RE: Inheritance that can add/remove and new occam?

[tjoccam@xxxxxxxxxxxxx (Lawrence Dickson)]

Twanging on my single harp-string again:

Much can be done with a simple dynamic OS that stacks occam processes
running as independent programs, connecting to previously defined
(allocated) channels which go idle after the child program finishes.

The big advantage of this over fully dynamic creation/destruction is
that there can never be fragmentation or unpredictable resource
conflict. I did it in DOS with 200 parallel programs on one 1-meg
HP palmtop.

Larry Dickson

> Message-ID: <C1B1A78EB953D211902D0040056AC48B075537@VRVCOM>
> From: Philippe Lemaire <philippe.lemaire@xxxxxxxxxxxx>
> To: occam-com@xxxxxxxxx
> Subject: RE: Inheritance that can add/remove and new occam?
> Date: Mon, 23 Oct 2000 16:34:58 +0200
>
> During my transputer work C004 were used to switch channel connections
> dynamically to allow one transputer to receive data from successive nodes...
>
> Philippe
>
> > -----Original Message-----
> > From: Nayef Baker [SMTP:n.baker@xxxxxxxxxxxxxxxx]
> > Sent: Monday, October 23, 2000 15:46
> > To: Oyvind Teig
> > Cc: occam-com@xxxxxxxxx; java-threads@xxxxxxxxx
> > Subject: Re: Inheritance that can add/remove and new occam?
> >
> >
> > I am not sure wheather this is very good idea or not!
> >
> > I used some quasi-dynamic approach to connect/disconnect process at run
> > time. Basically A calculated index of an array of channels allow various
> > processes to talk/end-talk on semi dynamic way..
> >
> > It would be great if Occam allows fully dynamic channel creation/
> > destruction..
> >
> > Nayef.
> >
> > On Mon, 23 Oct 2000, Oyvind Teig wrote:
> >
> > > Hi
> > >
> > > When I make library code of occam programs I end up wanting some
> > > action to be constrained. Either I don't want to send on a
> > > channel (because I have no server listening) or I don't want
> > > to listen on a channel (or want to listen on a channel that
> > > wasn't there originally). Now I use boolean flags (parameters)
> > > to get something like this.
> > >
> > > What I'm saying is: I want the local ALT to be "connected"
> > > with a new and dynamic branch, and perhaps even have removed
> > > a branch. I also want a local "!" (send) to be
> > > qualified/constrained by some kind of (dynamic?) flag; perhaps
> > > this is saying that I do or don't want to handle some part
> > > of the code (no: that output channel should really disappear).
> > >
> > > I would like a "new occam" to have this functionality, but
> > > is this possible without some kind of inheritance (like in
> > > SC++)?? (But then, inheritance can't remove, only add..)
> > >
> > > Is this possible in JCSP or CTJ?
> > >
> > > Which proves this is no religion: we do see shortcomings.
> > >
> > > --
> > > Oyvind
> > >
> > > @ Oyvind Teig (oyvind.teig@xxxxxxxxxxxx, oyvind.teig@xxxxxxxxxxxx)
> > > @   Navia Maritime AS, division Autronica, 7005 Trondheim Norway
> > > @   Tel: +47 73 58 12 68, Fax: +47 73 58 10 01
> > > @   http://www.autronica.no/
> > > @   Now part of world's largest company in maritime electronics:
> > > @   http://www.kongsberg.com/
> > > @ Publications at: http://www.autronica.no/pub/tech/rd/index.htm
> > >
> > >
> > >
> >
>