[track2] should we publish tsdoc-metadata.json

[chradek]

The conversation that started it all:
#8654

It looks like tsdoc-metadata.json allows api-extractor to know what dialect of tsdoc (if any) your package speaks. This can be useful for example when you're running api-extractor on package A, which in turn depends on package B that has types. If api-extractor finds tsdoc-metadata.json in package B, it should be able to respect the TSDoc comments...at least that's how I understand it:
microsoft/tsdoc#7

Currently many of our packages don't publish this file, since they include just the rolled-up types in the package.json files field. This package is to investigate (and potentially track efforts) whether we should be publishing this file.

[ramya-rao-a]

cc @jeremymeng, @xirzec

[deyaaeldeen]

Ok it looks like that the API extractor generating tsdoc-metadata.json is now undesirable as explained here: microsoft/rushstack#1628 (comment) (the origin of the design is here: microsoft/rushstack#959). Furthermore, in the same issue you are linking to, the idea of using a human-authored tsdoc.json was born and which I already introduced using in our repo across our packages to support the @hidden tag and the API extractor tool is already working on support for it here: microsoft/rushstack#1950.

Unfortunately, we are stuck with an older version of the api extractor for now for reasons I do not remember on top of my head so I do not think we will have access to that support anytime soon. Based on this I am gonna go ahead and close this issue and I recommend to take no action at this time.