| Newsgroups |
gmane.science.robotics.orocos.user |
| Organization |
Lund University |
| Message-ID |
<[email protected]> |
Herman and Daniel, the considerations you have put forward are understandable
and generally quite correct, I think. On the other hand, and that was the point
I was trying to make, there might be situations when we need to work
differently. For instance, people working with safety critical hard real-time
distributed systems often claim "that using time to synchronize two or more
activities" (or nodes) is required. Another issue is when closing feedback
control loops over networks, e.g. distributed sensing in robotics, then it
depends on protocols, dynamics, performance demands, and interfaces to included
legacy/proprietary systems, what approach is most suitable. There is also
programming on different levels; high-level specifications (end formal analysis)
is fine but somewhere on a system programming level the actual timing has to be
dealt with in such a way that high-level assumptions are fulfilled. About
sequencing and timing, as we all know, computers (so far, not counting the old
analog ones) execute sequentially and communication takes real time, so on a
system level we better have to deal with it explicitly.
I don't want to start any lengthy discussion about this, but perhaps we can
agree that there are rules and guidelines for general cases, and for development
of real-world systems we might have to adopt to actual demands. Orocos not being
dependent on a certain architecture therefore is good.
Bye, Klas
-----------------------------------------------------------------------
o--< mail: Klas Nilsson, Department of Computer Science
\ Lund Institute of Technology, Box 118, S-221 00 LUND, Sweden
O phone: +46-(0)46-2224304 (work), +46-(0)46-370160 (home)
+->>[_] mailto:[email protected] http://www.cs.lth.se/~klas
|
+-------<< Real-time << components << control << users << languages.