[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re. Budgets, etc.
> I couldn't agree more. Aside from the fact that I have proposed it (unless
> someone else has too), the main reason I like the idea of a JCSP port of
> the KROC webserver is that it is such a (in principle) well-defined, thus
> easily managed, less subject to huge dispute, more likely to get done fast,
> etc., kind of project upon which most of the members could probably easily
> collaborate without a huge expenditure of effort that could (should?) be
> devoted to more ordinary academic and industrial work
Porting the KRoC webserver into Java/JCSP is a trivial exercise - the design
needs no change and the implementation could be translated almost line by
line (modulo the usual Java verbosity).
But I'm not sure why to do this? The occam webserver already makes the point:
o such highly multithreaded servers can be built very quickly indeed;
o it is easy to scale on to a distributed memory server cluster;
o it is easy to scale its functionality.
These properties would apply to a JCSP version - except that its performance
scalability would currently not be so good (since JCSP is not yet implemented
on top of the occam kernel ... but occam is!).