From martin@loewis.home.cs.tu-berlin.de Thu Apr 26 20:23:25 2001 From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis) Date: Thu, 26 Apr 2001 21:23:25 +0200 Subject: [meta-sig] Retired SIGS, SIG ownership Message-ID: <200104261923.f3QJNPd01372@mira.informatik.hu-berlin.de> Looking at http://www.python.org/sigs/ I see loads of expired SIGs. In fact, except for the never-expiring SIGS, and the catalog SIG, all SIGs are expired. I propose to revive the following SIGs, with their old mission statement. The selection is based on the criteria that they had mailing list traffic on each month of 2001: db-sig distutils-sig do-sig doc-sig edu-sig image-sig pythonmac-sig xml-sig These SIGs should be extended by atleast 2 years (giving an expiration date of June 2002 for most SIGs) The following SIGs did not meat this criterion, but should probably be revived if since I expect that people will continue to work on it, and discuss this topic now and then: i18n-sig import-sig types-sig I propose the to expire the following SIGs, since they either have achieved their mission (thread-sig), or failed to do so with no visible progress: c++-sig compiler-sig plot-sig thread-sig In addition, it appears that the meta-sig owner is incorrectly listed; what happened to the fame that Guido promised to anybody taking over ?-) Regards, Martin From guido@digicool.com Fri Apr 27 18:21:16 2001 From: guido@digicool.com (Guido van Rossum) Date: Fri, 27 Apr 2001 12:21:16 -0500 Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: Your message of "Thu, 26 Apr 2001 21:23:25 +0200." <200104261923.f3QJNPd01372@mira.informatik.hu-berlin.de> References: <200104261923.f3QJNPd01372@mira.informatik.hu-berlin.de> Message-ID: <200104271721.MAA21468@cj20424-a.reston1.va.home.com> > Looking at > > http://www.python.org/sigs/ > > I see loads of expired SIGs. In fact, except for the never-expiring > SIGS, and the catalog SIG, all SIGs are expired. > > I propose to revive the following SIGs, with their old mission > statement. The selection is based on the criteria that they had > mailing list traffic on each month of 2001: > > db-sig > distutils-sig > do-sig > doc-sig > edu-sig > image-sig > pythonmac-sig > xml-sig > > These SIGs should be extended by atleast 2 years (giving an expiration > date of June 2002 for most SIGs) Done. > The following SIGs did not meat this criterion, but should probably be > revived if since I expect that people will continue to work on it, and > discuss this topic now and then: > > i18n-sig > import-sig > types-sig I've extended these through December 2001. > I propose the to expire the following SIGs, since they either have > achieved their mission (thread-sig), or failed to do so with no > visible progress: > > c++-sig > compiler-sig > plot-sig > thread-sig Agreed. I'll post to each of these sigs about their imminent demise. I expect that Geoff Furnish will yet again complain about plans to shut down the C++ sig. Possibly Jeremy Hylton has plans for the compiler-sig. > In addition, it appears that the meta-sig owner is incorrectly > listed; what happened to the fame that Guido promised to anybody > taking over ?-) Oops. Done. --Guido van Rossum (home page: http://www.python.org/~guido/) From jeremy@digicool.com Fri Apr 27 17:35:18 2001 From: jeremy@digicool.com (Jeremy Hylton) Date: Fri, 27 Apr 2001 12:35:18 -0400 (EDT) Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: <200104271721.MAA21468@cj20424-a.reston1.va.home.com> References: <200104261923.f3QJNPd01372@mira.informatik.hu-berlin.de> <200104271721.MAA21468@cj20424-a.reston1.va.home.com> Message-ID: <15081.40902.275316.739803@slothrop.digicool.com> No problem closing down the compiler-sig as it hasn't been active, except two general questions come to mind: 1) Why do we have to shutdown the compiler-sig list, which many people are subscribed to and which has periodic traffic? The general trend these days has been to take specialized discussions off python-dev and move them to mailing lists. If the compiler-sig goes away, the rare traffic on it will probably move back to python-dev -- or we'll have to make a new mailing list somewhere