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

Nitin Borwankar (nitin@borwankar.com)
Sat, 25 Mar 95 22:27:13 EST

In your message you, Rich Salz, graciously said

>
> > In short, I think a standard scripting language without a standard object
> > model is pretty useless, so lets propose a standard object model(s)
> > first, and then decide if we need a common syntax.
>
> I totally disagree.
>
> I see no reason for safe-tcl to compete with OLE/COM, OpenDoc, Fresco,
> etc.

I don't see the conflict - an object model is independent of the
language used to implement it.
OLE/COM, OpenDoc etc have object models and then they have language
bindings. The two are in different semantic domains and comparing them
or suggesting that they compete is comparing apples and oranges.

The author of the original statement was right on. Designing an object
model requires thinking through a number of modelling issues which
are language independent . By focusing on safe-tcl as the vehicle of
agent technology some of the precursor issues - object model being one
of them - are being ignored.

What are the objects in an agent framework ?
What actions do they perform ?
On behalf of whom do they perform them ?

- these are the kinds of questions involved in an object model.

What has safe-tcl got to do with this except being one of the languages
in which an object model is implemented ?

I agree with the original poster that an object model is required
before all this agent stuff will begin to make sense in an
implementation independent way.

Nitin Borwankar,
Borwankar R&D,
nitin@borwankar.com