[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: time and date


> >I found this useful reference on the web:
> >
> >http://www.ft.uni-erlangen.de/~mskuhn/iso-time.html
> >
> >
> >I suggest following ISO 8601:1988 and adding optional, 
> >arbitrary-precision decimal seconds, e.g.
> >
> >
> >YYYY-MM-DD hh:mm:ss{.s+}
> We used "_" for " " but otherwise this *is* our format.
> Is that one character difference a big deal to you?

I started this discussion, so I'll leap in.  No, it's not a big deal to
me.  I don't care one way or the other.  Now that ISO 8601:1988 has been
brought to my attention I'd prefer it to be followed, but it's rather

> >I question the wisdom of limiting the range of the spec
> >merely in order to prevent people from allowing their
> >precision to exceed their accuracy.
> Here I have to plead guilty.  I spent some of my formative years doing 
> performance analysis with rigorous adherence to Statistics practices, where 
> using more digits in a report than the process precision allowed was an 
> unforgiveable sin.

I'm still strongly in favour of *optional* fractional seconds.  Their
cost is slight, at most, and the flexibility more than makes up for it.
I fear that without them, future generations will curse us and have to
kludge their way around the limitations by encoding the fractional
seconds elsewhere.  If "our" format does indeed read ss{.s+} then the
argument is over, and I've won 8-).  The claim seemed to be that the
optional portion wouldn't actually be there.

(While we're on the subject, did anyone else notice the false claim in
the original article that the YYYY format would allow 9000 years of use?
9999-1997 == 8002)


Version: 2.6.3i
Charset: noconv
Comment: Processed by Mailcrypt 3.4, an Emacs/PGP interface


Follow-Ups: References: