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

Re: New language



"Rob Myers" <robm@tdv.com> wrote in message 392D0438.8A83997@tdv.com">news:392D0438.8A83997@tdv.com...
> The system bus idea was referred to in a previous posting. My take on it
was
> that it's a kind of central data pipeline for each process. This is in
> contrast to the callbacks and singletons used by languages that have to
> manually pass data between known nodes of functionality. It makes for a
very
> dynamic, environment-based system.

  Ok, is this something that's out there now?  Does it have some similarity
with the object freezing clipboard in NS?

> - You download the functionality as well as the data

  Right... suddenly Java just went up a few notches in my book.

> - The space-based programming paradigm is designed for distributed
computing,
> rather than squeezing an existing paradigm into the demands of a new way
of
> working as RPC/CORBA or DO/DCOM do. DO I like the most out of these, but
> JavaSpaces is still a tool designed to solve the problem, rather than
bring
> the problem to an existing tool.

  Well ok, but I'm not sure DO is "designed" for anything really.  I suppose
what I'm asking is what's different for the programmer?  I assume in both
systems you ID the space, id the object you want inside it, and then use it.
No?

> - It is a new way of thinking about structuring tasks. It's very flexible
and
> elegant, do read the JavaSpaces book online (even just the sample chapter)

  Ok, I'll do that.  The reason I'm asking here is that after attempting to
wade through the EJB stuff I finally saw a real explaination of it and EJB
sucks rocks.  Thus I try to avoid anything on the Sun site.  :-)

Maury





References: