RE: Re: dynamics: comment about MBdyn...

On Tue, 15 Jul 2003, Tiller, Michael (M.M.) wrote:

> > On Tue, 15 Jul 2003, Tiller, Michael (M.M.) wrote:
> > [...]
> > > > > The key thing is that "Syntax is not semantics".  Don't get me
> > > > > wrong, XML is a very nice technology and I use it every 
> > > > chance I get.  But
> > > > > it only covers the *SYNTAX*.  Modelica is much more than 
> > > > that.  Only a small
> > > > > part of the specification covers syntax.  
> > > > 
> > > > INdeed, and that's what the Schema's (formerly DTDs) are for: they
> > > > attach information about what is _meant_ with the parsed symbols.
> > > > So, I think Modelica would now talk about its XML and Schema
> > > > standards, instead of about its file format.
> > > 
> > > Not quite...schemas and DTDs only cover the structure of the file.
> > > They describe the grammar (what elements can appear nested what
> > > other elements, what attributes are associated with what nodes,
> > > etc).  They still don't cover meaning.
> > Of course they do! Look at a simple example, DocBook: when you use a
> > <section> ... </section> tag, you attach _meaning_ to the enclosed
> > block, and this meaning is, for example, used to determine what
> > typographical form to use when producing the DocBook layout.
> 
> I don't agree at all.  Should sections be numbered?  Should they be
> interpreted as ordered or do they float?
That different meaning that you give to the same file, by using
different Schemas. So, what't the problem? It's a nice example of
decoupling form and contents. (Really different from Modelica's
annotations :-) And you can attach different "forms" with the same
"content".

[...]
> I disagree.  You are associating a meaning for a word in the dictionary with meaning a computational sense.  I can feed:
> 
> <rigidbody length="5">
> </rigidbody>
> 
> Into any XML parser or even an XML validator.  It still isn't going
> to spit out a set of differential equations.
Of course, because the meaning is _not_ in the parsing, but in the
Schema! It's _there_ that the equations are.

Herman

-- 
  K.U.Leuven, Mechanical Engineering, Robotics Research Group
<http://people.mech.kuleuven.ac.be/~bruyninc> Tel: +32 16 322480