Re: Unifying ac-archive.sf.net and gnu.org

[Peter Simons]

Guido Draheim writes:

 >> (5) We leave everything as it is, and the GNU archive starts
 >> downloading the contents from the SourceForge archive as well. (I
 >> list this option for the sake of completeness. I do not want to
 >> include submissions into the GNU archive unless they come from the
 >> _author_ a.k.a. copyright holder of the macro.)

 > The wording of (5) makes me believe that you are not interested in
 > symmetrical/cooperative relationship at all.

The reading skills I learned in school suggest the wording of (5)
would make you believe that I do not want to include submissions
unless they come from the author of the macro. Downloading the macros
from a third-party site is not an option. Not to mention the fact that
two archives cross-updating each other is a maintenance nightmare.


 > To put the maintaince on two shoulders would be the best [...]. It
 > seems however that such is not an option for you [...].

If by "putting the maintenance on two shoulders" you mean that you get
commits rights to the GNU repository, then indeed, this is not an
option. I have had my share of experience with the quality of your
work on the archive, and I am reluctant to assume that everything will
be pretty and flowery from now on. You know how the saying goes:

    If you fool me once, _you_ must be ashamed.
    If you fool me twice, _I_ must be ashamed.


 >> (3) The macro repository (and I mean just the actual macros) are
 >> maintained in the GNU archive, and the SourceForge archive goes on
 >> to provide an alternate _presentation_ of them [...]

 > I do not want to limit that to just `non-gnu` or `experimental` or
 > some such, in fact I think that it should be some more status
 > levels [, ...] as examples it could be experimental, proposed,
 > semi-standard, or just alpha, beta, in-review.

You assume that there are macros that would be alright for the
SourceForge archive but not for the GNU archive. This is wrong. There
is no "quality standard" on the GNU archive that a macro must fulfill.
I include _everything_ that is submitted to me unless it is glaringly
obvious that the macro is useless. I do have neither the time nor the
competence to assess the quality of every submission. And I would be
very surprised if you did.

Thus, there is no need for an "experimental" area. There are no
"experimental" macros. Believe me, when I actually refuse a
submission, you would honestly not want it either.

As for the "status levels" that you speak of: In my opinion, the only
useful status information a macro can bear in the archive is
"peer-reviewed". But, if peer-reviewed macros are highlighted -- what
effectively makes them "better" than other macros is the eyes of the
user --, we need some standard process how a macro _is_ reviewed by
peers. We need rules, an open, public process to do it, we need a
review manager for each submission, and so on and so forth. And, to
make matters worse, we need to make sure that every submission has
chance to _be_ reviewed by peers. Meaning: we need reviewers.

Given the amount of feedback I receive today, I doubt that this will
happen. Unless, due to a staggering turn in evolution, the line "I
reviewed a macro in the GNU/SourceForge archive" is going to impress
chicks in the future.

Judging the quality of a macro in an arbitrary process is worse than
not to judge it at all. Thus, I accept _any_ submission as long as it
comes from the author.

A much better development of the events would be if the authors of the
macros received feedback from their users. But this is an different
story ...


 > I don't like to talk about (1) and (2) actually [...]

If I a got a penny for every paragraph in my postings, which you
refuse to address ... :-)

Peter