else for the same purpose. (In which cases there's a lot of wasted effort to move the list from python.org to somewhere else.) 2) What's the point of having sigs? In the absence of a maintainable Python Web site, I can't tell the difference between a SIG and a mailing list. We seem to create mailing lists without any meta-sig process, e.g. iterators, sets, crypto. Does the existence of SIGs make a difference to anyone? Jeremy From fdrake@acm.org Fri Apr 27 18:05:23 2001 From: fdrake@acm.org (Fred L. Drake, Jr.) Date: Fri, 27 Apr 2001 13:05:23 -0400 (EDT) Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: <15081.40902.275316.739803@slothrop.digicool.com> References: <200104261923.f3QJNPd01372@mira.informatik.hu-berlin.de> <200104271721.MAA21468@cj20424-a.reston1.va.home.com> <15081.40902.275316.739803@slothrop.digicool.com> Message-ID: <15081.42707.622102.711071@cj42289-a.reston1.va.home.com> Jeremy Hylton writes: > 2) What's the point of having sigs? In the absence of a maintainable > Python Web site, I can't tell the difference between a SIG and a > mailing list. We seem to create mailing lists without any meta-sig > process, e.g. iterators, sets, crypto. Does the existence of SIGs > make a difference to anyone? I think of SIGs as pretty much being a mailing list as well. The biggest advantage of Python SIGs is that there's a list so people can discover what lists are available. A maintainable Web site would be a big boost, however. Perhaps SIGs should be able to get a wiki (at the champion's/participants' discretion) for presenting information to non-members? That would be relatively easy to set up and shouldn't require much webmaster attention. How useful it would be would depend entirely on the SIG and SIG members responsible for managing it. -Fred -- Fred L. Drake, Jr. PythonLabs at Digital Creations From deirdre@deirdre.net Fri Apr 27 17:32:04 2001 From: deirdre@deirdre.net (Deirdre Saoirse Moen) Date: Fri, 27 Apr 2001 09:32:04 -0700 Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: <200104271721.MAA21468@cj20424-a.reston1.va.home.com> References: <200104261923.f3QJNPd01372@mira.informatik.hu-berlin.de> <200104271721.MAA21468@cj20424-a.reston1.va.home.com> Message-ID: > > I propose the to expire the following SIGs, since they either have >> achieved their mission (thread-sig), or failed to do so with no >> visible progress: >> >> c++-sig >> compiler-sig >> plot-sig >> thread-sig > >Agreed. I'll post to each of these sigs about their imminent demise. >I expect that Geoff Furnish will yet again complain about plans to >shut down the C++ sig. Possibly Jeremy Hylton has plans for the >compiler-sig. One question might be what to do about the off-site SIG (at least the one I know about): PyObjC, which seems to be far more alive than it's been in years, quite probably because of the shipping of MacOS X. There isn't, afaik, a link from python.org, but perhaps there should be a link to: http://sourceforge.net/projects/pyobjc -- -- _Deirdre Stash-o-Matic: http://weirdre.com http://deirdre.net "I love deadlines. I like the whooshing sound they make as they fly by." - Douglas Adams From deirdre@deirdre.net Fri Apr 27 18:17:27 2001 From: deirdre@deirdre.net (Deirdre Saoirse Moen) Date: Fri, 27 Apr 2001 10:17:27 -0700 Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: <15081.42707.622102.711071@cj42289-a.reston1.va.home.com> References: <200104261923.f3QJNPd01372@mira.informatik.hu-berlin.de> <200104271721.MAA21468@cj20424-a.reston1.va.home.com> <15081.40902.275316.739803@slothrop.digicool.com> <15081.42707.622102.711071@cj42289-a.reston1.va.home.com> Message-ID: > A maintainable Web site would be a big boost, however. Perhaps SIGs >should be able to get a wiki (at the champion's/participants' >discretion) for presenting information to non-members? That would be >relatively easy to set up and shouldn't require much webmaster >attention. How useful it would be would depend entirely on the SIG >and SIG members responsible for managing it. The way it's done on other sites I've seen is via CVS, where someone can propose a change by committing it and the web site is re-generated