[BUG] Timspan starting: Date today duration: 0 "gets negative duration" - 1 attachment

Avi Bryant avi at beta4.com
Thu Sep 23 09:46:26 UTC 2004


On Sep 23, 2004, at 4:15 AM, Ingo Hohmann wrote:

> Hi,
>
> I know that this is more like a philosophy problem:
> Chronologys philosophy is to _not_ include the end
> of the timespan. While I really like the Chronology
> work I don't agree with its philosophy.
>
> Attached is an SUnit test, which shows the problems
> I see.

Yes, I complained about essentially the same thing in the thread  
starting here:

http://lists.squeakfoundation.org/pipermail/squeak-dev/2004-February/ 
074397.html

I still think this is a bug.  My last comment on the philosophical  
difference was this:

" I have no problem with 1pm to 2pm not including 2pm, or with some  
method
like #last returning 1:59:59.  However, the #end of [1pm..2pm) is 2pm,
both intuitively and mathematically (time is not discrete, computer or
no, and you can't properly describe an exclusive range by the last
value it does include, you need to describe it in the limit of the
first value it doesn't).  As evidence of the intuition here, you need
only look at your own class side protocol - #starting:ending: takes as
its second parameter the first time that is not in the range, not the
last time that is.  IMO it's dangerously inconsistent to use the word
"end" in both of those places and give it different meanings."

I don't believe I ever got a reply.

Avi




More information about the Squeak-dev mailing list