It proposes to extent possible, the DA software be separated from other type’s code in a way which is obvious to external users and invites reuse across multiple RSS projects. Now it lists potential benefits of this separation and responds to possible criticisms. Both drawbacks and benefits are viewed through the prism of the two objectives.
Potential advantages
1. Wider reuse: it expects that the proposed change would allow community wide library based software reuse in addition to existing framework wide component based reuse. More reusable code would be available to any given robotic practitioner regardless of the chosen RSF ultimately. It is the main advantage of the proposed change.
2. Separation of concerns: more manageable unit is common software engineering practice factoring software into simpler. The expected outcome is increased code quality because it is easier to understand, write and maintain. Without the DA and CM software, an RSS project would contain single purpose RSF code. The mission of such thin RSF project would be clear; enable integration of external code of DA into a complete robotic systems using an external CM. smaller and more uniform projects are easier to maintain overwhelming job in its current form.
3. Code stability: the RSF software tends to be in constant flux in practice. The DA code is shielded from this volatility when written in RSF independent form. A stable code base contains fewer bugs.
4. More users: debugging rate scale favorably with the number of users. More users for a given unit of DA software mean better software quality and fewer bugs.
5. More contributors: contributions do not scale as well as bug reports. Nevertheless, we hope that contributions may increase because of specialization in software.
6. Lower entry barrier: thin RSF can be numerous.
7. More meaningful evaluation: the three types of software are very different and should be evaluated using appropriate criteria separately.
No comments:
Post a Comment