savannah-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Savannah-users] new project submissions disabled


From: Kaz Kylheku
Subject: Re: [Savannah-users] new project submissions disabled
Date: Thu, 10 Nov 2011 15:02:15 -0800
User-agent: Roundcube Webmail/0.4

On Thu, 10 Nov 2011 16:46:58 GMT, address@hidden (Karl Berry)
wrote:
> what's required?
> 
> Applying the (stringent) Savannah requirements to new submissions:
> https://savannah.gnu.org/register/requirements.php

These are political requirements. Of course nobody is going to
volunteer
to spend time enforcing a bunch of nontechnical rules that allow
junk projects to filter through, but reject a good project because
someone dared to write Linux unprefixed with GNU/.

You're missing something analogous to the Wikipedia's "noteworthiness"
standard.

A project should be worth hosting.  This means that it should have
a stated goal or mission statement describing something nontrivial.

Secondly, it should be in a reasonable state of completion of its
stated goals. It should be able to produce something that users
can download and execute, right from the day the project is
created.

Example: fails on account of being trivial:

 "A program for the POSIX environment which reads the files
  named on the command line and copies them to standard output,
  or else reads standard input if no files are specified."

This should be rejected even if it is a polished project
with bug-free code and complete documentation.

Example: fails on account of incompleteness:

 "An optimizing compiler for COBOL with object extensions."

 (But the only files are COPYING, and a Yacc grammar file with
 mostly empty actions.)

With a significance standard of this type, you can reject
most project submissions in 60 seconds. "Sorry, your project
is a weak idea without even the beginnings of
a working implementation. Sourceforge is that-a-way -> ."

And about the political rules, what's the point of being so
meticulous and consuming all that time, if you're not going
to police what happens in the projects afterward?

All that time spent checking the license, etc, but the day
the project is approved, the developers can change the
content.

So why put that much effort upfront?

You need to just skim through the project for anything
glaring. After that, you can rely on third party reports:
someone spots a project with the wrong license and reports
it.

Most of the projects on Savannah are ignored so it doesn't matter
what's in their repository, until someone who notices it
is bothered by it enough to report it.

If a tree in a forest enforces an EULA against a squirrel,
does anyone care? :)


You know, the submission process could be farmed out to
the user base via a kind of social networking gadget.
New submissions appear on some blog-like webpage where
users can check the submission and moderate it.

This way it does not require a set of dedicated volunteers
in fixed roles. At least not the grunt work of passing
a pair of eyes over everything.




reply via email to

[Prev in Thread] Current Thread [Next in Thread]