[ic] time tag with adjust in February bug?

Peter peter at pajamian.dhs.org
Mon Apr 1 22:49:48 UTC 2013


On 04/02/2013 05:13 AM, Mike Heins wrote:
> Quoting Paul Jordan (paul at gishnetwork.com):
>>
>> B) On January 31st, I ask you to meet me here one month from today,
>> likewise, you don't add any amount of days, but rather meet them at the end
>> of February.

Actually, as a human I would probably just ask you to clarify.

> That is your definition. It's fine, but you are basing it on assumptions
> which are not universal.

I agree, but that said, it does seem to be the most reasonable 
definition that I've heard for that scenario, and certainly more 
reasonable than what the code currently comes up with.  I think we 
should change the behavior of the code to do this instead of what is 
currently in place, and we should document it as such.

>> I am just saying, if we keep the "month" usage in the tag, it should work
>> how the vast majority of humans would expect it.
>
> I am not in favor of keeping it, because "vast majority" is not precise
> enough for computer language.

I'm only in favor of keeping it because I have code that uses it, and 
because it does not detriment from existing code that does not use it. 
Any code that does not use month or year is not affected by this at all. 
  I'm not opposed to changing it to be a bit more reasonable irt how end 
of month and leap day conversions are handled, though.


Peter



More information about the interchange-users mailing list