Hi,
I've just read most of that thread and found it rather disappointing.
I'd just like to add my 2 (or 3) cents:
a) I like to have the freedom to choose what software (under which
licenses) I'm using. Thus I'd like to see GNUTLS supported, as it adds
an additional feature to PostgreSQL per se: the option to choose between
different SSL implementations.
b) The other features of Martijn's patch got completely overseen. Can we
(can you Martijn?) break up the patch into smaller pieces and discuss
single independent features, like querying for parameters of the SSL
connection?
c) I'm disappointed at the way licenses are threated here. Being a
developer myself, I'm looking at licenses as a wish of the author about
how to treat his work and how to credit him. I'd like to follow these
wishes as good as I can, instead of stepping into the grey-area and
playing the 'hopefully-no-one-sues-us' game.
In case of the advertising clause, which is very strong, IMO, I think
most authors didn't want to be as strict as they made it sound in the
license. Or did any of the OpenSSL or libjpeg projects ever try to sue
somebody for not having mentioned them in their advertising materials?
You can ask the authors how they really meant it, probably they will
change the wording or even remove the advertising clause entirely. Or
probably they officially state how they meant their advertising clause
to be interpreted. (I'm not aware of the OpenSSL project doing so. While
the FSF states quite clearly that they don't consider such a restriction
to be respectful to their GPL.)
Following that 'better-safe-than-sorry' philosophy, one could ask if
PostgreSQL shouldn't better include the acknowledgements of OpenSSL (and
MIT Kerberos) in all of their advertising materials...
I fully understand and support Debian's point of view and I'd wish more
people would follow that spirit. We'd have much less cases to fight for
in curt and generally live in a better world (TM).
Regards
Markus