from CVS periodically. No need for a wiki and changes could be backed out. In the long run, probably less maintenance. Each sig could be a separate branch. -- -- _Deirdre Stash-o-Matic: http://weirdre.com http://deirdre.net "I love deadlines. I like the whooshing sound they make as they fly by." - Douglas Adams From martin@loewis.home.cs.tu-berlin.de Fri Apr 27 18:47:50 2001 From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis) Date: Fri, 27 Apr 2001 19:47:50 +0200 Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: <15081.40902.275316.739803@slothrop.digicool.com> (message from Jeremy Hylton on Fri, 27 Apr 2001 12:35:18 -0400 (EDT)) References: <200104261923.f3QJNPd01372@mira.informatik.hu-berlin.de> <200104271721.MAA21468@cj20424-a.reston1.va.home.com> <15081.40902.275316.739803@slothrop.digicool.com> Message-ID: <200104271747.f3RHlow01423@mira.informatik.hu-berlin.de> > No problem closing down the compiler-sig as it hasn't been active, > except two general questions come to mind: > > 1) Why do we have to shutdown the compiler-sig list, which many people > are subscribed to and which has periodic traffic? The general > trend these days has been to take specialized discussions off > python-dev and move them to mailing lists. I guess if there is interest in keeping the mailing list, the SIG should be kept alive. However, I was really questioning whether there is interest in keeping the mailing list. I think traffic should go to python-list (i.e. the newsgroup) - but I'm not on the list, so I cannot comment whether its readership would like to continue to get rare traffic. > 2) What's the point of having sigs? In the absence of a maintainable > Python Web site, I can't tell the difference between a SIG and a > mailing list. We seem to create mailing lists without any meta-sig > process, e.g. iterators, sets, crypto. Does the existence of SIGs > make a difference to anyone? For XML-SIG, we've managed to get a maintainable web site - just not on python.org. The question sounds backwards to me, and it should read: What can we do to make the SIG pages maintainable again? If keeping them on the python.org is not suitable, perhaps SourceForge would be happy to host them? Or Digicool? Regards, Martin From guido@digicool.com Fri Apr 27 20:44:38 2001 From: guido@digicool.com (Guido van Rossum) Date: Fri, 27 Apr 2001 14:44:38 -0500 Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: Your message of "Fri, 27 Apr 2001 19:47:50 +0200." <200104271747.f3RHlow01423@mira.informatik.hu-berlin.de> References: <200104261923.f3QJNPd01372@mira.informatik.hu-berlin.de> <200104271721.MAA21468@cj20424-a.reston1.va.home.com> <15081.40902.275316.739803@slothrop.digicool.com> <200104271747.f3RHlow01423@mira.informatik.hu-berlin.de> Message-ID: <200104271944.OAA24693@cj20424-a.reston1.va.home.com> > If keeping them on the python.org is not suitable, perhaps SourceForge > would be happy to host them? Or Digicool? Please not digicool. SF sounds just about perfect! --Guido van Rossum (home page: http://www.python.org/~guido/) From martin@loewis.home.cs.tu-berlin.de Fri Apr 27 20:02:20 2001 From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis) Date: Fri, 27 Apr 2001 21:02:20 +0200 Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: <200104271944.OAA24693@cj20424-a.reston1.va.home.com> (message from Guido van Rossum on Fri, 27 Apr 2001 14:44:38 -0500) References: <200104261923.f3QJNPd01372@mira.informatik.hu-berlin.de> <200104271721.MAA21468@cj20424-a.reston1.va.home.com> <15081.40902.275316.739803@slothrop.digicool.com> <200104271747.f3RHlow01423@mira.informatik.hu-berlin.de> <200104271944.OAA24693@cj20424-a.reston1.va.home.com> Message-ID: <200104271902.f3RJ2Kq01802@mira.informatik.hu-berlin.de> > > If keeping them on the python.org is not suitable, perhaps SourceForge > > would be happy to host them? Or Digicool? > > Please not digicool. SF sounds just about perfect! In that case, I'd propose to create a CVS module for the SIG pages in the Python CVS, and give all SIG coordinators write access. Alternatively, a new project could be created; I'd call it pythonsigs. I can volunteer to create the project and copy the current pages to SF if, at the end, somebody can verify that I got them all, and redirects http://www.python.org/sigs to http://pythonsigs.sourceforge.net. We have experience with that for xml-sig, but I could use some advice how to arrange every commit in the CVS resulting in an immediate update of /home/groups/p/py/pythonsigs on SF (this is a long-standing project of mine; if it fails once more, I would need to live with a cronjob, again). There is then also the issue of the SIGs' ftp areas, but we can come to that later. Regards, Martin From tim.one@home.com Fri Apr 27 20:33:25 2001 From: tim.one@home.com (Tim Peters) Date: Fri, 27 Apr 2001 15:33:25 -0400 Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: <200104271944.OAA24693@cj20424-a.reston1.va.home.com> Message-ID: [Guido] > Please not digicool. SF sounds just about perfect! Sorry, but this seems driven by passing on the maintenance burdens to "somebody else", not by what's best for the communities served by the mailing lists. The Geocrawler archive facilities are horrid compared to python.org's, or even to egroup's/Yahoo's: + infrequently and sporadically updated + no usable thread view + no author view + no way to download individual msgs in anything other than HTML format + a Search facility that appears 100% broken ("no matches" is the only response I've ever gotten from it) + no way to download archives in bulk (this is essential for "catching up" when first signing up for a list, and, thanks to the last point, the only way to *find* anything again) Since the *real* problem is that python.org hasn't been actively maintained in a year, and there's no way to open maintenance up to the community now, the real solution lies in addressing the latter. From klm@digicool.com Fri Apr 27 20:32:52 2001 From: klm@digicool.com (Ken Manheimer) Date: Fri, 27 Apr 2001 15:32:52 -0400 (EDT) Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: <200104271902.f3RJ2Kq01802@mira.informatik.hu-berlin.de> Message-ID: On Fri, 27 Apr 2001, Martin v. Loewis wrote: > We have experience with that for xml-sig, but I could use some advice > how to arrange every commit in the CVS resulting in an immediate > update of /home/groups/p/py/pythonsigs on SF (this is a long-standing > project of mine; if it fails once more, I would need to live with a > cronjob, again). I think ssh can do the trick for you. >From the cvs 'loginfo' hook (*) on the cvs repository host, use ssh to remotely invoke some harmless command on the web host - the command doesn't matter. The trick is to use ssh in RSA authentication mode, using the '-i' option to point at an RSA identity-file private key that's *not protected by a passphrase*. (This is necessary so it can run in the background without prompting. Alternately, you may be able to put the passphrase on the commadn line - either way, it sounds unsafe, but we'll restrict control which command is invoked on the remote side, to prevent anything harmful being done with this key. On the web-host side, restrict the authorized_keys public key entry to do the real command by leading the entry with command="cvs -q up -d -P" ... (where "..." is the RSA public key). This way, the only thing people can cause when using that key is invocation of that command. Someone with access to the key file may be able to cause lots and lots of updates, but so what. (You can restrict reading of the private key file to some suitable group, if that would help... Ken klm@digicool.com (*) http://www.cvshome.org/docs/manual/cvs_18.html#SEC169 From deirdre@deirdre.net Fri Apr 27 20:40:19 2001 From: deirdre@deirdre.net (Deirdre Saoirse Moen) Date: Fri, 27 Apr 2001 12:40:19 -0700 Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: References: Message-ID: >[Guido] >> Please not digicool. SF sounds just about perfect! > >Sorry, but this seems driven by passing on the maintenance burdens to >"somebody else", not by what's best for the communities served by the mailing >lists. The Geocrawler archive facilities are horrid compared to >python.org's, or even to egroup's/Yahoo's: > >+ infrequently and sporadically updated You forgot -- Sourceforge list archives are frequently not available at 2-3 hour stretches. Ugh. Most frequently when I need them too it seems. >Since the *real* problem is that python.org hasn't been actively maintained >in a year, and there's no way to open maintenance up to the community now, >the real solution lies in addressing the latter. Agreed. -- -- _Deirdre Stash-o-Matic: http://weirdre.com http://deirdre.net "I love deadlines. I like the whooshing sound they make as they fly by." - Douglas Adams From gstein@lyra.org Fri Apr 27 21:31:15 2001 From: gstein@lyra.org (Greg Stein) Date: Fri, 27 Apr 2001 13:31:15 -0700 Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: ; from deirdre@deirdre.net on Fri, Apr 27, 2001 at 12:40:19PM -0700 References: Message-ID: <20010427133115.R1374@lyra.org> On Fri, Apr 27, 2001 at 12:40:19PM -0700, Deirdre Saoirse Moen wrote: > Tim Peters: >... > >Since the *real* problem is that python.org hasn't been actively maintained > >in a year, and there's no way to open maintenance up to the community now, > >the real solution lies in addressing the latter. > > Agreed. Absolutely agreed. The answer is not to load up more stuff on the SF site, but to make python.org the site for the community to work with. Generally, this would mean wrestling the site away from CNRI, virtual hosting it on SourceForge (they have standard facilities for this; python.sourceforge.net would == www.python.org), and letting the community keep it maintained. Repointing python.org at an SF virtual server should be a "simple matter" of asking Bob Kahn. He has said that he'd repoint the IP. Actually assuming ownership of the domain is a lot trickier, and the PSF will hopefully be a way to get that done (as part of the Python source negotiations). Cheers, -g -- Greg Stein, http://www.lyra.org/ From martin@loewis.home.cs.tu-berlin.de Fri Apr 27 22:31:35 2001 From: martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis) Date: Fri, 27 Apr 2001 23:31:35 +0200 Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: (message from Ken Manheimer on Fri, 27 Apr 2001 15:32:52 -0400 (EDT)) References: Message-ID: <200104272131.f3RLVZE02358@mira.informatik.hu-berlin.de> > I think ssh can do the trick for you. > > From the cvs 'loginfo' hook (*) on the cvs repository host, use ssh to > remotely invoke some harmless command on the web host - the command > doesn't matter. I'm aware of the loginfo hook, so I can follow that part. The sshd command= option was the real news to me. That sounds very promising; thanks for pointing it out. Regards, Martin From thomas@xs4all.net Mon Apr 30 08:25:56 2001 From: thomas@xs4all.net (Thomas Wouters) Date: Mon, 30 Apr 2001 09:25:56 +0200 Subject: [meta-sig] Retired SIGS, SIG ownership In-Reply-To: <20010427133115.R1374@lyra.org>; from gstein@lyra.org on Fri, Apr 27, 2001 at 01:31:15PM -0700 References: <20010427133115.R1374@lyra.org> Message-ID: <20010430092556.A31571@xs4all.nl> On Fri, Apr 27, 2001 at 01:31:15PM -0700, Greg Stein wrote: > > >Since the *real* problem is that python.org hasn't been actively maintained > > >in a year, and there's no way to open maintenance up to the community now, > > >the real solution lies in addressing the latter. > Absolutely agreed. > The answer is not to load up more stuff on the SF site, but to make > python.org the site for the community to work with. > Generally, this would mean wrestling the site away from CNRI, virtual > hosting it on SourceForge (they have standard facilities for this; > python.sourceforge.net would == www.python.org), and letting the community > keep it maintained. Note that this wouldn't solve the problems of downtimes/disruptions in the services python.org would want to offer, like the list archives. XS4ALL has offered to host python.org on multiple, dedicated machines (configured however we wish) behind a load-balancing, high-availability Layer-4 ethernet switch. The offer still stands, and was my boss's idea, not mine, so it is not as volatile as my interest in Python, should that be a worry. -- Thomas Wouters Hi! I'm a .signature virus! copy me into your .signature file to help me spread!