Re: Usefulness and Limitations of XML
On Fri, Jul 28, 2000 at 01:13:54PM -0700, Fulton, Jim wrote:
> ...Neither BNF nor XML have formal semantics themselves, ...
BNF most definitely has a formal semantics (despite the fact that no one
seems to have bothered to write one up explicitly :-). The semantics of
a well-formed BNF is a grammar, i.e., a well-defined mathematical object
consisting of sets of symbols, strings, string concatenation functions
(production rules), etc. This can all be expressed very precisely.
Something analogous had darn well better be true of XML.
> ...or in and of themselves provide semantics for the languages defined
> using them.
Yes, of course, that is not the role of a BNF.
> > XML schema are not scheduled for realease for many moonths to come,
> > we cannot hold up our effort by waiting around. I stand by my view
> > that we need to develop our own SUO language (or adopt or adapt an
> > existing one). Separately form this, it is probably important to
> > XMLify our language so it can easily be ported. I believe that this
> > latter step is conceptualy trivial.
> >
> Wait a minute. We don't have time to work with a standard
> language-system that has a reasonable chance of industrial use, but we
> do have time to develop our own language!! Come on! Who's going to
> use it? Is this just another academic game?
Have a decaf, Jim! :-) Mike's point, I think, is that current XML
does not natively contain the constructs needed for an SUO (or for an
interlingua) -- in particular, the standard apparatus of first-order
logic. Hence, we need a language *anyway* with that expressive power.
KIF appears to be the one, at least for an interlingua. (Ontologies
proper, perhaps even the SUO, should perhaps be expressed in a more
tailored language -- see Pat's earlier note on this. I think there is
still some unclarity about KIF's *exact* role vis-a-vis ontologies, the
SUO, etc, but that's another discussion.) As soon as KIF's syntax has
been fully determined -- and it seems to me very close to that -- an XML
DTD can be defined, and folks can freely move between KIF and its XML
representation.
> Rather than waste our time with an utterly redundant language
> description language, let's use an existing standard language
> description language (playing the role of pilot project as we need to)
> to define XKIF.
Now I'm getting confused. Is it KIF or BNF that is the object of your
scorn? What needs definition is the syntax of KIF (or some equally
expressive language), whether by BNF, XML (?) or any other well-defined
means. That has to be done before there can be an XKIF, right? Where
exactly do you see the potential for wasted time? (I'm not denying the
potential is there, I'm just trying to get clear on where you see it.)
And what, exactly, are you proposing?
> Not only will we have our ontologies, we'll have a standard way to
> exchange information expressed in those ontologies.
I'm not sure what you are claiming here. Perhaps you could clarify.
Regards,
-chris
--
Christopher Menzel # web: philebus.tamu.edu/~cmenzel
Philosophy, Texas A&M University # net: chris.menzel@tamu.edu
College Station, TX 77843-4237 # vox: (979) 845-8764