One particularly great thing about web services is
that deployment problems of Lisp or Smalltalk
applications are not a big consideration. The old joke:
"on the Internet no one knows you are a dog" applies
here: consumers of web services don't know and don't
care what languages those services are implemented in.
With SOAP support, Lisp and Smalltalk based services
should play nice with .Net and J2EE.
I expect (and I talked about this in my new J2EE book)
that web services will open up interesting markets for AI
applications for many reasons:
1. reduce deployment problems - put everything on
your own server
2. allow the use of huge databases on a central server
that do not have to be deployed to end users
3. protection of intellectual property: your code and
data stays on your servers; clients only see (for
example) a SOAP interface
-Mark
-- Mark Watson: Java consultant and author
-- Open Source and Open Content at www.markwatson.com
-- Commercial software: www.knowledgebooks.com
> One particularly great thing about web services is
> that deployment problems of Lisp or Smalltalk
> applications are not a big consideration. The old joke:
> "on the Internet no one knows you are a dog" applies
> here: consumers of web services don't know and don't
> care what languages those services are implemented in.
>
> With SOAP support, Lisp and Smalltalk based services
> should play nice with .Net and J2EE.
I agree with that... I'm sure that using Lisp on server
side for Web Services will be a great advantage in the
future. The dynamic nature of Lisp make it perfectly
suited for this use (Paul Graham has made a great article
on that subject).
I'm working for a company which use J2EE technology
to deliver game services on mobiles phones. We use EJB,
JSP/Servlets and a kind of Web Service (not in XML but
proprietary) to exchange informations our platform
and the handset. Our front-end server (which role is talk
with the mobile) as a fairly stable, well-defined APIs
and implementation but our back-end server which is
implemented using EJBs has to deal with heavy marketing
changes to add new services and compute new statistics
informations. In this case, we take too much time to
integrates these changes and the deployment and debug
phase is a real pain...
IMHO Java has bring some solutions to common software
engineering problems which are usefull for us (such as
platform independance, a set on great APIs and tools)
I have to admit that but the language itself really
lack dynamic features and speed of development of Lisp!
In my point of view, re-developping our main server in
Lisp will speed-up the development, the deployment,
the test and validation phase.
I thunk future Web Services will be a second chance for
Lisp to prove its fantastic value to enterprises. I try
to push my company and my colleagues into that way!
--
Frederic Brunel
"Frederic Brunel" <······@mail.dotcom.fr> wrote in message news:<························@news.free.fr>...
> I thunk future Web Services will be a second chance for
> Lisp to prove its fantastic value to enterprises. I try
> to push my company and my colleagues into that way!
"Pushing" them that way might be counterproductive, because you might
lose credibility by being on the wrong bandwagon. A much better way
might be to implement in Lisp something very useful to them and then
try to persuade them to use it. Once they get in the habit of using
Lisp apps, Lisp will start to be taken seriously.
Or you could start your own personal web site using Lisp, and use it
as an example and demo. Whenver they want to add new features to
their web site, you can add those features to yours, while they're
still discussing them, and show the demo as part of the discussion.