Re: A missing feature of APL expressed in the Wikipedia

Somewhat thinking aloud here......

For "name" indexing to work it seems that four related capabilities would be
needed

a) Associate a value with a name

b) Retrieve a value by its name

c) Delete a name and its associated value

d) Obtain a list of names currently in use

All of these things can be accomplished with the namespace as implemented in
Dyalog APL, albeit with a smidgen of coding. Such coding however can be
embedded in defined functions (hopefully dynamic!) so that the user can thin
k
of these capabilities as being literally native.


-- James L. Ryan -- TaliesinSoft

TaliesinSoft a écrit :

> One could always produce the list from d and then find out if a given name
> (key) is present in it. But just being able to ask could be more convenient.[/colo
r]

Yes, and certainly more *efficient* and *readable* too. I do appreciate
being able to write "2+/A" in APL, but I feel very miserable when I have
to write things like "2 take X[gradeup X]" - for instance when buidling
a Huffman code or doing some data aggregation - when I think about the
unefficiency involved under the hood.