Re: Wish for MathML 3: version attribute

I started writing to this topic a few hours ago but had to
stop because of an interruption.  When I returned to
my draft, I notice most of what I was going to say had already
been said :)  

This is a summary of what I said in my earlier draft that now 
needs much less comment: "A new namespace and/or mime type 
for MathML3 is a bad idea. PIs would not be a very intuitive 
way of specifying MathML version since they specify a target
application."

But, to go a little further, it occurs to me that a "version" 
attribute to the <math> tag might solve a lot of problems.  
In particular it might also specify whether an application is or 
isn't obliged to read the DTD.

It seems to me that there are a number of DTD options,
e.g. whether to include or omit the list of character entities
or replace them with something else.  I can also imagine 
a p-MathML only DTD, or a DTD that combines p-MathML with 
OpenMath instead of c-MathML. There are also some small 
changes e.g. to the MathML prefix that can be conveniently 
made in the internal subset of the DTD.  Some of these options 
might be selected using DTD modules.  The document might signal 
via a "version" attribute to an application that doesn't normally 
bother to read DTDs whether it should do so, or might signal a 
summary of the DTD choices to the application in this way.

I don't have a specific syntax in mind, because I haven't got
a complete list of all the various options that may be needed.
Maybe this needs two tags. Something like
  <math version="MathML3.0" >...</math>
  <math version="p-MathML3.0+OpenMath2.0"  >...</math>
  <math version="p-MathML3.0"  >...</math>
  <math version="MathML3.0" dtd="important" >...</math>
etc

What do others think?

Richard




On Mon, 2007-09-17 at 08:12 +0200, Max Berger wrote: 
> Dear MathML community,
> 
> I've been reading a little bit of the discussion, and here are my 2 cts:
> 
> a) I think declaring MathML 2 obsolete is not a viable option.  
> Discouraged yes, but not obsolete. There is still plenty of software  
> which uses MathML 1.0.1 (OpenOffice, for example), and is very  
> unlikely to change easily. This would only hinder the transition. If  
> you look at HTML, there are still many documents written in 3.2 and 4.0.
> 
> b) I really like the idea of a different namespace: MathML is mostly  
> embedded in other documents, and namespaces are the way to identify  
> it. A new namespace would mean that existing software will not  
> immediately recognize MathML 3, but it will be easily adaptable.
> 
> To reason again for my originial post (version attribute) it does not  
> have to be an actual version attribute, but I would really like to be  
> able to detect which MathML version a document conforms to without a  
> given DTD.
> 
> Max Berger
> e-mail: max@berger.name
> 
> --
> PGP/GnuPG ID: E81592BC   Print: F489F8759D4132923EC4  
> BC7E072AB73AE81592BC
> For information about me or my projects please see http:// 
> max.berger.name
> 
> 
> Am 16.09.2007 um 04:46 schrieb Andrew Miller:
> > In the end, everyone was happy with using one MIME type, a different
> > namespace for each version, and defining an umbrella specification to
> > which all versions of CellML comply which essentially describes how  
> > the
> > version is encoded in the namespace.
> >
> > Best regards,
> > Andrew

Received on Monday, 17 September 2007 11:28:44 UTC