[Om3] CDSignatures/@cd, CDSignatures/@type: how to resolve?

[Christoph LANGE]

Dear James,

  OK, let me get it right. I confused you by mentioning these
application-defaults. Actually I don't need them. (But David?) So let me
propose the following:

If no cdbase is specified, http://www.openmath.org/cd is assumed as a
fallback. That makes arith1#plus point to the well-known definition.

And a practical advice which needs not be part of the specification:

If some application, for whatever reason, means something different by saying
arith1#plus, a cdbase different from http://www.openmath.org/cd must
explicitly be mentioned in any document where arith1#plus is intended to have
a different meaning. (After all, as far as I understand it, a symbol is only
completely and unambiguously referred to by the triple <cdbase, cd, name>, so
it might not be a reasonably choice of names, but one would not violate the
specification by talking about a symbol <some-obscure-cdbase, arith1, plus>,
which has a meaning completely different from our well-known arith1#plus.)

On Friday 09 May 2008 14:06:11 James Davenport wrote:
> Um - I am very unhappy here: I guess with the whole idea of an
> 'application-default'. If I write <OMS name="plus" cd="arith1"/> then I DO
> mean the standard one, and I DON'T want some application thinking it knows
> better. <OMS name="plus" cd="arith1"/> does NOT mean 'anythning you wnat
> to call plus'.
>
> ...
>
> This is my disagreement: I believe that if one sees arith1#plus it DOES
> mean the standard one.

Agreed -- _unless_ the author or the application explicitly defines a
different cdbase -- but not via some hidden defaulting mechanism, but
explicitly in the OpenMath document.

> > is officially defined in the openmath.org CDs is
> > (cdbase="http://www.openmath.org/cd", cd="arith1", name="plus"), or
> > http://www.openmath.org/cd/arith1.ocd, if the application chooses to use
> > URLs internally -- that's what I meant to say. Just that we make sure
> > that different systems that have their own local representation of CDs
> > are actually talking about the same thing when they say "arith1#plus".
>
> As I see it, this would be a significant change.

Maybe you misunderstood me here. I was not referring to application-default
cdbases here but just to different ways of encoding the triple <cdbase, cd,
name> in different applications. A semantic web application would prefer the
URL representation, other applications might prefer different representations.
E.g. one application would internally encode a symbol as
http://www.openmath.org/cd/arith1#plus, and another one with (symbolref
[cdbase->http://www.openmath.org/cd, cd->arith1, name->plus]), but both
applications must agree that this refers to the same symbol and has the same
semantics.

Best,

Christoph

-- 
Christoph Lange, Jacobs Univ. Bremen, http://kwarc.info/clange, Skype duke4701