[Olpc-open] Information flow problem

[Sameer Verma]

Hello everybody,

Information flow is a critical problem for any organization. Some
researchers even point out that an organization is shaped by how
information flows within and outside of it. Free flow of information
builds networks. Restricted flow of information builds hierarchies. In
the OLPC context, information flow happens over several channels:
mailing lists, IRC, Talk pages, Wiki pages, phone calls, RT,
face-to-face, and IM (did I miss anything?). We all have preferences
for channels and applications. One can largely divide the channels
into synchronous (IM, Phone, etc) and asynchronous (e-mail, wiki) and
the applications that support these channels. We also tend to have
preferences for applications: wiki, forum, mailing list, IRC etc.
Then, there's the element of public vs private conversations. As a
researcher in Information Systems, I find these problems very
interesting.

Two problems arise:
1) too many channels (example: if I wasn't on the phone conference,
I'll miss out the details via IRC) lead to lack of critical mass and
fragmentation
2) The application (wiki or IRC or mailing list) is a hammer and every
problem looks like a nail that it can fix. "Throw it on the wiki" is a
source of a lot of misery!

Then there is the element of fashionable social networking (flickr,
twitter, tumblr, etc)...as if e-mail, IM, IRC, and chatter at cafes
aren't social networking! That topic is for another day :-) My
approach is that we figure out the problem first, and then find a tool
to fix it. Activity centric as opposed to application centric. Sound
familiar?

So, this semester, I worked with five of my graduate students who
undertook a Information Systems Analysis and Design project to analyze
the OLPC information flow problem and come up with some design
concepts. All the students were new to the problem. This was useful
because their perspective was quite new and they asked some very good
questions.

They used phone interviews, e-mails, in-person interviews, and
observations on the mailing lists, phone conferences, and the RT
system to gather data. A huge thank you to Adam Holt, Seth Woodworth,
SJ Klein and a bunch of other who contributed and facilitated.

In brief, they have pulled together the following:

A general problem mind map (Freemind)
Context map (Dia)
Data Flow Diagrams (Dia)
Entity-Relationship Diagram (Dia)
Prototype (Drupal)
Report and presentation (OpenOffice)

Their semester ends next week, and the report and presentation are due
on the 21st. However, given that SugarCamp is this weekend, we'll try
to post bits and pieces on the wiki in the hope that it will help with
some of the discussion (marketing at sugarlabs cc'd). In the spirit of
keeping things open and generative, we have decided to release the
documents, slides and diagrams under a CC license and also release
source files to make modifications easier. We've also stuck with FOSS
titles and open formats for all documents - this was a bit of a
struggle because some of the tools are not as mature as their
proprietary counterparts (Dia vs Visio) and the students were a lot
more familiar with the proprietary ones (Visio vs Dia).

There are some unfinished pieces, which will hopefully be worked on in
the next few months to add better definition to the overall flow of
information. Stay tuned to this thread for updates.

cheers,
Sameer
-- 
Dr. Sameer Verma, Ph.D.
Associate Professor of Information Systems
San Francisco State University
San Francisco CA 94132 USA
http://verma.sfsu.edu/
http://opensource.sfsu.edu/