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

Re: Yet another list statistics for debian-project

On Sun, 18 Jan 2009, Steve Langasek wrote:

Right - the statistics themselves are interesting, but the assumption that
more list posting is better is an unhealthy one.

Could you please quote where I said this.  Perhaps I described
some things to simple - but I hope I did not this kind of
oversimplification explicitely.

For instance, I was gratified to see the statistics for debian-legal,
because they support my position that the discussion there is being DoSed by
non-DDs who are trying to use it as a forum to persuade Debian that their
interpretation of the DFSG is the correct one.

Perhaps I should be more restraining if it comes to lists I do not
know.  But hey, I'm fine that the graph supports another interpretation
and so it is not bad to have the graph, right?

And in the comments on debian-qa, Andreas, you conclude that QA is important
and it would be good to see a wider base - but the debian-qa list is only a
very small segment of the overall QA activity in Debian, most of which
doesn't need centralized coordination...

At least debian-qa is "maintainer" of an increasing amount of packages
and I always wonder who is a reliable coworker (is available and avtive
over a long time period).  This conclusion can not be drawn out of the
currently avaibale list statistics on lists.debian.org.

BTW, as near as I can tell the reports only show the figures for the top-ten
all-time posters on each list.  So there are probably a number of lists
where the top poster within a single year isn't represented at all - what
you're really capturing here is "when have the most active contributors to
each mailing list been the most active", which isn't a very good proxy at
all for "what is the list activity over time".

On small volume lists my measure works perfectly to sort out the people
who showed up for a short time and than left the project, it work to see
whether people are constantly joining etc.  If you are really active in
one year you are most probably amongst the top ten and people will see
whether you are active in only one year when you came and perhaps they
know reason why somebody left.  This is interesting information for several
projects.  The list activity over time is just graphed at


I did not wanted to reinvent the wheel.  I wanted to have answers to
questions which this graph can not answer.  I explained it in my
talk at Debconf and people agreed and asked me for more graphs.  That's
why I did them and well, I don't make stats about the pro's and con's
but more and more pros hit my mailbox and only a view cons.  That's
no proof that I did everything right and I'm meanwhile convinced that
I took some missinterpretations.

Kind regards



Reply to: