Not sure how relevant this will be for folks, but I found it an interesting read from Justin Gage.
AI-generated summary
A user yanokwa shared an article by Justin Gage about migrating a community from Spectrum to Discourse. However, pfaffman disagreed with Gage’s approach of using the API for data import, citing issues such as rate limiting and difficulties with importing certain data. pfaffman suggests that using existing import scripts would have yielded better results. yanokwa wonders if the community could better document the API’s limitations, but pfaffman notes that there are existing resources, including howto documents, that advise against using the API for imports.
It is an interesting read. As someone who does a score of uploads a year, I have to say that it’s a really bad idea to try to import data with the API. If they had used an existing import script and modified it they would have gotten a much better result.
Those things are all pretty easy or non-issues with an install script.
Another reason that trying to import data en masse with the API is a bad idea.
Most people maybe won’t know these best practices? I wonder if we could do a better job documenting what the API is good for and what’s not.
There are howto documents to import from most major forums. They aren’t hard to find. I’ve helped a bunch of people debug and modify them here. I remember several topics telling people not to use the API to do an import.
There is a long list of things that the API isn’t good for. I think it might be hard to think of all of them.