9. Contribute to / Recommend / Share pycos¶
If you find pycos useful, please consider contrbuting to the project.
The source code is available on GitHub. Here are a few suggestions (of course, you can implement anything else you need, or any pet idea!):
- More examples (real-world, academic, or demonstrative examples) can be added. Multi-agent systems, evolutionary / genetic computing, for example, can be implemented with Distributed Communicating Tasks (dispycos).
- Documentation can be improved.
- Support for fault tolerance can be added to Scheduler.
- Different schedulers (priority scheduling?) in pycos (for concurrency), or dispycos (for managing distributed tasks) can be tried.
- Name resolution of resources is not currently cached, so if, for example, a task is located again, pycos sends lookup request and waits for results. Caching can improve performance, but it means resources may need to be validated (periodically), pruned if too many resources are cached etc. Alternatley, external libraries such as Redis can be supported to store location information of resources.
- Comments in the code with ‘TODO’ are ideas / improvements / potential issues; these can be addressed.
- Recommend / Share pycos with the links below :