[Python-Dev] Announcing the python-static-type-checking google group

[Paul Boddie]

Stefan Behnel wrote:
> Paul Boddie, 01.07.2012 02:22:
> > Is there any reason why the compiler-sig mailing list wasn't chosen as a
> > venue
>
> Even I didn't know that this list even existed. And looking at the archives
> now, it's hard to see any relevant discussion in all the spam it received
> until it apparently died away in (almost) silence a couple of years ago.

Yes, although the mailing lists for special interest groups are advertised on 
python.org, there is no longer the focus on steering discussion to those 
lists. And I see that the compiler SIG is "retired", as is the related types 
SIG:

http://www.python.org/community/sigs/retired/

I seem to remember various procedures about SIGs and their retirement, but I 
don't really recall much discussion of such things recently. Still, the 
compiler SIG matches the scope of the Google group pretty well:

http://www.python.org/community/sigs/retired/compiler-sig/

There's even a link to discussion of some tools you may be familiar with.

> > It's obviously your choice where you host discussions and who you invite,
> > and I know that the special interest group mailing lists aren't exactly
> > well advertised these days
>
> True, but many (most?) of them are simply not very well frequented, which
> reduces the interest in joining them even further. Both SIG mailing lists
> that I read only receive a mail every so many months, often enough without
> any reply. And almost all of these mails deal with questions that would
> better be discussed on python-list to leverage the substantially higher
> number of eyeballs there.

Special interest group lists were always meant to be used as focused channels 
of communication where people are actively trying to get stuff done. The 
unfortunate thing is that they aren't as well known as they were. Another 
unfortunate thing is that getting stuff done of mutual benefit is frequently 
something that takes second place to whatever other motivations and goals 
people have, for whatever reason, good or bad. Thus, traffic drops away as 
people either do other things entirely or instead promote any related work in 
other channels instead.

> I think that's the basic problem: as long as more experts are lurking on
> python-list than on the dedicated SIG-ML, it's better not to use the SIG-ML
> for discussions but to go to python-list (or maybe python-ideas or
> python-dev) straight away.

I think we really have to sort out what python-dev is for, because currently 
there's a tendency to target the list when any kind of "expert" discussion is 
required, but there are a number of people who would rather see only 
CPython-related discussion here.

Another matter is that static analysis of Python is a topic that frequently 
hits the end of the road when one cannot, by definition, analyze Python in 
its most dynamic form, and when people refuse to accept that such analysis 
has anything to do with Python in its most pure, undiluted (and most 
contrived) form.

But as I wrote, I still intend to follow the newly created group and see what 
people have to say.

Paul