W3C home > Mailing lists > Public > www-math@w3.org > November 2006

Re: ci csymbol are confusing

From: <juanrgonzaleza@canonicalscience.com>
Date: Tue, 28 Nov 2006 00:28:31 -0800 (PST)
Message-ID: <3097.217.124.88.225.1164702511.squirrel@webmail.canonicalscience.com>
To: <paul@activemath.org>
Cc: <www-math@w3.org>

Paul Libbrecht said:
> Juan,
>
> This kind of blur between an identified-entity and a fixed symbol is
> common in mathematics I have studied; for example, when working with
> rings you can easily talk of a "2" (as 1+1) as a pretty general
> identifier and a bit later make it a symbol which you somewhat fix
> within some concrete representation of the natural numbers (or even with
>  some generality!).
>
> why not be more constructive and give us a concrete math example that
> would approach your problem in a more direct fashion ?
>
> paul
>
>
>
> juanrgonzaleza@canonicalscience.com wrote:
>> What is difference:
>> A)
>> ci ---> abstract identifier
>> csymbol ---> concrete symbol defined either internally or externally
>> (url)
>>
>> B)
>> ci ---> local identifier defined in the spec or in the doc via declare
>> or similar
>> csymbol ---> external symbol defined via external resource (url). Both
>> interpretations are different and i received both of them.

Thanks by reply Paul.

I would apologize if i am filling the list with a non-issue, but i think
that query is important.

You appears to think that my doubt is on the difference identifier vs
symbol, but is not. My main question focuses on the role of csymbol vs ci
regarding definition issues.

The spec clearly says that a csymbol would be used for the definition of
new  elements are not contained in the core when the definition is
external and referienced via a url with definitionURL.

The spec clearly says that <ci> would be used for extension of MathML with
elements locally defined.

Carlisle also remarked the role of csymbol for external definitions.

However, a MathML author wrote me that my early interpretation of
identifiers vs symbols was correct, but was no my emphasis on locality for
<ci>. Exact words were:

"Thus, <csymbol/> would be used to reference known or specific constants
or functions - possibly defined in the containing document, or even in the
MathML spec - the definition location doesn't matter."

The _definition location does not matter_ is very different from use <ci>
for local elements and csymbol for externally defined ones.

I provided some abstract examples from the spec where however, it appears
that distinction symbol vs identifier is not maintained but of course i
agree that this is somewhat a matter of opinion and of conventions used.

If you want a more concrete example ok.

I want to define the sharp product (think of it as close to the star
product of non-conmmutative geometry) for two functions f and g.

<apply>
  <??>#<??>
  <ci type='fn'>f</ci>
  <ci type='fn'>g</ci>
</apply>

I use ci for functions because are generic functions, not concrete ones.

_If_ difference between ci and csymbol is A) of above then

<apply>
  <csymbol>#</csymbol>
  <ci type='fn'>f</ci>
  <ci type='fn'>g</ci>
</apply>

and

<declare>
<csymbol>#</csymbol>
...
</declare>

for internal def.

and

<apply>
  <csymbol
definitionURL='www.canonicalscience.com/CanonFormal/sharp...'>#</csymbol>
  <ci type='fn'>f</ci>
  <ci type='fn'>g</ci>
</apply>

for externally referenced one.

But _if_ difference between ci and csymbol is B) of above then

<apply>
  <ci>#</ci>
  <ci type='fn'>f</ci>
  <ci type='fn'>g</ci>
</apply>

and

<declare>
<ci>#</ci>
...
</declare>

for internal def.

and

<apply>
  <csymbol
definitionURL='www.canonicalscience.com/CanonFormal/sharp...'>#</csymbol>
  <ci type='fn'>f</ci>
  <ci type='fn'>g</ci>
</apply>

for externally referenced one.

"Localization definition does not matter" is different from "use csymbol
for external definitions and ci for internal."
Received on Tuesday, 28 November 2006 08:28:53 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 20 February 2010 06:12:59 GMT