[om] CD questions on the occasion of permu1

[Mike Dewar]

On Wed, Oct 16, 2002 at 08:27:23AM -0400, Michael Kohlhase wrote:
> > 2. The name: There is a CD called permut1 by Andrew Solomon, so a I
> >    chose a variation. It could be abbreviated to perm1 or elongated to
> >    permut1 (if Andrew and James agree), permutat1, or permutation1 (not sure
> >    though that such long names are allowed).
> 
> I think that long names (unabbreviated are better, since they can otherwise
> get quite cryptic). "permutation1" would be my preference. I have looked in
> the standard, and I could not find a restriction (there was the intuition
> once that the file name should be the same as the cd name, but that should
> not be a problem still).1
I agree that long names are better.  The idea that the filename should
be the CD name does however suggest that it shouldn't be more than 8
characters to support ISO9660 (CD) file systems etc.  On the other hand
this seems a pretty archaic restriction tehse days.

> > 4. A symbol can stand either for a `real' function (like sin) or a
> >    constructor (like `rational'). At any rate, in the encodings you
> >    can often preform surgery and find the arguments back. But not on
> >    the abstract OpenMath level. So my question is: what is our general
> >    approach to finding arguments of contructors? In the example of
> >    `rational(123,456)', would you like numerator and denominator to be
> >    around so that the children (123 and 456) of the expression can be
> >    accessed?  the vector_selector is an example (in a case where the
> >    number of compnents may very...). We find the i-th argument of an
> >    OM list by applying the list to <OMI>i</OMI> using an <OMA>...

I'm not sure that the semantics of `numerator' is that it returns the
first argument of the constructor `rational' (not least because you can
say "numerator(r)".  At the moment we haven't tried explicitly to
standardise names but have inherited `selector' from MathML (giving 
`vector_selector', `matrix_selector', `list_selector'). 

> > 6. Regarding FMP's: there are many one could write. What are the
> >    practical guidelines here? James has written some guidelines. These
> >    are (roughly): have CMP correspond with FMP and don't put in too
> >    many. The current version of permu1.ocd is large because of
> >    examples, FMP's etc., not because of the number of symbols it
> >    defines.
> 
> I beg to disagree, the FMPs are part of the fun, since they go a long way
> to make the meaning and the usage of the symbols declared in the CD
> clear. How can they do any harm in a world of 20 Gb minimum hard drives
> even in laptops. 
I think the thought was that FMPs should be clearly understandable
(hence the associated CMP) and have a didactic purpose.  Saving space
wasn't the reason.
 
> I think that we want to use specific symbols for composition of
> permutations, even if they are PRESENTED the same as times, and if
> permutations with concatenations are a semigroup. If we wanted to stress
> this, it would be better to add this as a FMP or CMP.
I agree.  We don't have any concept of inheritance in OpenMath.
 
Mike.

_____________________________________________________________________
This message has been checked for all known viruses by Star Internet
delivered through the MessageLabs Virus Scanning Service. For further
information visit http://www.star.net.uk/stats.asp or alternatively call
Star Internet for details on the Virus Scanning Service.
--
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