[SGVLUG] Meeting recap 4/14: Webification + May/June meeting announcement

[Lan Dang]

Hi all,
 

Announcement: I need a break.  The SGVLUG functional anarchy is in charge of the May/June meetings.

tl;dr Thank you Bobby and Akash and Wendell from OpenX.  We had about 56 folks show up--lot of new and returning people. Zhangfan's goal with his talk was to excite community interest in webification and he appeared to be successful. Swante gave away leftover t-shirts acquired at SCALE.  We had fun talking to each other after the event.  People went to Du-Par's afterwards.  

The webification* notes may help for people who had issues following the talk as the audio volume was variable depending on whether Zhangfan was using the microphone or facing the audience.


Lan


ANNOUNCEMENT
============
I am burnt out.  I'm taking a break from organizing the May and June SGVLUG meetings. I am relying on the functional anarchy that is SGVLUG to pick up the slack.  All I am going to do is show up on the day and enjoy hanging out with people.


The May meeting will be at Du-Par's and consist of short talks by LUG members.  The Meetup page has already been updated:

http://www.meetup.com/SGVTech/events/229463153/

The June meeting will be Peter Linnell speaking on Open Build Service.  I believe James is his contact and that Peter has to travel to come speak with us.  I am unclear how much of the logistics have been handled.  Last I heard, the venue has not been determined yet.  



LINKS
=====

* Webification slides - http://scifari.org/talk/talk-20160414.pptx

* NASA Space Apps Challenge Pasadena info - http://bit.ly/1oanT7H (writeup by Lan with links to resources and the videos from the past two kick-off meetings)

* SGVLUG website contains links to our mailing list, twitter, etc - http://sgvlug.org

* Hackaday Los Angeles has cool talks and events for makers and inventors - http://www.meetup.com/Hackaday-Los-Angeles/



ATTENDANCE
==========
I'm listing the attendance stats because it's helpful in guestimating actual turnout.

According to our sign-in sheet, there were about 58 people, of which 7 were marked as OpenX employees, who don't normally RSVP.  We had a total of 55 positive RSVPs on Meetup with about 15 no-shows. There were 2 RSVPs with me, 6 SGVLUG members who didn't RSVP, and 4 people were walk-ins.


I saw a lot of familiar faces as well as a lot of new faces. 



INTRO
=====

Despite the very last-minute promotion** for this month's SGVLUG meeting and it turning out to be just me running the meeting, the event turned out rather well.  Bless Bobby and Akash and Wendell for having everything setup and providing A/V support. 



I had OpenX start off the night.  Bobby gave his spiel for OpenX and mentioned that they are currently hiring for Site Reliability Engineers.


I had to be emcee and run Linux in the News.  Not my favorite role.  At least I gave people a very thorough introduction to SGVLUG website/twitter/mailing list and also talked about the umbrella nature of our SGVTech Meetup group.  Also announced the NASA Space Apps Challenge which is happening next weekend.  


Thank you Carlos, Jaime, and others who provided Linux in the News items, discussions, and clarifications.  


There were some great questions during the talk, and a lot of people lingered to chat with each other afterwards.  We finally got them out of the door so they could continue their conversations at Du-Par's.


WEBIFICATION
============

Zhangfan's talk was actually called "Simplify Data Use on Web Platform" and the core of it is the concept of "webification" or w10n.  It is a topic that is very close to his heart and that he has been fortunate enough to take from a hobby project to a technology that is currently being used in a variety of places at NASA.  What was obvious during his talk was that he has a lot of passion and vision for w10n, but he is too busy to pursue it. 


His goal in speaking at SGVLUG is to excite the interest of the open source community and take this concept beyond NASA, and his last slide basically says "Seeking help from enthusiasts to move to next level." 

The idea behind webification is to provide a simple API for people to access data via a semantic URL.  Webification is a specification, and while he has developed software implementations of it, people can develop their own or extend his software.  He doesn't believe there are any limits to what you can webify.  As long as you can map it to a tree structure of nodes and leaves, you should be able to webify it.  

This ties into his work with w10n-sci, or an application of w10n to scientific data, which comes in formats like images, NetCDF or HDF5***, and tend to span multiple files and take up gigabytes and terabytes of disk space.  In the past, scientists have had to mirror this data to their local servers to run analysis on it and they would have to write software to read in the data, pull out the fields they want, and reformat that into a more useful format for them to do their analysis. 

Wouldn't it be wonderful if, instead, they could just go to the provider site, make a request via URL and retrieve only the data they are interested in and have it be returned in a standard format like CSV or JSON?  You can extend this and ask, wouldn't it be even nicer if there was some kind of web widget that could interface with the data, and all they would have to do is to navigate to a URL and be able to display their data selection on a map or a lineplot?  


The above pretty much describes his demo.

  

He laid out a specification that fits on one slide, which Liz commemorated here:
https://twitter.com/LearningNerd/status/720814051148713984
There are three categories of users for webification: 

* the data providers  - to expose their data in a user-friendly way

* the developers aka the data innovators - to build the cool visualizations, web widgets and/or tools 

* the end users - who analyze or consume the data


It was a pretty personal talk as well.  We learned that he did his undergraduate work in China, his graduate work in Holland and was at Caltech for some time before eventually going to JPL.   The concept of webification is something he had developed back in 1999, during a half-year time period when he was applying for a green card, but his visa had expired, so he was not allowed to work.  He started work in scientific modelling****  His big break came when they had difficulty disseminating the huge quantities of data they were generating.  He has a background in both science and IT, so he was able  to cobble up a solution using his old hobby project of webification.


His programming background is in Java and Python and he whimsically names his software after fruit, and they start with a P if they are Python-based and a J if they are Java-based. He is a great believer in Unix philosophy.

https://en.wikipedia.org/wiki/Unix_philosophy

You should check out the slides.  They are minimalist and contain all the links necessary to get started with webification as well as his contact info, if you want to help him take it to the next level.  There are a bunch of links to NASA open datasets that support webification. He has links to a couple of public web widgets that you could use to browse and visualize this data.  Or a clever developer could build some pretty awesome visualizations or applications based on this webified data.  



FOOTNOTES
=========

* I am a big fan of w10n and its potential and have talked to Zhangfan many times about it; my summary is going to contain a few extra details that may not have come out during the talk.



** My goal with these Meetups we do at OpenX is to provide an appropriate audience for our speaker, who may be talking about a particularly niche topic.  I do very targeted promotions to the communities who may be interested in this topic.  This month, I targeted folks at JPL, the Learn to Code LA community, the folks excited about the NASA Space Apps Challenge, CS students from PCC, and the local Big Data group.  Judging by the level of questions, I was pretty successful.

*** NetCDF and HDF5 are data formats that are popular with scientists.  NetCDF is a simpler format and is very popular with the scientific modelling community.  HDF5 basically allows you to put whatever you want into the file and can be challenging for scientists to read, depending on how complicated the structure is and how well their implementation or programming language supports HDF5.


a) NetCDF is a set of software libraries and self-describing, machine-independent data formats that support the creation, access, and sharing of array-oriented scientific data.
http://www.unidata.ucar.edu/software/netcdf/

b) HDF5 is a data model, library, and file format for storing and managing data. It supports an unlimited variety of datatypes, and is designed for flexible and efficient I/O and for high volume and complex data. 

https://www.hdfgroup.org/HDF5/

**** There was a point in his talk where he mentioned modelling and observational data.  Scientific modelling is about analyzing a bunch of data and coming up with a model to explain a certain phenomenon.  There is a level of abstraction to them.  

https://en.wikipedia.org/wiki/Scientific_modelling

Observational data refer to data based on actual measurements, usually taken by instruments or sensors on spacecraft or aircraft.  (Or robotic geologists on Mars.)

END RECAP