Morten Høgholm wrote:

>> Question 1. The various counter-creation macros in expl3 work like
>> \newcount.  There doesn't seem to be anything like \newcounter, for
>> creating something with a user-facing \the<counter>.  Should counters
>> that go near the user stick with \newcounter for the moment, or should
>> one define \c@<counter> and \the<counter> "by hand" so that things look
>> the same for the user?  I guess this is partly a philosophical question!
> 
> LaTeX counters (\newcounter) are document level. The expl3 counters
> are so far low-level only. I think it is important that you write your
> low-level code separate from the user interface, e.g., use xparse or
> similar to map between document usage and low-level.

Okay, I shall "try the exercise again"!

>> Question 2. Why is \clist_map_break:w "weird"?  It doesn't take an
>> argument from a use point of view, so I'd have expected \clist_map_break:.
> 
> I think the original reasoning was that the user had to know exactly
> where it could be used (\if level) but that of course applies to many
> things... I recently added (but did not yet check in I think) two
> variant forms, namely, as you say, \clist_map_break: and also
> \clist_map_break:n. The latter will carry an argument to the end of
> the loop, typically something involving the last value found. Same for
> other loop types. The l3prg case-switches use something like this.

This seems to make sense.

>> Do tell me to stop if I'm asking too much!
> 
> Not at all. The reason expl3 has grown so much over the past years is
> because it was discussed and used. So keep 'em coming!

Will do!
-- 
Joseph Wright