Re: bytecode modules

JP Dinger <[email protected]> Mon, 4 Apr 2005 19:07:15 +0200
Newsgroups gmane.comp.graphics.y.devel
Message-ID <[email protected]>
On Mon, Apr 04, 2005 at 04:08:15PM +0200, John Olsson wrote:
[I miss something here]
> >What you _need_ is, as OP said, something restricted to what things on
> >your screen do. OpenGL primitives or somesuch. Not another virus engine.
> 
> And when this same problem was solved by NeXT some 20 years ago they 
> opted for DPS (Display PostScript) which was also used by Sun in the 
> later cancelled NeWS project.

And apple uses DisplayPDF in their aqua, if I'm not mistaken.

DPS is nice, in a way, but it is also (c)ed and [tm]ed and whatnot.
If only because PS is.


> James Gosling was one of the persons 
> behind NeWS. You could basically send a small PostScript program 
> implementing, for instance, a button from the server to the client. The 
> small button program would then tell the server if the user had clicked 
> on the button or not instead of having all that heavy communication 
> between the client and server. Sounds familiar? ;)

Yes. What I'd like --iff we're going that route-- is something that can
ideally be done in hardware, like a series of OpenGL primitives. What
I'd _not_ like is a random pet language that forces the server and, for
bonus points the client, to include an interpreter and probably the
(bytecode-)compiler and the rest of a toolchain. As eventually it'll
become /non-optional/, and I don't want to go there with pyton. Or perl.
Or java. Or lisp. Or BCPL. Or whatever else.

The idea behind this idea, as I see it, is that it might be used to
increase execution speed by providing ``canned messages'', or ``stored
procedures'', or whatever you want to call them, to decrease traffic
and context switches and whatever. And that those messages ideally but
not necesairily exclusively can be executed by hardware. What I do NOT
primarily look for or care about is increasing flexibility.


> You can read a bit more here
> http://pepper.idge.net/disaster.html

Ah, the unix-haters handbook.
I'll add[1]

  http://weblogs.java.net/jag/wsd.pdf

Different angle, different ideas. I'm not saying I fully agree with
everything the man says, mind you. He's throwing out low-bandwidth
support out, for example, and I might not want to do that.


[1] I thought it had been mentioned on this list before, but couldn't
    find it in the short time I looked.

-- 
  j p d (at) d s b (dot) t u d e l f t (dot) n l .