SPUG: New Meeting Location: Update

[Tim Maher/CONSULTIX]

>> Message submitted at: Sun Oct 17 09:57:49 PDT 1999
X-Mailer: ELM [version 2.4 PL25]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 4771      

According to William Julien:
> 
> Hmmm... The geographical center would be an interesting statistic. But, like
> Barry said, that would probably place the optimum meeting place in
> the middle of Lake Washington. 
> 
> Given that the possible meeting places are finite, I would think driving
> distance and time to be more practical approach. 
> 
> >You could also weight the average by the probability each member would
> >actually attend a meeting at the population center. For example, if
> >A would attend 10 out of 10 meetings and B would attend 5 out of 10,
> >you could multiply A's latitude by 10 and B's by 5, then add them and
> >divide by 15 (to get the weighted average latitude).
> 
> Good Idea! But I would multiply by the probability of attendance. "A" has
> a probability of 1 (10/10) and "B" has a probability of .5 (5/10). 

Sounds like you're using an assumed value of a dependent variable
(likelihood of meeting attendance) to help predict the appropriate
weighting of an independent variable  (lat/long, etc.) in predicting
a location to maximize the value of the dependent variable (attendance).

Not too kosher! 8-}

> Nomally a major consideration for me, is the length of the meeting vs
> the time it takes for me to get there. I am much more likely to go to
> an hour length meeting if it takes 5 minutes than if it take an hour.

So we should then make all meetings 4 hours, to minimize relative drive
times? 8-}
 
> So, maybe this could be used for ranking the relative worth of a meeting
> location. Something like the sum of the differences between the

Sure, William's Perl program to calculate driving distances is cool, and
as the Instructor for his 2/99 "Perl + CGI" class I'm proud to see him
using Perl creatively, and it's fun to go hog-wild imagining measurable
parameters that might relate to meeting attendance, but the conspicuous
lack of smiley emoticons in these recent posts suggest that you guys
are taking these objective parameters much too seriously!  After all,
we're trying to predict human behavior here (likelihood of attending
meetings), not the amount of fuel used by a robot to get to the meeting.

Whether or not "SPUG-lurker A" will attend a meeting at "Location B"
might be influenced to some degree by driving distance, driving time,
ratio of driving-time to meeting duration, what shows are on the car
radio at the time he/she has to start the drive, whether the optimal
route would be convenient for a stop at Burger Queen along the way, and
what adult beverages are available at the post-meeting hangout, but our
goal here is not to develop a theory for predicting meeting attendance.

Instead, our goal is to provide reasonable access to meetings for
potential attendees throughout a wide geographic area.  They can then
choose to attend or not, as they see fit.  Quality is more important that
quantity - we could have huge turnouts if we picked up all potential
attendees in a big bus at their work locations and gave them free beer
for two hours, but nobody would learn much about Perl that way! 8-}

Currently, our greatest need is not in identifying web sites that handle
additional geographic parameters of dubious predictive value, but rather
in finding new options for the Seattle locations (since we have several
apparently attractive choices for the Eastside already).

I believe our best plan is to provide at least one good choice on each
Lake-side, and possibly choices on Mercer Island and the North end,
and take votes from live members for their preferences.  Depending on
the outcome, we'll probably end up running meetings on both sides of the
lake, perhaps 3rd Tuesdays on the East and 3rd Thursdays on the West,
so everybody could attend from 0 to 2 meetings per month, as they wish.

Current List of Meeting-Place Choices:

West:
	Fred's Cancer Center
		(for $100/night - a big problem)
	SpeakEasy, Downtown Seattle
		(although some have doubts about basic suitability,
		cheaper than Fred's, and with better drinks)
East:
	Lucent, in Redmond (520 and Redmond Way)
		fancy conference room, reportedly
	Maxim Group, Bellevue (520 and ?; near Bellevue Way, I think)
		only basic conference room?  still awaiting details . . .
North:
	A Chicken-Coop in Woodinville 
		deluxe ($50/night - a problem)
	ATL-Philips Ultrasound(Advanced Technology Labs)
		522 and I405 in Bothell

*==================================================================*
| Tim Maher, PhD  CEO, Consultix &    (206) 781-UNIX/8649          |
|      Pacific Software Gurus, Inc.   Email: tim at consultix-inc.com |
| "The UNIX/Perl Training Experts"    http://www.consultix-inc.com |
|Classes: 11/1 Shell/Utils  11/15 Adv Shell  12/7 LINUX  12/13 Perl|
*==================================================================*

 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    POST TO: spug-list at pm.org        PROBLEMS: owner-spug-list at pm.org
 Seattle Perl Users Group (SPUG) Home Page: http://www.halcyon.com/spug/
 SUBSCRIBE/UNSUBSCRIBE: Replace ACTION below by subscribe or unsubscribe
        Email to majordomo at pm.org: ACTION spug-list your_address