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

Re: another reason why requiring NMs to be sponsored is a bad idea



On Wed, Jan 16, 2002 at 02:44:12PM -0900, Britton wrote:

> I just read in DWN that there are now 30 people waiting for sponsors.
> So now we are, essentially, turning away volunteers.

To examine two of these cases of people who can't find sponsors:

<snip>
David Parker <david@neongoat.com> (he's looking for a sponsor !)
Database Independent Abstraction Layer for C:

libdbi implements a database-independent abstraction layer in C, similar 
to the way DBI/DBD does in Perl.

Writing one generic set of code, programmers can leverage the power of 
multiple databases and multiple simultaneous database connections by 
using this framework.

Right now, I have packages for
- the library runtime
- the development files and API documentation
- the MySQL plugin
- the PostgreSQL plugin

Packages (apt-gettable) are at http://www.neongoat.com/projects/libdbi/
Webpage and docs are at http://libdbi.sourceforge.net 
</snip>

'Database Independent Abstraction Layer for C' is the most interesting 
misspelling of 'ODBC' that I've ever seen, but for all that, I don't 
think we're missing anything by not being able to get it into Debian 
more quickly.  The advantage of a database abstraction layer is to allow
a single, familiar API that people can code to for all of their database 
needs.  What, please, is the advantage of *two* database abstraction 
layers, for the same language?

Debian already has good support for ODBC (you know, the database 
abstraction that's so widespread that it's even used under Windows), 
probably better than any other Linux distro right now.  I've personally 
invested quite a bit of time into improving that support.  So while I 
can conceive of a time when libdbi might rival ODBC, I don't think it's 
a grave concern that there are delays getting it into the archive right 
now given the current state of things.

And I would certainly have sent these comments to the ITP for this 
software -- if there was any sign of an ITP...

<snip>
Javier Kohen <jkohen@users.sourceforge.net> (sponsored by Janos Lenart 
<ocsi@debian.org>)
GtkGamma is a simple program that allows the user of an XFree86 station 
to easily alter his/her monitor's gamma correction. If a monitor's gamma 
curves haven't been tuned colors half in the range from black to maximum 
tint (either red, green, blue or white) tend to look different from a 
uniform pattern formed by an equal number of black and maximum tinted 
pixels, although it should look exactly the same. GtkGamma provides a 
simple GUI that gives the user the feedback needed to correct these 
curves.

Binary (i386) package as well as a tarball (from where the binary .deb 
was built) can be found at http://gtkgamma.sourceforge.net/ . 
</snip>

I'm pleasantly surprised to see from that web page that Javier has found 
a sponsor -- I imagine this explains why he didn't respond to my offer 
of sponsorship.  It looks like he and two others have found sponsors 
since Raphael's call went out; who knows, maybe more -- there may be 
more people on the list now looking for sponsors who weren't there 
before.  So if not everyone can find sponsors, it's not because no one 
is sponsoring; it's because there's natural selection of what packages 
get sponsored in and which do not.  If it's not something that any 
current developers can make sense out of, naturally it's not going to 
get sponsored in.

> In looking at the list of people waiting the first thing I see is a guy
> who has packaged some ruby libraries.  I know nothing about ruby and so
> naturally am not inclined to sponsor those packages, since my sponsorship
> would be pretty useless.

> In fact the only people who are really worthwhile sponsors for him are
> people who know ruby, and while I'm sure there's more than one developer
> who qualifies, I doubt there are that terribly many.  What if they are all
> ~MIA (please note that I am not in any way suggesting that this is the
> case)?  I guess ruby on debian is in for a tough time.

And if there's sufficient overlap between the ruby and the Debian 
communities, don't you think this situation would sort itself out on its 
own?  As it happens, 'apt-cache search ruby' turns up 106 matches.  I'm 
sure you can find some maintainers in this list who aren't MIA.  If not, 
what exactly is the point of letting *more* ruby packages into Debian 
when there would already be so many in need of adoption?

> The current sponsorship system has serious potential to damage the
> succession process for specific areas of expertise.

> Sponsorship doesn't buy any more trust in developers than the old signing
> scheme.  I don't see what use it is.  If a new maintainer's packages are
> broken and useless, just cheerfully dump them, they are only in unstable.

What sponsorship buys is quality control.  Debian has a reputation for 
quality, and if we're interested in keeping that reputation, it stands 
to reason that every package uploaded to the archive should be inspected 
by someone we know is familiar with our standards for quality and is
committed to upholding them.  Uploading to the Debian archive is not 
a right; there's no reason to think that not being allowed to upload to 
ftp-master.d.o prevents an NM from being able to distribute .debs that 
are useful to the Debian community.  There's /plenty/ of reason to think 
that letting people who haven't been approved as Debian developers 
upload their own packages directly to Debian would be harmful to the 
quality of the distro.

There's already plenty of cruft in the archive that's used by almost no 
one.  It doesn't make sense to add to the pile.

Steve Langasek
postmodern programmer

Attachment: pgp3hKXBbawiF.pgp
Description: PGP signature


Reply to: