[nSLUG] Jigdo File

Peter Cordes peter at llama.nslug.ns.ca
Wed Mar 5 17:32:57 AST 2003


On Wed, Mar 05, 2003 at 04:12:30PM -0400, Donald Teed wrote:
> On Wed, 5 Mar 2003, Ben Armstrong wrote:
> 
> > There is an alarming trend in Debian that if a package exists and it looks
> > cool, we must package it.  This gives users a bewildering choice of broken
> > pieces of software to install.  With projects like Debian Jr. I'm fighting
> > that trend, trying to zero in on a subset of everything Debian offers that
> > is useful (and solid) for a particular user base.
> 
> I'm glad a developer in the Debian community perceives this.
> More power to you...
> 
> > "Private meaning" in that you assert what 1.0 means, and then assume that
> > everyone else takes it to mean that too.  There's a good reason companies
> > don't sell software less than 1.0, and that is marketing.  They want the
> 
> Commercial software development has been around for a long time and
> has established traditions for what version numbers mean.
> The way Linux has developed a system with odd and even point kernel versions
> is a cool enhancement on these traditions.  Putting out software
> with versions *far* less than 1.0 and saying it is "useful" or maybe
> even robust, but please check the mailing lists, bugzilla, and google
> and everyone you know first, is not a practical expectation.

 I really think you're overstating the problem with version numbers.  I
agree there is a problem, but if you step back for a second, you might agree
with me that the actual problem is how to determine the quality/stability of
a project.  Some project web pages tell you which version is stable, and
which is beta, and some have CVS snapshots.  A site like that pretty much
answers the question.  (As long as their opinion of "stable" is reasonable!
That can also be a problem).  It might be nice if you could just go by the
version number, but you can't unless you know something about the project
and how they number releases.  I totally agree that having to dig up bug
reports and mailing list messages is excessive, but I think that putting
decent info on the main web page is good enough.  Maybe if there was a more
objective way to measure software quality, consistent version numbering
might make sense.  As it is now, you can only approximately figure out how
solid some software is, so even if you required people to number versions
according to some sort of standard quality index, it would still be a matter
of opinion exactly what quality number you would have.  (You'd still need
separate version and quality numbers, because quality goes up and down as
features are added and debugged.  I guess if you just measure functionality
ignoring non-fundamental bugs, you'd be more likely to see a steady
progression.)


 As for problems with jigdo, remember:
 Free software is only free if you don't value your time
 -jwz
 
 I guess I'm a Debian apologist, but it seems to me that if you want a CD,
you can either figure out what's wrong with jigdo and fix it, or pay for a
CD.  If you need a CD right away, then I don't know, I guess you're out of
luck.  

 I do get your point that this is not just you bitching, you're trying to
suggest that Debian would be better as a whole if jigdo was better.
I think you're right, and so do the people who are working on it.

 I don't know all the details behind the decision to move to jigdo, but
presumably mirror sites were unwilling to carry ISOs, given the rapidly
increasing release size.  I don't know if there was a choice between a beta
jigdo and nothing, or what, but that might be part of the story.

 It's very easy for Debian users to forget that the installer process
matters, since hardly anyone ever re-installs Debian, and the whole
mentality is built around apt repositories, not CD images.  That is
certainly part of the problem, so it's good that you bring it up.  The
debian-installer package is chronically under-loved, and often holds up the
release.  (To their credit, Debian doesn't go ahead with a release until the
installer does actually work right.)

-- 
#define X(x,y) x##y
Peter Cordes ;  e-mail: X(peter at llama.nslug. , ns.ca)

"The gods confound the man who first found out how to distinguish the hours!
 Confound him, too, who in this place set up a sundial, to cut and hack
 my day so wretchedly into small pieces!" -- Plautus, 200 BC



More information about the nSLUG mailing list