ‘credit for units of work’, I.e. not as conferring ‘credibility’ in every context, but def signaling credibility in a given context. A user node’s cred after combining all their work should convey a sense of credibility. One of the baseline definitions of cred is as a ‘reputation’ metric, and it’s designed to be a social technology, so I think as cred evolves, ‘credibility’ will become a more prominent feature, and as you say, time-based cred will go a long way towards putting the ‘credibility’ in cred.

@noman I’d like to hear more about the difference between “all-time cred” and “current credibility”. What are some examples of when they’re very different? I think @brianlitwin brings up a good point that credibility is contextual, e.g. your credibility as a researcher and your credibility as an implementer might be different.

I guess an example here might be: suppose we are looking at cred in physics. Isaac Newton might have a very high all-time cred and a low current credibility.

SourceCred should be able to output an answer for both, it depends on how you parameterize the algorithm. Eventually I’d like to be able to answer very locally scoped questions like, “who has cred in the Graph implementation?” or “who has overall cred for emotional labor and community support?”, as well as questions like “who has cred in the past month” and “who has cred if we time decay old contributions” and “who has cred across all time”.

1 Like