[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: FW: It was a stressing weekend!

Hoi Gerald Hilderink,

> >It's unacceptable to rely on a particular VM's implementation of threads;
> the
> >behaviour may differ accross platforms,
> Does this mean that you wonn't use Java threads anymore?

I stopped using them as soon as I left Bakkers' group.

The idea of cloning every communicated value is unacceptable to an object
oriented mind.
It may serve well for control purposes, but not for mine (plain OO design).

> I agree! That's why we build our own scheduler that tightly specifies
> similar behaviour across platforms.

This probably means your JavaThread-exploiting code is not Java Compliant,
because it cannot run on an arbitrary standard VM.
This sounds like the lawsuit between Sun and Micro$oft...
Better be careful with your claims of Java Compatibility!

> >Once again, I disagree.  The synchronisation constructs function well for
> >building conceptual models of systems.
> I disagree, because the Alternative bug is a result of this.

I'm afraid not all implementation bugs can be pointed at the language...
Except, of course, where there is a disagreement between language specs and
VM implementation.
But it certainly is possible to define a correctly functioning ALT in Java
in spite of the partiality of its specs --- even though it is not easy.

> By experience, I have learned otherwise.

Your experience, I am sorry to say, is limited to a very specific angle to
Java --- one for which the language was not primarily intended.  That means
you can only claim experience facts in that particular domain.

Keep in mind that you are implementing one concurrency philosophy in terms
of another one --- such actions are deemed to be somewhat

> Is this acceptable? If Java is *not* designed as a fully controllable and
> entirely deterministic language then is there a futhure for EmbeddedJava and
> Real-Time java?

That is a very good point.
Unless RealTime java starts from other basics and basic constructs than you
do. But this is outside my territory; better let others answer that.


Rick van Rein,
	Spiritus flexibilis in corpore flexibile
Universiteit Twente * INF-3027 * Postbus 217 * 7500 AE Enschede * 053-4894291