E-Cell 3.1.106 Released
Workshop Topics
Things that needs to be discussed.
Main page of the workshop is here.
Organization
- Who does what?
- Infrastructure -- web, wiki/cms, ML etc?
- How design/implementation decisions are made and approved, and by whom?
- How new developers are recruited, trained, rewarded?
Style
- Development as refactoring? E-Cell 3 code tree as the start point?
Code Management
- Coding Standard
- Is SourceForge-SVN good enough?
General
Goals
- What are the problems to be solved?
- Who are the intended users?
- Intended timeline.
Features
- Portability.
- Multi-Space.
- Dynamic model structure.
- Parallel computation
- Distributed computation.
- Models as Python scripts.
- Simulation algorithms -- what needs to be there in the first version?
Simulation Model
Simulator Core
- Rename System as Entity, Entity as something else.
- Entity-Variable-Process-Stepper.
- Is Stepper-Process relation clear enough?
- 'Views' for multi-spatial representation.
- How to support models with multiple spatial representations interacting with each other?
- Dynamic model structure API (Nathan).
Modeling Schemes
- Models as scripts.
- Modularity -- building larger models from parameterized smaller models.
- Persistence?
- How domain-specific semantics can be better handled?
Software Architecture
- C++/Python
- Straight Python wrapping of C++ classes in backend using boost-python.
- Flat functional API as in libemc is obsoleted.
Core
- Logging support as 'hooks'. Loggers are implemented in the frontend layer.
- Support for parallel computation.
Frontend
- Command/component/application suite? Same as ecell3 or something new?
- GUI toolkit?
- Logging.
- Support for distributed computation.
