Since I probably kind of triggered this discussion with the last api issue ((https://forge.typo3.org/issues/83689) I thought I might comment as well.
There are several sides to this as has already been mentioned:
- Is the api.typo3.org necessary or is browsing the api sufficiently or better done with other tools (mostly IDE)?
- Does it provide any additional advantage?
Personally, I’ve used api.typo3.org several times, I like browsing through it, reading code and jumping back and forth between code and structure. You can probably do that more or less just as well with an IDE, but it’s kind of a convenience thing. Not everyone works 100% with TYPO3 and has everything available to set it up. People who evaluate, might want to peek in the code first. Sometimes you just want to browse a specific version but not download it. Sometimes you don’t have your laptop with you or your phpstorm license key etc. etc. I’d find it annoying if Drupal has it and TYPO3 does not.
That being said: It is also a question of resources and priorities. If the resources are simply not there and the priorities of api.typo3.org is quite low, I think you should put the effort into other things. If it is not possible to fix the current problems (e.g. estimated ~390 public functions simply missing), it should be shut down.
One other thing: the people who replied here are mostly core team. Those are the people who find it least important. If and when you shut it down, you will find, there are suddenly other people who do miss it. For those, as has already been mentioned:
Redirect in a non-obtrusive way and communicate the other resources (github, using IDE).
P.S. For the issue I reported, I pretty much tested this extensively, however only locally using standard doxygen configuration so I am not sure my results are identical, but anyway I could reproduce the missing public functions and fix it with 1 line of change in the doxygen configuration (but we all know it is usually not that easy and new problems might pop up).