	       INTRODUCTION TO THE SOAR-CMU ENVIRONMENT
		    Logistics you'll want to know

		   Bob Doorenbos and Virginia Peck

		  created: Mon Jan 20 10:33:12 1992
	     last modified: vapeck Tue Jan 21 21:32:37 1992

       /afs/cs/project/soar/member/vapeck/new.people/fact.sheet

-------
WHAT'S THIS?
This document lists various facts, people, guidelines, and pointers to
life in the Soar environment at CMU.  It is aimed toward newcomers to
the Soar group here but should also be useful for more peripheral
people, like grad students doing elective projects or temporary
visitors.  It can also be useful as a reference for existing users on
what to expect on other machines. It was compiled by Bob Doorenbos and
Gin Peck in Winter '91-92, with comments from Allen Newell and Frank
Ritter.

-------
WHO CAN I TALK TO?
If you've gotten this far, you've obviously been in contact with
someone already (a project leader, an advisor, etc.) in the Soar
group.  This person will tell you details on the how-do-I-learn,
what-do-I-learn, and who's-going-to-teach-me questions.  There's been
a strong interest recently on how people learn Soar; in order to
find out what the difficulties are in the learning process, you may be
asked if we can monitor your progress.

Gin Peck and Bob Doorenbos are currently volunteering as the contact
people if you have general or logistical questions.  They can point
you in the right direction, explain acronyms and Soar buzz words, and
refer you to other sources for even better answers to your questions.
(Gin Peck, vapeck@cs, 7205 Wean, x3620; Bob Doorenbos, doorenbs@cs,
8101 Wean, x3060)

-------
WHO'S THE SYSTEM PERSON-IN-THE-KNOW?
Tom McGinnis, tfm@cs.  He can help with machine problems, accounts,
and Soar bugs.  He's an all-around helpful person; it pays to
be nice to him.

-------
WHO'S MARGE?  
Marge Profeta, mp3r@cs, Allen and Bonnie's secretary.  She keeps track
of the library of Soar documents and papers, arranges mass events, and
tracks down advisors.

-------
I DON'T UNDERSTAND SOMETHING.  NOW WHAT?
In order of preference,
1) Ask a contact person
2) Ask the other one
3) Ask Tom McGinnis
4) Ask at a Hacker's Lunch
5) Mail to soar-bugs@cs


-------
HOW DO MOST PEOPLE RUN SOAR?  
There's no one typical environment for programming Soar.  Some people
run Soar in a regular shell window and edit code in a gnu-emacs.  Some
people work completely in a gnu-emacs, editing code in one buffer and
running Soar in another.  You can flip back and forth between the two
buffers when in Soar-mode.  Then there's TAQL-mode, for easy editing
of TAQL code in gnu-emacs, and Soar in X (SX) for a visual
display of Soar runs.  If you want to use any of these tools or need
help setting up a fancy environment, contact a contact person or Tom
(tfm@cs).

-------
WHERE DO MOST PEOPLE PUT THEIR SOAR CODE?
On the andrew file system, since it's the easiest place to store code
when using multiple machines.  Most members of the project have a
/afs/cs/project/soar/member/<userid> directory.

-------
DO WE HAVE PROJECT MEETINGS?
Yes, two every week.  Both are confirmed weekly by email.  

Hackers' Lunch, Tuesday, noon - 1 pm, 5120 Wean (the small conference
room next to Allen's office).  Informal, get help with coding problems, 
throw ideas around, bring lunch and questions.

Soar Lunch, Thursday, noon - 1:30 pm, 4605 Wean.  Presentations of
research, roundtable discussions, practice talks.  This is the big
meeting of the week.
		
-------
WHAT ARE THE VARIOUS MAILING LISTS? 
	soar-cmu@cs   	Soar people at CMU, plus a few others
	soar-group@cs 	Everyone, everywhere, doing Soar, and then some
	soar-doc@cs	Documentation bug reports. 
	soar-bugs@cs	Bug reports, help requests, etc.  If you're not 
                        sure where to send, send here.
	soar-taql@cs	Experienced TAQL users

-------
WHAT ARE THE SOAR WORKSHOPS?  
Every 8 months, the Soar community gets together for a 2-3 day
workshop.  Typical format includes lots of short presentations,
discussion sessions, and demos.  There's usually 60-80 people in
attendance and the site rotates between Michigan, CMU, USC-ISI,
and Ohio State.  Everyone from CMU goes and usually doesn't want to
talk about Soar for days afterwards.  The latest one is Soar 10,
February 7-9, 1992, at Michigan.  All transparencies from the
workshops are collected, copied, and handed out in bound form to the
participants.  Ask an "older" Soarer if you would like to see an
overview of what goes on.

There's also the European version called Euro-Soar, and CMU usually
sends a few representatives.  Soar is being used and developed
literally world-wide -- even in Fiji, which leads to the obvious
suggestion for where the next workshop should be.


-------
IS THERE A BIBLIOGRAPHY OF SOAR PAPERS?
Check out /afs/cs/project/soar/member/biblio/soar.bib for a list of
most Soar papers.

