Re: [DotGNU]Defining Webservices => Language

[Gopal.V]

> Yes I wold also like to see this happening. However, GNUe is a little
> more set on python as the language for most things and I am sure how
> much dotGNU has looked at python. I like it more than Java and it
> doesn't have the same "issues", however I don't wanna go on a language
> flame war.
        Since GEAS would not need to be run in a SEE , the verification issues
are not present. Since we are using GEAS on the server side, we can use any
language there, as long as it has a compliant interfacing. As I said before,
it is only a matter of providing the plugins, and a standard interface ( what
ever it is ).

        We need an enterprise level Application Server, just like any other 
server
we communicate over a network. As long as we need not run any dotGNU apps ( ie
IL,Java,Parrot bytecodes ) in the server, language would not be a problem. 

        Also the Jython project can compile Py to Java class files. As a last 
resort we also have a java based python interpreter , JPython. If they are
GPL'ed, maybe we could just add a python extension some time in the future.
But those may cause performance/security concerns.

The GEAS would otherwise need a client side plugin ( which would be a LOT of 
extra work ). Bill Lance, please note this point please.

> We can return whatever types...be it an object reference or a SOAP
> object, etc. ODL/IDL allows for all of this sorta thing and we are using
> our own class definition format right now (loosely based on ODL). 
The nice thing about standards is that there are so many of them to choose from
                                                      -- Andrew S. Tanenbaum
        No hard feelings , but if everybody has his own format what are 
standards
for ?. I think ,(*my opinion*) that dotGNU would like a common standard but 
you are the experts in this field. I am presenting an outsider's look at the
issues at hand.

G'Bye
        Gopal
--
 //===<=>===\\
|| GNU RULEZ ||
 \\===<=>===//