W3C home > Mailing lists > Public > www-rdf-interest@w3.org > November 2000

Re: Semantic Document Framework(s)

From: Aaron Swartz <aswartz@swartzfam.com>
Date: Mon, 06 Nov 2000 15:36:46 -0600
To: "Sean B. Palmer" <sean@mysterylights.com>, <www-rdf-interest@w3.org>
CC: <www-talk@w3.org>
Message-ID: <B62C7F06.1C291%aswartz@swartzfam.com>
Sean B. Palmer <sean@mysterylights.com> wrote:

> The current outline of a Semantic Web (and RDF/Schemas etc.) is all well and
> good [if a little fuzzy], but at the end of the day we will still need a
> Semantic output format for the SW, i.e a document format that is
> semantically compatable and more to the point, works well, with XSLT and RDF
> etc.

While this is certainly useful, I see it only as two pieces of the puzzle.
In my view, there are two simultaneous realizations of the Semantic Web.
First, there is "smarter" documents. This, I believe, was the original use
of XML, and is certainly the goal for XHTML and related formats. While still
documents, like those on the Web today, such documents will have richer
semantic meaning. Foremost, such documents are intended for human
consumption, with the semantic definitions (for computers) added afterwards.

Second, there is the syndication and publication of pure data, which is what
XML is being mostly used for today, and which is definitely a possible
(intended?) use for RDF. This data is primarily intended for other
computers, which then interpret it and publish it in a human-readable form
as an XHTML/HTML document.

> I hope people follow what I am saying here: a semantic output medium, but
> based on HTML tags so that it works in current WWW browsers, as well as
> whatever upcoming SW browsers there are. Put more technically, but
> innaccurately: for document structure use XML DTDs; for semantic filling and
> content use Schemas and RDF.

While a great goal, I think this could be very difficult. Prose, as humans
are used to reading, is very difficult for a computer to understand. Machine
data, which computers love to read, is not the loveliest thing for a human
to read. If you could find a way to combine these two, it would be great,
but I doubt it could be done without either a lot of duplication or a heavy
bias to either (computer or human) side.

> O.K., this does take these ideas on quite a few steps further, but nothing
> at all like this has really been developed yet. I guess what I'm trying to
> say is "let's stop using class=[...] for RDF, and get this show on the
> road!"

How do you suggest we do such a thing? For a simple example, let's say my
homepage reads:

Ora wrote this page.

In RDF this would be:

<rdf:Description about="">
    <s:creator>mailto:ora.lassila@research.nokia.com</s:creator>
</rdf:Description>

To encode this information in the original, we'd need some markup along
these lines:

<statement>
    <subject>
        <a href="mailto:ora.lassila@research.nokia.com">Ora</a>
    </subject>
    <predicate="s:creator">wrote</predicate>
    <object><a href="">this page</a>.
</statement>

Seems like a lot of work to me.

> Of course, the bottom line is that the SW/RDF *has* to be made for accessible
> to the public...

Of course! If it wasn't, no one would use it, and there wouldn't be much of
a web!

-- 
[ Aaron Swartz | me@aaronsw.com | http://www.aaronsw.com ]
Received on Monday, 6 November 2000 16:37:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:46 GMT