Acceptance

[at88mph]

No description provided.

[msdemlei]

I think this would about do the job, and we can probably ask the Registry chairs to perform the test.

Right now, there are still many publishing registries missing. I think we should migrate their records manually, as getting them all to pass the enrollment validation is probably too much effort.

What I think we should split up between the two of us: run the validation, comparing old and new on a few "interesting" publishing registries (EURO-VO, IRSA, CSIRO, CADC, STScI, Chandra, APPLAUSE, WFAU, GAVO, CEFCA, JVO, NOIRLab; I think that about covers the various implementations, except for VizieR, which I don't want to touch for this purpose). There will be some differences, but at least we should understand what they are.

I considered building some broken entries, but... Ah well, I think that would be overdoing things.

[tsdower]

I've looked a little bit at the code, and installed/run it from the readme, and run the acceptance tests against the new site. This looks really good!

I was happily surprised to see the validation code does call the still-http-only STScI old directory interface at all. The results suggest I do have a couple of records that need fixing on my end (mostly they are prototype HiPS/HATS things). Once I get a chance to fix that I can try to register it, and we can find out if the issue with the ST registry not showing up as a publishing registry in the listing persists from the old RofR, and what to do about it.

[at88mph]

@tsdower I'm sorry, I hadn't realized there was an issue in the old RofR. I can happily help diagnose that. We have seen recently issues relating to IVO IDs not matching the case exactly, which may be a part of this. It is likely that the new RofR will be a bit better at reporting what went wrong in any case. I look forward to seeing the results anyway!

[tsdower]

@at88mph No worries, as I recall I broke something with an update to the stsci registry Identify record, then got it back up to the point where the registry is listed at all, and given the state of the old codebase, we left it there. There is another small update to the record in the works, so I'll reach out then and see what we can do!