W3C home > Mailing lists > Public > www-archive@w3.org > June 2005

Re: semweb syntax engineering

From: Sandro Hawke <sandro@w3.org>
Date: Wed, 29 Jun 2005 13:04:34 -0400
To: Dan Connolly <connolly@w3.org>
Cc: Dan Brickley <danbri@w3.org>, www-archive@w3.org, Eric Miller <em@w3.org>
Message-Id: <20050629170435.AC3F34EF61@homer.w3.org>


> "The machine needs to dereference the namespace/tag URIs to tell which
> tags are which kind of name, unless it has been given local information
> or some kind of over-ride."
>  -- http://www.w3.org/2005/06/semantic-xml/
> 
> that violates a basic design requirement of RDF/XML: that the syntax
> be locally evident. This was an operational requirement of PICS
> ("I can't afford to fetch a schema...") but I think there's more
> to it than that.

Right.

> Hmm... I thought this was written up in
>   http://www.w3.org/TR/1998/NOTE-webarch-extlang-19980210
> or
>   http://www.w3.org/1999/04/WebData
> but I don't see it.
> 
> On the flip side, if you're willing to give up the locally-evident
> requirement, you might as well go nuts. I have long thought that
> the constrained RDF/XML syntax was only one part of the story;
> that pretty much any syntax should be usable, as long as there's
> a schema somewhere that tells you how to treat it as a logical
> formula. This notion was in various drafts of various RDF WG
> charters, but we have not yet pursued it in a standards-track way.
> 
> GRDDL is sorta the extreme of this "go nuts" idea, with a
> totally turing-complete mapping to the constrained syntax.

I don't want to go nuts, I want something that's so simple that
even... uh.... people can use it.

What I really want is instead of using dereferencing is to use the
case of the first character.  But that's as cringe-worthy as using
indenting like python does....  :-(

For what fraction of applications does one need to parse the data
without having the ontologies present?  I guess this is the age-old
debate about whether you need the DTD/schema present to parse the
SGML/XML file.....

Any new insights on that one?

> Meanwhile, also in the syntax engineering space, I wonder
> if you ever saw...
> 
> data, an RDF syntax born of jetlag and nxml-mode
> From: Dan Connolly <connolly@w3.org>
> Date: Wed, 19 Jan 2005 13:30:42 +0200
> Message-Id: <8D710417-6A0D-11D9-A29A-000D9338C596@w3.org>
> To: public-cwm-talk@w3.org
> http://lists.w3.org/Archives/Public/public-cwm-talk/2005JanMar/0018.html
> 
> It's a response to the "tags are holy; only the standards
> gods can create them" XML orthodoxy, i.e. things like
> 
> RPV http://www.tbray.org/ongoing/When/200x/2003/05/21/RDFNet
> 
> and
> 
> TriX http://swdev.nokia.com/trix/TriX.html

Yeah, I think that's the style the rule language will end up using,
unless someone can pull a rabbit out of ... somewhere.    (Which is
kind of what I was trying to do, but....   it doesn't feel like it's
getting convincing yet, and as usual, there's not enough time....)

        -- sandro
Received on Wednesday, 29 June 2005 17:04:40 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 7 November 2012 14:17:52 GMT