Re: Keyword Arguments

Jeremy Tregunna <[email protected]> writes:

>
> This isn't true. You can write a sort function which iterates over your list,
> applying a message which takes one argument (i.e., a comparator) to one element
> and inserting the next element as its argument. For instance, a use case might
> be:
>
> list(1, 5, 2) sort(<) # Would yield: list(1, 2, 5)

Yes you're right it's however done so  e.g. in Smalltalk. And yes I know
it's done also like that in functional languages, but I think I still
prefer having the arguments "explicit" but without doubt your solution
has some elegance in it. 

It's still an area where I think a to argument block is more
readable. YMMV of course. 


>
> This uses higher order messaging and is entirely possible in Io. This would
> work quite like Io's sortBy() which takes a block and two arguments at present.
> This works great for simple comparators. For more complex comparators I can see
> you needing a different solution however. However, if you always yielded an
> "elements" name in the context sort evaluated its arguments in, or some other
> name, you could write:
>
> list(1, 5, 2) sort(myComplexAction(elements at(0)) < myComplexAction(elements
> at(1)))
>
> And pretty much solve your problem. You can set up symbolic names if you want,
> but it's certainly doable and not so bad either. If it wasn't clear, "elements"
> above is a tuple of two items in the list.
I understand but I honestly dislike having it  that way but you're right
it's doable.

Regards
Friedrich

-- 
Q-Software Solutions GmbH; Sitz: Bruchsal; Registergericht: Mannheim 
Registriernummer: HRB232138; Geschaeftsfuehrer: Friedrich Dominicus