#599 maplist inconsistent with args

[Stavros Macrakis]

maplist(foo,a[i]) => error

while

map(foo,args(a[i])) => [foo(i)]

I'd think these should do the same thing, especially
since maplist does act like a "part" function in other
ways (e.g. in respecting inflag).

-s

[Robert Dodier]

Logged In: YES
user_id=501686

In 5.9.3cvs, maplist(foo, a[i]) => error and also map(foo,
a[i]) => error.

maplist(foo, a[i]) should yield [foo(i)] as suggested above.
map(foo, a[i]) should yield a[foo(i)] (by analogy with
map(foo, a(i))).

[Kris Katterjohn]

I agree that this seems inconsistent:

(%i1) map (foo, bar (i, j, k));
(%o1) bar(foo(i),foo(j),foo(k))

(%i2) map (foo, bar [i, j, k]);
map: improper argument: bar[i,j,k]
 -- an error. To debug this try: debugmode(true);

But since bar[i,j,k] is a mapatom, isn't this the expected behavior? To change this we would need to change the definition of a mapatom.

Was changing mapatom implicit in the previous comments, or...?

[Kris Katterjohn]

Hi Stavros, Robert. Can we close this as not-a-bug based on my comment from 5 years ago?

From what I see, Stavros was initially suggesting changing maplist only. Then Robert generalized this to other mapping functions. It looks like my comment from 5 years ago was based mostly on Robert's comment.

I do see Stavros' point about the inconsistency, but I think maplist being different than other mapping functions w.r.t. mapatoms is also inconsistent.

I think we should keep the current behavior.

[Kris Katterjohn]

I left my first comment in this ticket over 5 years ago and I left another about 6 months ago.

I'll close this ticket as not-a-bug soon (probably next week) unless anyone wants to comment in favor of the original proposal (about maplist only) and/or the follow-up proposal (about other mapping functions too).

[Kris Katterjohn]

Closing as not-a-bug, per my previous comments.

[Robert Dodier]

Haha, it's funny to think I first encountered this bug report almost 15 years ago ... how time fled.

Anyway I agree that changing the behavior of map(foo, a[i]) implies a change in the definition of mapatom, which I guess I'm not in favor of. So it makes sense at this point to close this ticket as not a bug.