|
June 13, 2007
Participants: Jones, Barseghian, Gallagher, Potter, Altintas, Cornillon
- Nathan finished actor that reads opendap data, all data in one complex data type
- memory problems with even a moderate opendap dataset
- need to focus on the optimizations now -- matrix optimizations needed (e.g., for 2d array of int, long, double)
- probably 2 new types needed in Kepler
- Grid type
- N-dimensional arrays with optimized storage
- next step to change actor to add dynamic ports based on the type, so that the ports can reflect the structure of the source
- Ilkay
- Matlab actor, using the code from Zhang at the workshop
- Still really need Matlab actor to be as robust as the R actor
- Ilkay will write up a list of Matlab actor issues and requirements
- Only supports passing input variables -- need output too
- graphical output needed too
- issue with Matlab being local or remote (command line versus ssh)
- should we support octave? probably a good idea
- Derik
- hardware for the terrestrial use case -- setup a few days
- CR Basic code to write to PC will take a bit
- then move on to CR Basic to RBNB -- not sure if there is existing code for this -- need to search
- may be able to use other languages over serial directly
- probably still need to use CR Basic to configure the sensors
- Terrestrial usecases
- Ocean usecases
|
