[CLUE-Talk] OSCON, final thoughts

[Tom Poindexter]

First, thanks for the notes folks have sent who have found my ramblings
useful.

A few notes on my final day (Friday)
Tcl sessions: 
  -Karl Lehenbauer presented a project that embeds a Tcl 
   interpreter into a ruggedtized handheld computer for test automation.
   The handheld runs MSDOS on a 80186, so it was quite a trip down
   memory lane for the problems one encounters working in that type of
   environment.
  -David Gravereaux present a method for embedding Tcl/Tk into existing
   Windows applications by running Tcl in a new thread and using sending
   message back and forth from another windows thread.
  -Bowers, Parker, and Tanenbaum presented a 'killer application' in Tcl.
   They work with the Army Ballistics Research Laboratory, and presented
   work on putting a Tcl/Tk front-end onto BRL-CAD and MUVES, the former
   is a CAD & 3D rendering package, the latter is modeling tool for 
   predicing damage to people and things from projectiles.  A Tcl controlled
   multiprocessing (Beowolf-style) cuts modeling time from days down to
   minutes.
Other sessions:
  -Richard Shank, CrystalSpace - a game engine for building high-frame-rate 
   first-person style games.  Unfortunately, the presenter didn't really
   get much across that isn't otherwise found on the web pages.

Overall: I will compare to the annual USENIX General Technical conference, 
which I have attended for the last several years (except this year.)

  -This was my second O'Reilly Open Source Convention (1999 and 2001).  Not
   much was very different than from the first conference.  Attendance
   was down, I heard they were expecting up to 2500, but had 1300-1500
   that actually attended, probably a sign of weaker budgets and cutbacks.
   That's still a lot of people in my book for a conference of this nature.
  -Among the keynote speakers, Moore (Wed.) had the most interesting to say.
   The 'great debate' between Mundie & Tiemann (and panelists) didn't
   really break any new ground, although Mundie seems a little less hard-
   line than earlier accounts.
  -It's pricey!  The conference fees are about twice the USENIX General
   Technical Conference ($1100), and hotel is expensive ($200/night).  
   Tutorials are priced reasonable ($345 for a single half-day tutorial, 
   $950 for four).  I didn't attend any tutorials.
   For this extra price, you do get good food, continental style breakfast
   and lunch (except for the last day), bagels/cookies/brownies 
   soda/coffee between sessions.  
  -Quality of papers and presentation are still a cut below USENIX.  USENIX
   has recently had a 'Freenix' track where there is a lot of overlap
   between these two conferences.  The big missing factor is any substantial
   number of academic papers from students and faculty.  I enjoy reading
   about research in computer science; OSCON has more 'practical' 
   presentations.  Not that that in itself is bad, but OSCON is not 
   well balanced.
  -It's always nice to network with colleagues, and make new contacts in
   BoF's. 
  -Vendor presence was down.  The exhibit hall is never really a draw for
   a conference like this, but a nice diversion to talk with some vendors
   (and to pick up on toys & freebies).  The number of exhibitors was
   definitely down, and only one vendor had late-night smooze-networking-
   fests (ActiveState).  
  -Speakers seems like they were choosen to elicit controversy (e.g. Mundie
   from Microsoft) rather than to enlighten.  Again, the USENIX conference
   generally has a keynote speaker and an entire track of invited talks
   throughout the day.
  -Organization was pitiful.  Only the Perl Conference papers were published
   into proceedings.  The Tcl/Tk, Python, Linux, etc. tracks didn't not
   get published, except for download on the web.   The conference locations
   were between two of the hotel buildings (a 7-8 minute walk) which cut
   down on the number of presentations one could reasonable attend.
  -A large number of simultaneous tracks and presentations.  I found it
   both good and bad.  A lot of interesting stuff presented, but many times
   at the same time :(


I'll probably attend the USENIX conference next year, and pass on OSCON.

Be sure to read others' accounts:
 http://slashdot.org/article.pl?sid=01/07/27/1656232&mode=thread
 http://slashdot.org/article.pl?sid=01/07/26/1823233
 http://www.zdnet.com/zdnn/stories/news/0,4586,5094814,00.html
 http://www.zdnet.com/zdnn/stories/news/0,4586,2799210,00.html?chkpt=zdnnp1tp02
 http://news.cnet.com/news/0-1003-200-6690267.html
 http://news.cnet.com/news/0-1003-201-6686198-0.html
 http://idg.net/ic_659965_1794_9-10000.html
 http://idg.net/ic_660022_1794_9-10000.html

see some videos and listen to audio at:
 http://www.technetcast.com/
 http://www.oreillynet.com/oscon2001/

-Tom
-- 
Tom Poindexter
tpoindex at nyx.net
http://www.nyx.net/~tpoindex/