[2557] in java-interest
Re: The MetaMedia Project
daemon@ATHENA.MIT.EDU (Matthew Fuchs)
Wed Oct 4 13:55:59 1995
From: fuchs@cerc.wvu.edu (Matthew Fuchs)
To: mmccool@cgl.uwaterloo.ca (Michael D. McCool)
Date: Wed, 4 Oct 1995 11:44:45 -0400 (EDT)
Cc: java-interest@java.sun.com
In-Reply-To: <9510022135.AA11837@cgl.uwaterloo.ca> from "Michael D. McCool" at Oct 2, 95 05:35:51 pm
Michael:
I've been working on SGML over the Web for a while. I've been arguing
that we need to move from using a small set of languages (HTML, jpeg, even
Java) for shipping information, to a small set of metagrammars (SGML,
PDES, regular expressions, even LR(1)), where we can separately ship
the document, the language definitition, and the semantic actions.
My goal is to enable intelligent clients. At present, even with Java
(as currently used), it is extremely difficult to develop client-side
applications (i.e., developed by the client, not the server) because
the client has no way to "understand" the information arriving,
because HTML only had display semantics. However, with SGML, the
information has a syntax corresponding to the domain, which the client
can use to supply its own semantics, or use default (display)
semantics from the server.
A discussion of this can be found in my dissertation (chapter 4) and
in a soon-to-appear article found on my home page at:
http://www.cerc.wvu.edu/~fuchs/fuchs.html
Take a look, and I'd be happy to discuss collaboration.
Matthew Fuchs
fuchs@cerc.wvu.edu
>
>
> I've been lurking on the mailing lists for a while, and thought I should let
> people know that we've begun a project at the University of Waterloo to
> create a "MetaClient" that can understand arbitrary SGML documents. Basically,
> the idea is to provide the structure of a document class in a DTD, then
> provide the semantics of all elements of that document type with a collection
> of Java objects that share a standard interface. A grammar-driven frontend
> would allow the interpretation of ARBITRARY languages as well as arbitrary
> SGML markup. The only "standards" involved would be a fairly unrestrictive
> interface used by objects defining the semantics of each element of a DTD,
> and SGML and Java themselves.
>
> Bottom line: I don't think Java should be used to distribute "applications",
> it should be used to distribute the semantics of the *languages* used to
> describe the content. Applets are handy if you WANT monolithic content;
> but...
>
> Our prototype, which we hope to release in late November, will be a
> "mega-applet" which should run inside the Beta version of the applet
> interface. Basically, it would use HotJava to bootstrap itself.
> Depending on how flexible the applet interface proves to be, and how
> efficient this approach is, this may or may not be our final
> implementation approach.
>
> I don't think this is a totally unique idea, and if anyone else is working
> on a similar idea I think we should hook up. We fully intend to make the
> result of our work available at the source code level, in the absence of
> other external constraints. We're also going to be using the system
> (hopefully) to support a distributed learning environment at the University
> of Waterloo.
>
> The URL http://www.cgl.uwaterloo.ca/~mmccool/Meta/index.html provides some
> links to documents on the project. Oh yeah, the index page contains a "logo"
> applet with the beta api.
>
> Michael McCool.
>
> PS Leo Chan, who posted earlier about 3D APIs, is also working on this project.
> In theory, the system we are proposing should make implementation of VRML,
> or any other 3D markup language, "trivial" if an appropriate 3D graphics
> backend is available in the client.
> --
> http://www.cgl.uwaterloo.ca/~mmccool/ mailto:mmccool@cgl.uwaterloo.ca
>
>
> -
> Note to Sun employees: this is an EXTERNAL mailing list!
> Info: send 'help' to java-interest-request@java.sun.com
>
>
-
Note to Sun employees: this is an EXTERNAL mailing list!
Info: send 'help' to java-interest-request@java.sun.com
