[Om3] Math in office documents/Providing an implementation for a symbol

[Paul Libbrecht]

Bryce,

This an interesting thread... quite a pan-galactic one.

I tend to agree with the vision described in:
   http://lists.oasis-open.org/archives/office-comment/200812/msg00016.html
(I haven't read attachments yet)
One seems to be able to propose even a lot more with the web at hand.

One thing that really bothers me is that, on the one hand, you wish to  
have a universal representation of mathematical objects and, on the  
other side, you propose to carry a *single* implementation.

- why single? (I guess the answer is "because it's the simple way" but  
that doesn't convince me... in many bizarre cases alternate  
implementations are useful

- why not hope that arbitrary FMPs of OpenMath CDs can make it that  
tools devise implementations that satisfy these FMPs? (note: devise  
does not mean execute, at least, not in all cases)

The big big big issue with a single implementation is the "topology of  
symbols" which makes it that if you specify your implementation P in a  
document you send me I have to take your implementation and cannot  
trust my software that my implementation is also applicable: it makes  
the whole math organized as a tree with central things without freedom.

I wonder if you don't wish to port this to om at openmath.org which is  
slightly broader than om3.

paul







Le 15-déc.-08 à 23:58, Bryce L Nordgren a écrit :

> I am indulging in a thought experiment and I was wondering if I  
> could enlist your aid, at least to determine whether I am barking up  
> the correct tree or not.  The central question is: "How can an  
> office application suite better manage mathematical information in  
> an integrated fashion?"  Succinctly, I suggest that all occurrences  
> of mathematical objects have a uniform representation so that  
> cutting and pasting works between presentation, text, and  
> spreadsheet docs.  I also suggest that the formula document  
> (currently undefined) be defined as a workbook modeled after the  
> CDGroup and each sheet is a CD.  An outline can be found here, if  
> you're interested: http://lists.oasis-open.org/archives/office-comment/200812/msg00016.html
>
> So here's where I would like to ask your help: I'm fairly new to  
> OpenMath/MathML and I'm exploring how to define simple named  
> functions with a formal parameter list, which can later be used  
> (e.g., a symbol).  I've written a brief summary of what seems to me  
> to be the most important concepts, given examples of most of the  
> concepts, and culminated in the writing of a CD containing a really  
> dumb function called "myfunc".  Could I get someone here to look at  
> this and see if I'm on the right track?
>
> One thing I did note is that there does not appear to be an orthodox  
> way to specify a symbol's implementation, even when the expression  
> can be completely defined using other symbols.  (e.g., the equation  
> implementing Planck's law for blackbody radiation) Yes you can  
> include it as a "formal mathematical property", but there can be  
> more than one of those, and a function can only have a single  
> implementation.  The attachment explains this concern in more  
> detail...<OpenMathFunction.pdf>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://openmath.org/pipermail/om3/attachments/20081216/c23b052e/attachment.htm 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 2203 bytes
Desc: not available
Url : http://openmath.org/pipermail/om3/attachments/20081216/c23b052e/attachment.bin