In practice I think many requests go through automatically and are available quickly, but some require manual intervention and take a couple of weeks. At the moment StackExchange voluntarily make them available globally, but I could imagine that changing if they started losing significant users that way.
Data portability is designed precisely for this purpose: moving the data you gave one site to another site. So it seems appropriate to me.
Anyway, just an idea for improving the transition model from S.E.
Iām with @hsenag in that I think itās appropriate logically but practically SE could/would interfere with the process.
More importantly is that votes, reputation, and privileges are going to be very different on Codidact than they were on SE. From the discussions around this I donāt see a practical way to use that information meaningfully inside Codidact, which is why I agree with @cellio on the āfresh startā idea.
IMO it seems the only value is using the votes and knowing who voted is to retain the score. But Iāll note that weāre talking with TopAnswers about having a full-fledged API between them and us. In other words, we may have posts in three places - SE, TA, and Codidact. Perhaps even more in the future.
Having a way to establish the scores from other platforms will likely serve us well in the future. Iām thinking how Plex Media Server shows the ratings from IMDB, Metacritic, and RottenTomatoes, as well as your personal star-rating on the server. Or heck, even how RottenTomatoes itself distinguishes between the āaudienceā score and the ācriticā score.
So importing content but having a way to share what itās value is elsewhere is beneficial - highlight the voting score on Codidact, but list nearby what itās votes/stars are on TopAnswers and SE as well.
And since different systems will have different voting systems, this probably means we donāt try to model it but instead have a varchar for āuseful info from source siteā that we can display on our site. For SE that would be upvotes and downvotes; for TopAnswers that would be number of stars; for hypothetical other systems it might be something else.
Basically yes. Except not a varchar - actually a set of data elements - e.g.: {"source": "stackexchange.com", "sourceid": "8675309", "author": "User1234", "created": "2019-01-01 08:17:15", "updated": "2019-09-15 17:13:45", "upvotes": 3, "downvotes": 1, "HNQ-start": "2019-01-01", "HNQ-end": "2019-01-07"}
etc. with the details varying by source site and a blob of code in Codidact that extracts whatever we end up deciding is useful and displaying in an appropriate way. In other words, potentially a lot more than just votes/stars/etc. But all of which is āfinal implementation nitty gritty detailsā and not important right now. For now āblob of imported dataā to be treated as ādisplay onlyā.