Re: daitch_mokotoff module

Tomas Vondra <tomas(dot)vondra(at)enterprisedb(dot)com> writes:

> On 2/8/23 15:31, Dag Lem wrote:
>> Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> writes:
>>
>>> On 2023-Jan-17, Dag Lem wrote:
>>>
>>>> + * Daitch-Mokotoff Soundex
>>>> + *
>>>> + * Copyright (c) 2021 Finance Norway
>>>> + * Author: Dag Lem <dag(at)nimrod(dot)no>
>>>
>>> Hmm, I don't think we accept copyright lines that aren't "PostgreSQL
>>> Global Development Group". Is it okay to use that, and update the year
>>> to 2023? (Note that answering "no" very likely means your patch is not
>>> candidate for inclusion.) Also, we tend not to have "Author:" lines.
>>>
>>
>> You'll have to forgive me for not knowing about this rule:
>>
>> grep -ER "Copyright.*[0-9]{4}" contrib/ | grep -v PostgreSQL
>>
>> In any case, I have checked with the copyright owner, and it would be OK
>> to assign the copyright to "PostgreSQL Global Development Group".
>>
>
> I'm not entirely sure what's the rule either, and I'm a committer. My
> guess is these cases are either old and/or adding a code that already
> existed elsewhere (like some of the double metaphone, for example), or
> maybe both. But I'd bet we'd prefer not adding more ...
>
>> To avoid going back and forth with patches, how do you propose that the
>> sponsor and the author of the contributed module should be credited?
>> Woule something like this be acceptable?
>>
>
> We generally credit contributors in two ways - by mentioning them in the
> commit message, and by listing them in the release notes (for individual
> features).
>

I'll ask again, would the proposed credits be acceptable? In this case,
the code already existed elsewhere (as in your example for double
metaphone) as a separate extension. The copyright owner is OK with
copyright assignment, however I find it quite unreasonable that proper
credits should not be given. Neither commit messages nor release notes
follow the contributed module, which is in its entirety contributed by
an external entity.

I'll also point out that in addition to credits in code all over the
place, PostgreSQL has much more prominent credits in the documentation:

grep -ER "Author" doc/ | grep -v PostgreSQL

"Author" is even documented as a top level section in the Reference
Pages as "Author (only used in the contrib section)", see

https://www.postgresql.org/docs/15/docguide-style.html#id-1.11.11.8.2

If there really exists some new rule which says that for new
contributions under contrib/, credits should not be allowed in any way
in either code or documentation (IANAL, but AFAIU this would be in
conflict with laws on author's moral rights in several countries), then
one would reasonably expect that you'd be upfront about this, both in
documentation, and also as the very first thing when a contribution is
first proposed for inclusion.

Best regards

Dag Lem