OpenEuphoria

Re: Proposal for 'math.e' (2007-08-23)

• Posted by Jules <jdavy at dsl.pi?ex.c?m> Aug 24, 2007
• 693 views

Derek Parnell wrote:
> 
> Jules wrote:
> > 
> > > Routines for sets should be in a separate include file, not in "math.e".
> > > 
> > > Regards,
> > >    Juergen
> > 
> > I'm not going to make a big fuss about it, but why? sets are the foundation
> > of all mathematical systems, underlying all functions, relations, and logic.
> 
> That they may be, but the file will contain set OPERATIONS, and they are not
> commonly used. In other words, just because sets are the basis of
> mathematically
> systems it doesn't follow that therefore set operations are used all the time.

they are, but you just don't realise it. :)

In fact, you mentioned set operations in a recent post in this thread, replying
to CChris:
--------------------------------------------------------------------------
"CChris wrote:
> I have a data series which decreases, then increases, then decreases even
> more,
> then increases less than before and so on. Think of global winter and summer
> rainfall over the long run, for instance.
> Then I'm interested not only in the absolute maximum (which start=1 will
> return),
> but also the local maxima further down the road, for which I must filter the
> first points in the series.

Got it. You have a set of data which contains various intersecting and
non-intersecting
subsets. Such as annual rainfall which has various subsets for months, seasons,
and ad hoc so one can ask questions such as ...

   highest rainfall for the year, for March, for Summer, since "that dust storm
back on July 17th", etc ..."
---------------------------------------------------------------------------

I think you're taking too narrow and limited a view of set operations.
Euphoria's power and flexibility comes from sequences, which can be regarded as
sets (collections of objects). If you look at the descriptions of the set
operations in sets.e and mentally replace each occurrence of 'set' with
'sequence' you'll get a better idea of how useful they can be, and not just in
computational applications. In fact maybe that's their 'weakness', they're so
generic we take them for granted.
Take 'Intersection' and 'Union' for example. These are closely related to 'and'
and 'or' respectively, but are more powerful in the sense that they return the
actual sequence elements, not just 1's and 0's.