[om] Specfun CD's (was: library)

[Bruce Miller]

Richard Fateman wrote:
> 
> I found James' paper at
> 
> http://icm.mcs.kent.edu/research/iamc2001.papers/davenport.pdf
> 
> Having scanned through it, I do not see what questions
> of style are being raised.  Perhaps I must read it more
> carefully, but a hint as to a page number (1-14) would help.

Hmm, I must have already gotten the revised version ---
See page 20, under "Future work on CDs".

Two points are explicitly raised:
 1) Multiple CDs with short function names (eg. Bessel:J)
   vs. 
    A single CD with longer names (eg. BesselJ)

 2) Currying (eg. \BesselJ_\nu)
   vs. 
    not (eg. forall z: \BesselJ_\nu(z)).

And an implied one: 
 3) deciding on `standard' branch cuts.

*** Some questions:

Point 1: What is the impact of such a choice?
Clearly, a finer granularity might make it
easier to revise (sub)CDs in the future, at the
cost of having more CDs to cope with.
Are there other issues?

Point 2: It's not clear from the example (on page 21)
how the CD itself depends on this choice, as opposed
to the OpenMath core or the corresponding STS.

Perhaps I need to re-read the OM specs, but if F is
intended for 2 arguments, is
  <OMA><OMS name="F"/><OMV name="x"/></OMA>
not implicitly curried?  Or an error?
If the latter, can you point to examples of how
currying is explicitly handled?

Point 3 is difficult.  Clearly something is needed.
I have a _definite_ bias, here, but I think that the
only viable resource for these determinations are
the Editors & Authors of the DLMF project.

However, when I've raised such questions in the past
the answers have been somewhat ambivalent "Oh, the
branch cuts are whereever you need them -- use
analytic continuation"    Perhaps I'm misrepresenting
thier point-of-view, and the comment really translates
into "Oh, please dont make me think about that _now_"
But, it may be difficult...

*** Some Comments on my submitted CD.

I developed my CD in parallel with a LaTeX style.
This led to certain choices in the above; they
are not manditory, by any means, as long as I can
get from here to there! :>

Since latex doesn't have any usable namespace system,
I put everything in one large CD.

And I am in favor of currying; J_nu is an interesting
thing in itself.  

As an aside; the latex macros I defined make an 
artificial  distinction between `parameters' 
(sub/super/pre/whatever)-scripts and `arguments' 
(fenced arguments).  The `parameters' are arguments
to the macros; the `arguments' are added by the
author, if needed.
This is largely a presentation distinction, but
it impacts the ease authoring and eventual parsing .
[The normal latex markup of parameters is exceedingly messy 
and random making parsing difficult.
Authors need the freedome to markup argument lists 
differently: with different fences or none,
with different separators.]

----------------
Bruce Miller
<bruce.miller at nist.gov>  http://math.nist.gov/~BMiller/
--
om at openmath.org  -  general discussion on OpenMath
Post public announcements to om-announce at openmath.org
Automatic list maintenance software at majordomo at openmath.org
Mail om-owner at openmath.org for assistance with any problems