[stem-dev] Notes from today's STEM call

["Judy Douglas" <judyvdouglas@xxxxxxxxxxx>]

STEM Community Call October 27, 2011

On call: Jamie, Kassaye, Matthias, Matt, Stefan, Kun, Judy

 

RELEASE PLANNING: Code Freeze Date. Matt will get STEM 1.2.3 RC1 up by Nov. 1

Response to query from Matthias: Code Freeze Date does not affect work his group is doing on the shape file importer
Release Date for STEM 1.2.3 is Nov 21  
Target date for 1.3 is 1/07/2012
* Ten years of Earth Science Data (2000-2010) available as STEM ''Features''
* Malaria Disease Model (''Anopheles'' calibrated)
* Dengue Fever Disease Model (''Aedes'' not yet calibrated)
* Additional NLS + User ability to switch languages
* Logger Framework with new loggers
   ====Release new loggers for 1.2.3
   ====Delete old loggers and views for 1.3
* Shape File Import Utility (   *********  EXPOSURE: Eclipse has not made progress yet on Open Map CQ. Possible Interim step - write import from KML) See additional info under Participants; Matt to re-contact Eclipse
* Integrating external models for study of food based transmission
* New Differential Equation solver(s) from commons.math library
* Population Rescaler
* Bug Fixes from 1.2.2
*** Need presentation slides for 1.3
NEW DOC FOR 1.3: No additions this week
*  Dengue
*  Initializers
*  Loggers
*  Malaria (done)
*  Mosquitoes VCAP (done)
*  Population Rescaler
*  Shape file importer (recently added)  
*  Solvers
BUG OF THE WEEK: BUG 258989 REFACTORING IS BROKEN

Actually a Feature request

Decision: Will try to address in January release by refactoring across projects  
Issues involved: renaming disease models, scenarios may create errors; may need deep copy
NEWSGROUP DISCUSSION TOPICS
STEM Runtime: Compiling Scenarios for high performance runs if want to run in different environments

Stefan: Has prototyped for dengue fever; welcomes comments & input on Newsgroup on what platforms users want for running smaller binaries
UPDATE ON DENGUE MODEL STATUS: Kun
Longer runtimes resolved the problem locating bifurcation point

Now running, have data, doing experiments (e.g., re immunity).
Simone visit Nov 1st                
 NEW MALARIA MAPPING ALGORITHM  
Jamie: Have paper ready to send to Hopkins  
Running data from Waraporn with both Dengue and Malaria data from Thailand; will add malaria data to paper when ready      
IDEAS FOR FUTURE STEM DEMOS: New item added this week; will continue to add as needed
>FD vs Integration
>Data Import example (playback)
>Data input example (initialize from csv)
>Data logging image and data examples
>Dengue Fever example(s)

>Disease Playback (NEW)
>>> Can we do a YouTube on each one? Stefan, what did you use to do the screen captures?
>>> Code freeze November 1st allows 1 month to build & test scenarios

NORWAY MAP: Kassaye

Map for/from disease prediction service for Northern Norway goes down to municipality level

One mistake in current map (19 counties, not 18)

Jamie: Kassaye will send us data; we will replace; will need approval from Eclipse [Use KML to create property files (identifiers and polygons), store in internal data, create plug-ins

MORE ABOUT GETTING BETTER MAPS OF EUROPE (ADDED)

Kassaye will check with his colleague to see if he has other maps that are consistent w/ the Eclipse license

Matthias will email his contact at the European mapping agency again

ITEMS FROM PARTICIPANTS

Queries from Kassaye:  Jamie will send him a playback disease model; referred him to wiki tutorials

Issues raised by Matthias: (1) Status of Openmap CQ w/ Eclipse? New contractor needs to know for conversion. Matt: Will renew efforts to get this through Eclipse Legal, get a quick read. Jamie: If contractor indicates what’s missing, we can identify tools to coordinate transformation. (2) Query re edges and shape files. Discussion: Shape file importer need not have edges; user can instruct; can use code as a graph import tool. Matthias will ask contractor to email Jamie to discuss. (3) Transmission rate varies by region for food mediated disease (cattle-slaughterhouse-distribution of meat in central Germany-disease in humans); not reflected by using two different population models (cattle and human). Possible solution: Use demographic model for cattle (alive and dead); transmission rate is not a property.