Re: Wanted: Debian package vitality index
On Sat, Feb 02, 2002 at 03:36:31PM -0400, Ben Armstrong wrote:
> On Sat, Feb 02, 2002 at 10:58:55AM +0100, Johan Walles wrote:
> > I would like to see a Freshmeat like vitality index for Debian
> > packages. Basically, it is a measurement of how actively a package is
> > being maintained. My intended use for it is that when there are
> > several packages providing the same functionality, I can try the most
> > actively maintained one first.
>
> It sounds more like a "volatility index" as you describe it. As a very
> coarse measure of how actively maintained a package is, it might be of
> some value.
This is exactly what I'm after.
> But it says absolutely nothing about the quality of a
> package which in my mind would be the number one factor that would be of
> use to the average user in choosing one package over another.
It is more like a measure of the package quality delta than an absolute
measurement of package quality. For new packages, no information about
package quality exists, so this would be a lot more informative than
nothing (as is currently available).
> Software
> that has reached maturity will score much lower on the "vitality index"
> simply because most of the bugs have been shaken out, and few new
> features need to be added.
Yes. As this is in practice kind of a "delta" measurement, that follows.
However, if a mature package comes out with a new release it will get a
high score because it has been around for long and have had lots of bugs
fixed. New packages will never get into the vitality top ten because of
those factors.
> Changes alone do not indicate that a package is
> of any better quality than another. A thoroughly rotten design may need
> constant changes to make a workable piece of software, whereas a well-
> thought-out design may need only occasional work to keep the software
> useful to its user base.
You are right. However, if a package changes frequently and gets lots of
bugs resolved, it increases the chance for me to get help when I
encounter bugs. I value that.
> The whole idea reeks of "marketing" new, unstable software.
Possibly, but that isn't what it is about. It markets old software that
gets lots of bugs fixed, has had a new release recently and is actively
(admittedly not necessarily productively) maintained. Those are the
exact factors that the index consist of. Being new lowers the vitality,
not raises it as you suggest. Being unstable or not is not a factor
either way. Even though I would like to measure it I don't know how :-(.
Cheers //Johan
--
Johan Walles WWW: http://www.student.nada.kth.se/~d92-jwa
Vultejusvägen 40 Home: +46(8)371657 / Great minds run |
168 56 BROMMA GSM: +46(70)7106277 | in great circles. /
Reply to: