[technology-pmc] Proposal for STEM to be a project under the Technology

[Daniel Ford <daford@xxxxxxxxxxxxxxx>]

Following up on Wayne's note regarding
the OHF project shutting down.  The STEM project http://www.eclipse.org/ohf/components/stem/
(an OHF component) would like to propose its inclusion as a separate project
under the Technology PMC.

STEM is an acronym for "Spatiotemporal
Epidemiological Modeling" system.  It is a Eclipse based RCP
designed to be a globally accessible platform for disease modeling.  The
short summary is that it combines mathematical disease models with extensive
(i.e. "complete") global geographic and population data sets
and geographic visualization.  It leverages Eclipse's plugin architecture
to allow other researchers to contribute their own mathematics, data and
other extensions.  It should be noted that while the initial target
for STEM is disease modeling, it is general enough, and is implemented
to facilitate, other types of modeling scenarios, some of these could be
disaster planning and response, real-time situational awareness, military
modeling, public policy decision support and layered economic models.

One of the main innovations in STEM
is use of a "composable" graph as the representational framework
for its simulations.  This means that a researcher can create a model
by combining various components of the graph from different sources.  STEM
includes  "built-in" graph components representing all of
Earth's 244 "countries," as defined by the ISO-3166 level 0 standard.
 The components include political subdivisions, geographic relationships,
population data, transportation infrastructure and other aspects of the
political and physical geographic of the Earth.  These "building
blocks" can be incorporated into any model and their general availability
avoids the need for disease model researchers to "reinvent the wheel"
by having them find and manage their own data sets.  In addition,
a researcher's own models can be exported from STEM and imported by others
to facilitate collaboration.  STEM includes extensive geographic visualizations
of disease spread.

STEM currently has four Eclipse committers
who are funded by both IBM Research and the United States Air Force.  The
committers work in the United States and Israel.  In the future, additional
developers may join the project from IBM Research's labs in Asia. The project
has extensive connections with academia including the University of Vermont,
MIT, Tel Aviv University and Johns Hopkins Bloomberg School of Public Health.
The project is also ramping up its publishing efforts with a recent acceptance
of a paper at BioSecure 2008 (joint with University of Vermont) and a submission
of a paper on its transportation model to PLoS One.  A book about
the project is in the proposal stage.  STEM was presented at the CODIGEOSIM
Disease Modeling workshop at York University in Toronto last month, and
will be presented at the INFORMS conference http://meetings.informs.org/DC08/
 in Washington DC on Monday, and at the IBM Global Pandemic Initiative
(GPI) workshop in Washington DC, later in October. We'll be proposing a
talk on STEM at the next EclipseCon.

The STEM code base makes extensive use
of EMF, with all of its core components being modeled and then generated
by EMF.  It also leverages BIRT for many of its visual components.
There is an extensive set of JUnit test cases and the implementation has
integrated run-time self testing in the form of a project specific DBC
"class invariant" implementation.  The project's IP Health
is good, all of the data sets have passed through the Eclipse IP process
and all of the code has been written by Eclipse committers. STEM's feature
set has stabilized in recent months.  The project's builds are stable.
 The documentation is good and improving (the project has hired a
dedicated technical writer to help with this area).  The current version
is 0.3.0a.  No project is perfect, or ever done, there are still many
things to improve.  To that end, we're working on improving our release
engineering, our usability and we are developing more advanced mathematical
models and techniques for inclusion in the code base.

The project has been well sheltered
under the OHF umbrella as it has matured, but given that that project is
now "shutting down", the project members are very happy with
Eclipse and think it would be a good time to move directly under the Technology
PMC.  We would use such a shift as an opportunity to revamp the project
website and supporting documentation as well as refactor the code base
to reflect our experience and hindsight.  It would also be a good
point for the project to expand it's efforts to attract more developers;
having a more visible, independent, presence under the Technology PMC should
help with that.

Many thanks for our consideration of
this proposal.

Daniel Ford
Jamie Kaufman
Yossi Mesika
Stefan Edlund

IBM Research