Re: Credits For v19

This is an update on the Credits for V19.

I would have liked to have attached the files that have the git messages
with the already-matched names redacted out, but those files are just too
big for the mailing list. I can do that for a smaller subset of commits if
need be.

In the mean time, here's three datasets, each one is a gzipped tsv file,
all of them have the same columns (name, list_of_emails, list_of_commits).
This isn't how the data is stored, it's just the most compact
representation I could come up with that allowed a reviewer to easily
reference all the places where a given name could be found.

The items in preferred_names are names where there is some sort of direct
confirmation that the name is correct:
- it appeared along with the email address in an email they sent
- the person (or entity, more on that later) has responded to a direct
email confirmed that the name is how they like it
- someone else has asserted that the person's name should be thus

This set is mostly for the purpose of checking against the less-strict set:
confident_names.

The set confident_names includes all the the preferred_names, plus names
that appear sufficiently name like (2+ initcapped words, not just initials,
etc). Basically, I'm confident that the names are at least mostly correct.
This is probably where most of the things needing corrections will be found.

Last is the "mystery" set. It's ones which did not make the confidence
requirements stated above. In most cases I've reached out to them directly
and haven't yet gotten a response. There are currently only 8 of them, so I
got the sense that we're close to done here.

Something interesting has come out of this, and I think it should at least
be mentioned. Overall, committers are pretty good about citing
contributions with all of the details available, so when I went digging in
the discussion threads for additional info, usually none was to be found.
But when a citation is just a name, it is unclear if that name is a person,
a person who is going by a professional handle, or the tool or service that
was used to report a bug. That much isn't new, and it can usually be
cleared up with an email to the address in question. What is new, however,
is getting a response from that email that all but confirms that it is an
AI tool, and that tool has a preferred name and is proud to be in the
credits.

I have no strong opinions about the degree to which a contributor must be
carbon-based in order to make the list, but it is perhaps something the
community should decide.

Anyway, here's the data sets, happy reviewing!