Different max behavior from commercial product.

I would prefer to take inspiration from the definition
in Common Lisp.

(max )  ;; i.e. no arguments at all --> error.

If you want to use some particular ordering, I suppose
something like dominate(theTransitiveBinaryPredicate,theList)
would take all pairs from theList and find the winner.

Then dominate(">", thelist)

would be max.

RJF

Stavros Macrakis wrote:

>>The commercial product returns minf when max is either called
>>with a empty list, or applied to one, or called with no arguments.
>>If called with a non empty list, it applies max to the list.
> 
> 
> I agree that max() == apply(max,[]) should be minf.
> 
> What does commercial Macsyma do with max([2,3])?  with
> max([2,3],[3,2])?  I do not think it is a good idea for max([...]) to
> be equivalent to max(...) because there are various possible orderings
> you could apply to a []-list: lexicographic, max of various norms
> (which may not yield a unique max, of course).
> 
> In general, I am opposed to features which sacrifice mathematical
> consistency for supposed convenience.  Experience shows that the lack
> of consistency always causes problems later on.
> 
>              -s
> 
> _______________________________________________
> Maxima mailing list
> Maxima@www.math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima