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

Re: Should we divide Debian to usable and unusable



On Mon, Jun 05, 2000 at 10:33:11AM +0300, Sami Haahtinen <ressu@uusikaupunki.fi> was heard to say:
> On Mon, Jun 05, 2000 at 02:47:35PM +1000, Anthony Towns wrote:
> > On Sun, Jun 04, 2000 at 03:59:06PM +0300, Sami Haahtinen wrote:
> > > we have 1371 pre 1.0 packages, against 5309 packages total..
> > > (approx 25%)
> > 
> > This is an absolutely meaningless statistic. Pre-1.0 version numbers don't
> > mean the package is unusable or useless anymore than post-1.0 version
> > numbers mean a package doesn't crash every second time you use it.
> > You'll note, for example, that Apt hasn't made 1.0.0 yet.
> 
> the statistics were there to show where we are.. 
> this percentage should be much smaller, just because it makes linux
> more professional (No flames about this please!) it makes a whole lot of
> difference to the end users what is the current version number.. think
> of it.. would you use Debian if it was 0.6.. or 0.2.4? the point for
> those figures was that the suftware that is considered finished, shoudl
> be 1.0 it would be more interesting to look for software older that 4
> or 5 months that hasn't changes since and still has pre 1.0 version...

  I'd like to point out that the following software is either pre-1.0 or just
out of pre-1.0, and this is only out of what I happen to have installed:

  debconf
  apt
  mutt (1.0 recently released)
  xmms (1.0 recently released)
  glade
  quake (GPL version seems to have cranked the version number back)
  ftp/ftpd
  telnet
  grub
  gettext
  modconf
  whiptail
  doc-base
  ed
  bzip2
  libident (used by exim, our standard mailer)
  biff
  finger/fingerd
  rsh
  talk/talkd
  libssl09/openssl
  ash
  fakeroot
  gkrellm
  wmaker
  sawmill^H^H^H^Hfish
  enlightenment

  Most of these programs are fairly important to the system in one way or
another, or else very useful user-level programs.  I don't think that removing
them based on version numbers (which are, after all, entirely arbitrary) is
a particularly good idea.  All that a <1.0 version number means (usually) is
that the author still wants to add stuff.

  Daniel

-- 
/----------------- Daniel Burrows <Daniel_Burrows@brown.edu> -----------------\
|  The sigfile hits! | "You keep on using that word.  I do not think it means |
| You feel confused. |  what you think it means."                             |
|                    |   -- "The Princess Bride"                              |
\-------- Classes are first-class objects. -- http://www.python.org ----------/



Reply to: