[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: A call for reason, another go [SML et al]

• From: Vane Lashua <vlashua@R...>
• To: xml-dev@i...
• Date: Tue, 30 Nov 1999 12:17:12 -0500

In other words, SML may fill a very narrow niche that is probably better
served by paying a small penalty in complexity and size. Is SML a
three-week-old spam? Don Park: publish a document and ask for comments; or,
start the sml-dev list and see who comes; or, don't mention it for a couple
of weeks and see if it rears its head again on its own. Meanwhile, I'd like
to propose a much simplified version of XML called Virtual ML ...
Vane

-----Original Message-----
From: Dimitris Dimitriadis [mailto:Dimitris.Dimitriadis@l...]
Sent: Tuesday, November 30, 1999 10:27 AM
To: xml-dev@i...
Subject: RE: A call for reason, another go [SML et al]


I've been following this discussion, and have a few, non-technical points to
make (being a philosopher of language/logician I do not qualify for making
technical claims).

The analogy one could have in mind in trying to decide whether or not to
follow the SML-line is quite simple:

Given that the important question is whether, and if so, to what degree, XML
can model complex information structures successfully (and if SML, a simpler
version of XML could do the same job without added complexity), one need
only bear in mind that those kind of structures seem to be able to convey
meaning when uttered in natural language and heard by a human being. This
same level of semantics would be achieved if:

(i) XML (or any other markup language) had semantics --in and of itself--
which it of course does not (and this is the key to the entire
problem),therefore we have to opt for going for the least incompete *ML
(substitute * for a letter of your liking), and

(ii) Whether this more or less incomplete *ML can be maximized in order to
get as much structure as possible out of it. Given that no *ML is "complete"
in the sense mentioned above, we should opt for a high ration of
complexity/expressibility

If these two are accepted, and the claims made in (ii) are indeed worth of
pursuing, I see no point in taking one, fairly standardized *ML, namely XML,
and simplifying it to the degree where the cost of losing expressibility
clearly overweighs the gains concerning complexity by switching to a simpler
*ML, namely SML.

I hope I have made my point clear. I'd be glad to receive feedback.

Kind regards,

D. Dimitriadis

-----Original Message-----
From: Thomas B. Passin [mailto:tpassin@i...]
Sent: Tuesday, November 30, 1999 2:38 PM
To: xml-dev@i...
Subject: Re: A call for reason


Don Park wrote:
> ...
> Lately, I have been trying to gain a different perspective
> by thinking what if SML was here first and XML was actually
> SML 2.0?  This line of thinking adds a rather interesting
> appreciation of attributes.
> ...
> Don Park

I'm starting to like Don's thinking here.  What a neat idea!  But there is
one thing against the notion, which probably is responsible for  a lot of
the desire for an SML.  XML has got some historical baggage.  If it were
really SML 2.0, SML 1.0 would have had the same baggage.  But it seems that
the baggage is what people want to eliminate.

Entities are an example.  You could replace <, &, and the rest with
some escape mechanism (like the backslash, I don't mean CDATA), and get rid
of all other entity types.  That would probably simplify things.  But then
you aren't a subset of XML anymore.  On the other hand, if you omit DTDs,
you could still be a subset.  I don't sense a consensus on the importance of
being a subset of XML, but I don't think things can really go anywhere until
this is pretty well settled (to first order, anyway).  Look at the number of
arguments about attributes that are being posted.  The people who actually
write parsers have been saying that there is virtually no cost in memory or
complexity to having attributes.  There's obviously a lot of people who find
them useful.  I think that anyone who sees an attribute in HTML for the
first time understands how to use them and their syntax.  It doesn't confuse
anything to have attributes. So leave them in and move on!

Regards,

Tom Passin

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@i...
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To unsubscribe, mailto:majordomo@i... the following message;
unsubscribe xml-dev
To subscribe to the digests, mailto:majordomo@i... the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa@i...)

• Follow-Ups:
  • RE: A call for reason, another go [SML et al]
    • From: "Don Park" <donpark@d...>

• Prev by Date: [SML]My opinion - SML in general
• Next by Date: Unsubscribe
• Previous by thread: RE: A call for reason, another go [SML et al]
• Next by thread: RE: A call for reason, another go [SML et al]
• Index(es):
  • Date
  • Thread