[local-time-devel] A new release & possible extensions

[Attila Lendvai]

> - Lack of support for non-Unix platforms
> - No separate types for dates and timestamps, no interval type

i think we are dealing with these concepts (with an "uneducated
guessing" warning):
 1) a point in time (this is a local-time struct currently)
 2) a date (i use local-time structs with UTC zone and minimized time part)
 3) intervals (two local-time objects, maybe a simple cons cell? probably a
     struct or a CLOS class would serve better)
 4) duration (elapsed units of time; only secs/usecs or support for
all calendar units?)

i'm not sure about separate date objects. the reason is that a
local-time in the UTC zone and minimized time part works fine with all
the local-time specific arithmethic operators. if we introduced a
separate date object then it would introduce a lot more code unless
there's an easy way to capture the shared logic.

otoh, i can see that sometimes it would be useful to have a standalone
object to be able to dispatch on it. but then i would still keep a
low-level struct (what local-time is currently), and introduce CLOS
objects that use the low-level struct for implementing the higher
level concepts (if we decided to introduce them).

but enough of this rumbling, i should definately read more on the
subject before making deeper comments.

-- 
- attila

"- The truth is that I've been too considerate, and so became
unintentionally cruel...
 - I understand.
 - No, you don't understand! We don't speak the same language!"
(Ingmar Bergman - Smultronstället)