RE: [boost] Re: date_time -> ptime

From: fred_at_[hidden]
Date: 2003-09-06 15:17:43

> From: Joel de Guzman [mailto:djowel_at_[hidden]] 
> 
> David Abrahams <dave_at_[hidden]> wrote:
> > Yeah, it does to me too.  The Euro format is a lot less ambiguous.
> > The problem is that if you allow 1970/Jan/1  and 1/Jan/1970 you have
> > an ambiguity problem which you can only sort out at runtime, and
only
> > if the user isn't using years in the first century A.D.
>
> Right. If the user is willing not to use 1..31 A.D., we can support
both.
> Thoughts? I do not have an opinion on this matter. I just need the ET
> to work.

I'd like to keep things as unambiguous as possible. ISO 8601 recommends
separating the parts of dates with an operator - instead of operator /.
Perhaps the format that most closely resembles ISO 8601 should require a
-. (I.e.: 1970-Jan-12.) If there is a need for a US or European-specific
format, that could use the / operator. 

Of course, ISO 8601 wants things to be written 1970-01-12, presumably to
make things language neutral, so there wouldn't actually be more than a
passing resemblance to that standard. I don't see any reason why a C++
library has to be language neutral, though.

-Fred