On Fri, Jan 22, 2016 at 10:34 AM, John Cowan <xxxxxx@mercury.ccil.org> wrote:

Shiro Kawai scripsit:

The logic is that I don't want people to say "Lists have the richest set
of library routines, so I'll just use lists because they are the easiest
to use", at the expense of contorting their data structures. By making
all libraries equally rich, the choice of data structure can follow other
criteria.

I see. I'd imagine the ideal is something like STL.

The difficulty is that we define it one by one so it's tricky to keep

the consistency. But if we all know the goal, we can manage it, I think.

I try. There are two forces pushing me toward incompatibility: One is
that sometimes the old definitions are just bad: not having a failure
thunk for SRFI 1 find and returning #f was a mistake. There's a
workaround for this mistake, so it's not fatal, but it's discouraging.

I agree that 'find' was mistake, albeit useful for most situations.

I'd like to have a proper one, but I wish it be named differently;

search? locate?

If you see particular operations that seem to be missing, please let me know.

Now I understand the motivation, I'll try to adapt Gauche's to the draft and

see if I can find anything missing.