[SGVLUG] Thu 11/14 Meeting Recap

[Lan Dang]

tl;dr, Another great speaker from JPL, talking about why climate science is important and how open source tools helped the effort, with nice use of IRC and IRC bots.  Slides available later.  SGVLUG BBQ this Saturday; details at sgvlug.org.   SGVLUG e-mail is being flagged as spam by some providers; check your spam folders.  Some people have offered short term solutions to our meeting room situation, but if we don't resolve it soon, we may stop having meetings period.  Junaid continues to be awesome in his efforts to help us out.  Bonus Youtube link at the end: "Bohemian Gravity"

I don't know why it seems that people are coming later and later.  Still, we had a decent turnout.  I knew most of the people there, but there were a couple of new faces.  I hope they will return.


BBQ
====

I advertised the SGVLUG BBQ and even posted it to sgvlug.org


SPAM ISSUE
==========

Also, I've been hearing that SGVLUG e-mail is being marked as spam or 
rejected by some ISPs.  I think there was a big problem with Gmail in 
particular, because we were considered spammers.  (James said something 
about the problem being an old SIG mailing list that allowed *anyone* to post, and so it was being used for spam.)   Check your spam folders for SGVLUG mail, 
and if you think you missed a message, check the SGVLUG mailing list 
archive.
http://sgvlug.net/pipermail/sgvlug/

MEETING ROOM ISSUE
===================

I've been talking to people individually about the meeting room situation, 
but I was no good at leading a discussion on the topic.  Basically, if 
we don't find a place to meet on a regular basis, that could be the end 
of our meetings.  


Junaid is trying to find us a Caltech sponsor.   He will try to get us a room for January, since we've already scheduled a 
speaker from OLPC.  Doug has offered us use of his grandfather's house 
in an emergency.  He has also suggested DropLabs, though I think it is 
too far away for most of us, especially given traffic.  


Another thought is that we could be willing to pay for a meeting venue.  One model that people have used is to collect money from attendees to cover the price of the venue.  This can be done through Meetup, Paypal, or EventBrite.

I prefer to go with the option involving the least financial burden and/or commitment, because I think it is the most sustainable option.  I prefer our functional anarchy to having to elect officers who are responsible for things.

PRESENTATION
============

James was responsible for recruiting this month's speaker, and because of that, I had to step up my game and recruit more JPL speakers than he did :)

Our speaker this month is Chris Mattmann, a man who wears many hats  and a great proponent of open source.  Chris is also a colleague, and I keep running into him because I use software he has developed. Aside from working at JPL and teaching at USC, he is also heavily involved in Apache projects, including OODT, which is a data management and processing framework that was originally developed at JPL by his group, and which is  still under active development and use at JPL.

His talk was about using open source software to build a data processing system operating on a mobile computing cluster to quickly process data taken from airborne instruments measuring snow and snow melt.  The goal was to process the raw data through the final product within 24 hours, because it was useless to their customer otherwise.  I think it was used in weekly forecasts of snow.


I'll let the press release for the Airborne Snow Observatory speak for itself, as I was more interested in the computing technology.

http://www.jpl.nasa.gov/news/news.php?release=2013-154


In any case, this is really quick turnaround of data.   Because the site was in a place with not a lot of broadband, they had to process the data onsite, instead of transferring the data back to JPL.  This required that they build a computing cluster onsite, and come up with a way to move the data from instruments to LaCie eSATA drives to a mobile computing cluster and NAS.  One of the things they did was use QR codes to tag their assets, and smartphones to scan the QR codes, so the asset could be tracked, no matter where it was going.


As is with many things done at JPL, it was a project that required them to do things that had not been done before, so they learned a lot from the project, and it will spin off more projects, as well as useful technology for future missions.  One of the highlights of the talk was when he described getting everyone to use IRC to coordinate, and then added value to IRC by using the Notifico service and an IRC bot to report things like e-mailed status, JIRA tickets, Subversion commits, and even status from the data processing system.  Their processing strategy was just to launch multiple copies of jobs, and when one of them finishes, it will kill the other copies.  They also learned to make multiple local copies of executables and data on their compute cluster, to increase performance and avoid issues with file locks.


My feeling from the audience is that the topic was interesting, but there was some material that they would have understood better if they dealt with science data and scientists on a regular basis.  This is understandable as this talk was originally a JPL technical seminar. 


DINNER
======
Dinner was again at Hamburger Hamlet.  I think there was 8 of us.  It was a small, friendly group, and we talked about movies, the NSA, and other things.

Since I mentioned this to Kevin, the physicist, at dinner tonight,  I will end with a link to "Bohemian Gravity" by Tim Blais and A Capella Science.  It is a song about string theory, to the tune of Queen's "Bohemian Rhapsody" and the music video features sock puppet Einstein and Feynman diagrams.  

http://youtu.be/2rjbtsX7twc

Cheers,
Lan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sgvlug.net/pipermail/sgvlug/attachments/20131115/abfcd242/attachment.html>