Re: [stack] Survey: Your impressions and thoughts about concatenative languages

On Dec 3, 2007, at 3:24 AM, Manfred Von Thun wrote:

> 1: Language L is concatenative if and only if the concatenation of
> programs P and Q in L computes the composition of the functions
> which P and Q compute. Note: the functions must be unary,
> otherwise composition makes no sense.

Must they really be unary? I think it is possible to have a language  
where functions my return multiple values and may consume multiple  
values. Essentially, functions would have varying numbers of input and  
outout "ports". Two functions could be composed only if the output  
ports of the first lined up with the input ports of the second. I'm  
not sure such a language is especially useful (unless we allow other  
operations more complex then composition -- see languages like Max,  
etc), but it at least seems possible.

- John