Re: agents [was: Client-Side Scripts & Web Scripting languages

On Mon, 27 Mar 1995, Rich Salz wrote:

> You can get an amazing amount of useful work done without an object model.
> Look at all computer programs before, say, the late 1980's.

Ok. What is *required* is an architecture. 
( What I said in an earlier message - I then made the arguable assumption 
that "The Architecture" would be an Object Oriented one. )

Having a layered architecture has been one of the things that has
enabled the internet to grow and evolve, even though most of the
technologies it was originally built upon are now obsolete ( or
rapidly growing so ) [ I don't think you can claim that programming
language design is a mature art, or that tcl is the last word in
that art! ] 


> And you want to stop progress until you can figure design a new one,
> or pick among the competing ones?

I don't want to stop ANY progress. 
However, since there are already a dozen or more incompatible
research or commercial ventures into the distributed agents 
language niche, I think it would be worth spending a bit of
time thinking about whether there is a posibility of some sort
of common architecture. ( Wouldn't you consider interoperability 
to be some sort of progress ? Besides - we were discussing mostly
capabilities that aren't yet part of the enabled mail version
of Safe-Tcl. I understand that several groups are working on
various safe-Tcl extensions. Don't they need some standardization,
at least within the Tcl world to work ? Or are you talking about
"stopping progress" until John Ousterhout looks at all of the
different version and "blesses" one of them as standard ? ;-) 

> Foolishness.
> 	/r$


However, I've recently been informed on www-talk that all this
discussion is moot, as Java has already won. ;-)

Despite statements from people at Sun that the two languages are
not in competition with each other, they clearly are. 

IF there can only be a Single Universal Network/Distributed/Agent
language, THEN it can't be both Java and Tcl ( and/or others )

IF there is room for more than one language, THEN there is a
need for *some* sort of interoperability plan. 

Java suggests *A* plan: compile other languages into Java Virtual
Machine bytecode. ( I'm not sure that it's the best plan, but 
it's a plan. I haven't yet read enough of the VM spec. or the Class
Library to judge. ) 


---|  Steven D. Majewski   (804-982-0831)  <sdm7g@Virginia.EDU>  |---
---|  Computer Systems Engineer          University of Virginia  |---
---|  Department of Molecular Physiology and Biological Physics  |---
---|  Box 449 Health Science Center    Charlottesville,VA 22908  |---

Received on Tuesday, 28 March 1995 02:56:38 UTC