On Mon, Mar 23, 2009 at 12:59:53PM -0400, Yaakov Nemoy wrote:
My main issues are perceived ideas of privacy on IRC. It's a
quasi-ethical one that has to do with what people expect from an IRC
channel vs. mail.
I've never encountered this perception. Private messaging directly
from nick to nick is clearly perceived as private by it's very name.
Do you have some examples you've seen where e.g. a FLOSS project had a
group of otherwise sane people who perceived a public IRC channel as
private?
Sometimes a topic in a channel specifies it might be logged, or that
it is logged and posted at a URL.
Merely to avoid rising ire in anyone, we could use IRC logs from the
past in an aggregate format that is less granular than after we put a
message in the topic that the channel is logged and analyzed.
But I wouldn't put a single stone in our pathway about this because
folks might have the mistaken impression they are having a private
discussion when they are speaking into a microphone in a room full of
people.
PISG does its analysis completely differently from EKG. Ethical
issues
aside, getting IRC logs into EKG is *very* simple. I could do a test
demo using logs i have on my shell server. My comments were about
adapting the graphs and reports that PISG generates. Since it's
written in Perl, i couldn't reuse anything there easily without just
rewriting.
If you're interested, we could set up PISG and EKG side by side for
any reason in particular, but i think alot of the general purpose
analysises i have planned for EKG would rival what PISG can already
do. I'm definitely not planning on rewriting what i have to work with
PISG code in Perl, 2 rewrites is enough. PISG can't handle mailing
lists, and EKG has the capability but not ability to handle IRC
channels.
Sure, and I see that PISG tracks individual work in a way that is less
useful to the overall health analysis of a community.
My position, though, is one seeking features and having limited
resources and means to make them happen. (Feeling a bit like grasping
for any tool since I cannot have all of them in one tool by 7 April.)
Should we prioritize the list of gathering-and-analysis modules we
want EKG to cover? From what I've seen so far, IRC analysis should be
further down the list.
If anyone has an idea of an EKG module, it should be a ticket in
Trac.[1] Then we can prioritize in the ticketing system (or on the
wiki.)
- Karsten
[1] E.g.
https://fedorahosted.org/ekg/ticket/9
--
Karsten 'quaid' Wade, Community Gardener
http://quaid.fedorapeople.org
AD0E0C41