Re: Converting seconds past midnight to a time

On Sun, Dec 18, 2005 at 07:29:45AM +0000, Karl O. Pinc wrote:
>
> On 12/17/2005 10:21:39 PM, Michael Glaesemann wrote:
> >
> >On Dec 18, 2005, at 13:25 , Karl O. Pinc wrote:
> >>On a related note is there some reason why
> >>interval + int
> >>does not result in the interval plus int number
> >>of seconds?
> >
> >Why should the int necessarily represent seconds and not some other
> >amount of time? It's just a unit-less value.
>
> Good question. I guess it's because I couldn't corece
> an int into an interval number of seconds. ;-) But if it was to be
> anything it should be seconds or milliseconds as those
> are the only ones that make the math easy, in the sense of
> working with whole numbers.

Well, generally speaking people work with time either in some native
format (ie: timestamp) or as a number of seconds. I don't think there's
enough consistency between time being a number of milliseconds or
microseconds to warrant that.

Personally, I don't think it would be unreasonable to allow timestamp +
int and timestamp + double (with int and double being treated as
seconds). I don't recall ever seeing an email on the lists from someone
expecting time/timestamp + bare number to mean 'add X hours' or 'add X
fractional seconds', but people do ask about adding X seconds pretty
often.

Another option would be creating a set of timestamp math functions; that
would probably help cut down on the number of questions about this.
--
Jim C. Nasby, Sr. Engineering Consultant jnasby(at)pervasive(dot)com
Pervasive Software http://pervasive.com work: 512-231-6117
vcard: http://jim.nasby.net/pervasive.vcf cell: 512-569-9461