-------
WOW, THAT'S A LOT OF PAPERS.  IS THERE A SUGGESTED READINGS LIST FOR
GENERAL REFERENCE OR PEOPLE JUST MILDLY INTERESTED IN SOAR?
Allen compiled a list just for that purpose.  The latest version is
available on-line as /afs/cs/project/soar/member/biblio/suggested.{doc,ps}

-------
HOW CAN I FIND OUT WHAT BUGS ARE ALREADY KNOWN TO EXIST IN SOAR?
Send mail to soarhack@cs with the "Subject:" line "bug list".  You
will receive an automatic reply consisting of the current bug list,
one-line-per-bug.  There are other keywords you can use for the 
subject line as well:

 Subject Line                         Action
   Subject: help                        Returns list of subject keywords
   Subject: keywords                    Returns list of subject keywords
   Subject: bug list                    Returns the current Bug List.
   Subject: bug references <N>          Returns the mail messages cross-
					  referenced to bug number <N>,
					  where <N> is an integer.
   Subject: digest                      Returns a short-form of the Bug
					  List (one line per entry),
					  sorted in priority-order.
   Subject: priority summary            Same as "digest", above.
   Subject: bug summary                 Same as "digest", above.
   Subject: <anything else>             Is delivered to soarhack@cs.cmu.edu.

-------
HOW COME I CAN'T FIND A MACHINE TO RUN SOAR?  THEY ALL LOOK BUSY.
The Soar computing environment is unique, in that our pool of shared
machines aren't located in a lab but are instead spread out into
members' offices.  Though a machine may sit on person X's desk, it's
not that person's private machine - the cycles belong to everyone in
the group.  As a member of the Soar group, you should have accounts on
most, if not all, of the Soar machines and can run Soar jobs on any
Soar machine that's free.

BUT even with all the machines around here, it's difficult to find a
free one.  It's easier after 5pm, but sometimes people leave Soar jobs
running during the night too.  You'll want to run on either a
Decstation 3100 or 5000, NOT on an RT.  

Centro.soar is a special case.  Most people in the group read their
mail and bboards, etc. on Centro, so it's usually quite crowded.
Running Soar on Centro is usually a no-no.


-------
GUIDELINES FOR FINDING A PLACE TO RUN SOAR?

1. Guess a machine to try.  One heuristic you can use is to finger at
each machine (using /afs/cs/project/soar/bin/pmstat or the fingerps
command in /afs/cs/user/ritter/.soar-aliases) and look for one
with very few users.  Another heuristic is to try machines with large
idle times, but this is iffy, since the accuracy of idle times varies
from machine to machine.  Try the 5000s before trying the 3100s, since
they are faster.  

2a.  If you choose to try a 5000...
	 Mixo.soar.cs		Rick Lewis	  	 96 Meg
	 Sacro.soar.cs		Jill Lehman	 	 96 Meg
	 Salta.soar.cs		Thad Polk	 	 96 Meg
	 Shamo.soar.cs		Frank Ritter	 	 32 Meg 
	 Spino.soar.cs		David Steier	 	 96 Meg
	 Stego.soar.cs		Bob Doorenbos	 	128 Meg 
	 Styraco.soar.cs	Gregg Yost	 	 32 Meg 
	 Triceratops.soar.cs	Gary Pelton	 	 96 Meg

This is what the configurations will look like in the "near future",
as of 1/19/92. Although we have not done the experiments, it should be
possible to run at least two Soar jobs on a 96 Meg machine, even
conceivably three Soar jobs.  Once we find out, then running a job on
the 96 Meg machines when another Soar job is present will be okay --
BUT until then, follow the one-running-Soar-per-machine rule as listed
below.

Exceptions...Use Frank Ritter's machine (Shamo) and Gregg Yost's
machine (Styraco) as last resort, because they only have 32 Meg.  Bob
Doorenbos' machine (Stego) may be running VERY large programs and you
should never run when one of these is running.  Use the 'top' command
to check before you run Soar.

2b. If you choose to try a 3100...
	 Phyto.soar.cs		Jill Larkin		 24 Meg
	 Pisano.soar.cs		Brian Milnes		 24 Meg
	 Plateo.soar.cs		Alonso Vera		 24 Meg
	 Probacto.soar.cs	Virginia Peck		 24 Meg
	 Psittaco.soar.cs	Xiaoping Li		 24 Meg
	 Ptero.soar.cs		Thomas McGinnis		 24 Meg
         Taj.soar.cs		Dhiraj Pathak		 24 Meg

When you telnet over to a 3100, first type 'ps ua' and look for
another Soar job.  If there's no running Soar, check if someone is
working on the console (either by idle time or by checking active
processes).  Starting up Soar can really slow down the console to the
point of annoying the owner.  If the coast is clear, then go ahead and
start up Soar.

We know it's confusing.  Good luck in your quest.

-------
ANYTHING ELSE? 
Of course, lots, but you can just ask us when a question arises.  This
document subject to constant revision. 

-------
THE END
