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

Re: ideas....

On Friday 06 April 2007 21.43:09 John Watson wrote:

> I would suggest having two releases of Debian, one "really stable" which
> could be released every 2 years, another one "stable" released every 6
> months by taking a freeze of the current "testing" distro and spending a
> month (no more) fixing any major bugs.

Many people are using testing or a mix of stable and testing for desktop use 
and have found that very good.  For bigger deployments and other more 
sensitive areas, stable is great - I don't want to fiddle around with my 
fai setup every 6 months even if it means I run Oo.org 2 while 2.1 is 
already out etc.  1.5 to 2 years is a reasonable timeframe in a commercial 
surrounding, the major problem Debian needs to solve in that area is 
support of new hardware, like SATA.

> 2)One real issue with Debian is a lack of admin tools, (such as yast is
> for SuSE). Considering starting a project to develop a range of gui based
> admin tools for Debian.

Please don't start yet another project.  There's webmin, there's the YaST 
for Debian project and I believe there were some others, too.  And then 
there is the idea of building up debconf into a full admin tool.  I set up 
a SLES10 server a few months ago, see 
<http://fortytwo.ch/blog/archives/2006/10/#e2006-10-30T14_22_18.txt> for a 
few comments.

To reiterate again here: any configuration tool that modifies configuration 
files, especially at unexpected times like from an init script (gaah!  
That's really a traumatic experience.) is evil to some degree.  A full 
config file parser that also preserves comments is much harder to do, but 
after all Debian has no marketing department saying that this config tool 
mus be finished in two weeks ...

> [money]

As the past year has proved, money is a difficult topic in a volunteer 
organisation.  Would I want Debian to accept a 10 million dollar donation?  
Probably not.  What would I do if I had 10 million dollars to spend on 
Debian?  I guess the wisest course would be to hire some people and get 
them to improve Debian so that it does exactly what I want.  Added bonus if 
they can get the improvements integrated in Debian.  If it's just 10 
millions, that would be probably 3 to 5 people, and if I could resist the 
temptation to try to force those people into prominent roles within Debian 
it might even work out in a much improved distribution without too much 
controversy, and without touching onto some fundamental values of the 
project (that people are sometimes paid for Debian work is a very old fact, 
so that part of this idea should work out fine.)

Now somebody get me that money :-)

-- vbi

featured link: http://fortytwo.ch/smtp

Attachment: pgpE676OeBJU2.pgp
Description: PGP signature

Reply to: