This is an idea that came out of the core team public meetings at the TCL 2003 conference.
The gist of it is that it's difficult to put on "small" conferences for both logistic and financial reasons.
Clif Flynt did a fine job with TCL 2003, but his life would have been a lot easier if there had been some overseeing body such as
USENIX which could carry the logistic load. We bounced around various ideas and the limitations of the current tech marketplace for a while, and the end result is that I volunteered to try something.
Assuming I can get active co-operation from representative folks from at least two other
scripting languages (
python and
ruby are my first targets), I will organize a three-headed proposal and take it to
USENIX to see if they'll fund it. The general idea is to have a scripting languages conference that does for tcl, python and ruby what used to be done with individual small conferences. This has been done before by
USENIX with their 'little languages' conference, so it's not unprecedented. The goal is to get all three conferences going simultaneously, thereby allowing a single logistics staff to cover the conference with the same expenses as for a single conference.
The stake-in-the-ground idea is to get one conf chair for each language. That conference chair would do all the things they'd normally do with respect to gathering papers, getting them reviewed (or not), finding special speakers, etc, etc (this is the USENIX conference model). The logistics group would worry about registration, in-hotel networking, food, room rental, etc. That lets the tcl/python/ruby people focus on what they're best at, while
USENIX focuses on what they're best at.
One thing that would separate this proposal from the earlier scripting conference is the deliberate inclusion of some cross-language
GUI toolkit issues.
Python,
ruby,
perl and
tcl all use
TK, and there's been occasional talk over the years about going to a common source/toolkit/whatever for the GUI widgets for these different scripting languages.
Gtk is another obvious case. This conference would provide some neutral ground for workers and users of those kits to meet, share problems and solutions, and leverage off each others' work, and (in the best of all worlds) come to a common, well-supported GUI code base which is independent of each of the languages.
As per some discussion with
Mark Roseman I'll suggest that each language core group pick their own line chair by whatever method they feel appropriate. Then the line chairs should team up and pick an overall conference chair. It could be one of them, it could be someone else. That'll form the group, and we can then take the proposal to
USENIX.
I've already kicked off the task of finding some representative python folks to discuss this with, and am looking for ruby folks and other possible languages. If you have any suggestions, please throw them my way.
I'll track progress on this page.
January 4, 2004 Hey, I'm back. My apologies for the long delay; after 16 months of unemployment I found a job (didn't even have to move!) and promptly got buried in it. That phase seems to have come to an end, and I'm picking up some of the thinks I'd dropped.
As you can see below, it's apparent that the Python et. el. communities were already too far down the road on their 2004 conferences to do something worthwhile as a joint conference for 04. However, IMHO all the things that made a joint conference a good idea for '04 remain so for '05. I'll respond individually to a few comments below, and welcome any feedback.
August 8, 2003 18:42 EST: Clif Flynt informs me that he had an informal talk with Ellie Young, executive director of
USENIX. She said she'd be receptive to a proposal, but didn't seem to make any other commitment than that at this time. This is pretty standard for Ellie and USENIX, who are always willing to listen to ideas but don't put time and money on the table until they're sure your both serious and able to carry off your end.
August 8, 2003 19:04 EST: I've dropped a line to the Python Marketing Special Interest group (
mailto:marketing-python@wingide.com
) giving a general outline of the tentative proposal and asking for contact data for recent chairs or committee folks for python conferences.
JH: As noted at Tcl'2003, one of the overall tie-ins for such a conference that fits with the general dev focus for the short term is
Tk, which is used across the languages.
Steve Simmons: Absolutely, and I've folded that into the proposal above.
August 8, 2003 21:46 EST: I've found the Ruby core group (
mailto:ruby-core@ruby-lang.org) and sent them mail about the proposal. The proposal content at the top of this page has been updated as per the suggestions from
Mark Roseman and
JH. Many thanks!
davidw - don't know how much you followed it, but these guys (Larry, Guido, the ruby guy, lots of others) were all at the O'Reilly conference:
http://www.artima.com/weblogs/viewpost.jsp?thread=7589My guess is that you may have to demonstrate why a new conference is more valuable/worth going to.
scs response 1/4/04 - There's a difference between tcl developers attending a conf and the conference having a focus on tcl. The feedback I got from folks who attended the recent O'Reilly Open Source conferences pretty consistently said that perl was the front runner, with python a distanct second and nobody in third place. The conference we've proposed here doesn't specificly exclude perl, but does attempt to put no one language before the other.
Guido van Rossum - PyCon 2004 was just announced, we made a downpayment on the venue, and I don't think the venue can hold an additional several hundred of Tcl + Ruby developers. But you should talk to Steve Holden who is the PyCon 2004 conference chair. I like the idea of discussing the future of Tk with you all; but maybe we can start a mailing list for that?
escargo 15 Aug 2003 - Not that I expect to be coming, but it would be very speculative to think that a joint conference might bring "an additional several hundred" people. Certainly, one could always limit the conference to what the capacity of the venue is, first registered, first served. The question then remains, if several hundred additional people are too many, how many more
could the venue hold? (And where and when is it?) (I hope these are respectful questions, or at least questions that are asked in a respectful manner.)
Joe Mistachkin I would like to propose Portland, Oregon as the venue for 2004, details to be forthcoming.
Michael Jacobson First I would like to say that I really enjoyed Ann Arbor area (good food and very nice people) but I think that this conference was so small due to the choice of this location.
Therefore I propose that we should choose the next location based on where people
want to visit, where the
cost is not to high and has lots of direct
flights to the city (no 30 mile drive to airport). To me
Las Vegas (NV) meets all of these requirements including having lots of space for a conferences. In addition, if we wanted even more people we could make this a Tk Conference with separate language tracks (
Tcl,
Python,
Ruby, ...) to draw lots of attention to the updated Tk.
Tom Krehbiel seconds
Las Vegas because it is with in driving distance of Phoenix and his wife would enjoy the vacation ;-)
CL has an intense ideologic aversion to Las Vegas, but recognizes his views on venues are not mainstream.
RLH Well I would agree with you CL.
