1) Bringing up Java 2)Promela 3)Limbo 4)SPOC

[Oyvind Teig <Oyvind.Teig@xxxxxxxxxxxx>]

Still alive..
The Java Threads group, 29.9.97

-----------------------------------------------------------
Ted Lewis is now writing for Internet Computing, and 
has an article that describes "what went wrong" with
Java Aug.97:

1. "Java ain't got no state"
   Only low-level and minimal facilities for distributed 
   computing available:
2. He proposes to introduce a "fan" construct for
   doing the same operation on different data..
3. ..a "tree" construct for multivalued input and scalar
   output and..
4. .."par" & "pipe" constructs. It "is better" than a fork
   in Unix because "the compiler and interpreter can
   deduce things about the computation's state that a 
   C compiler cannot deduce about a fork-join in Unix. It
   is safer and hides a lot of error-prone code involving 
   sockets and threads."

-----------------------------------------------------------

By the way, has anybody looked at Bell Lab's "Spin", "XSpin"
and their "Promela" language? It's supposedly based on CSP, 
(procs and chan-of's (?)) but is only a meta-language, 
usable for analysing for deadlocks etc. An article 
in Dr.Dobb's pointed to it. Is has a nice
graphical wrapping, and it produces PostScript outputs of
MSC's (Message Sequence Charts). 
They had a conference at UOfTwente a week or two before us
now in April. 
http://netlib.bell-labs.com/netlib/spin/

-----------------------------------------------------------

Another by the way, has anybody looked at "Limbo",  by
Lucent Technologies? Also another chan-of-protocol language.
I think somebody in this community pointed to it.
(I imagine Lucent is also Bell Labs?)
http://inferno.lucent.com/inferno/limbo.html

-----------------------------------------------------------

Autronica's first system with SPOC-code will be delivered on a 
new-build at Kvaerner-Gowan in a month. 20.000 lines of automatically
generated C-code - much occam on top of that (chan-of-protocol
language of my preference!) and a lot of hand-written C-code
for DSP handling in the bottom. We run all this on a Texas
TMS320C32 DSP. 

----------------------------------------------------------- 

Cheers,   > 0yvind Teig, Autronica, Trondheim, Norway <
Oyvind    > Oyvind.Teig@xxxxxxxxxxxx                  <
          > Tel.: +47 73 58 12 68                     <
          > Fax.: +47 73 91 93 20                     <