- From: Evan Wallace <ewallace@cme.nist.gov>
- Date: Fri, 25 Jul 2008 15:44:13 -0400
- To: Uli Sattler <sattler@cs.man.ac.uk>
- CC: public-owl-wg <public-owl-wg@w3.org>
> On 25 Jul 2008, at 13:33, Deborah L. McGuinness wrote: > >> >> ps. if i do not have a time zone, i can look for the location of the >> instrument / observatory that generated it and then figure out time >> zone based on lat / long and date of instrument. >> we currently do not have moving instruments yet. >> > > good - so if you were to do this, you could (with no overhead to speak > of), > > - instead of adding the timezone information > - convert the time in the 'owl-standard-time-zone-time' > > Hence not having time zones shouldn't be a huge problem?! > > Cheers, Uli Two things: 1) I don't understand why supporting Times both with and without timezone offsets is such a burden. They would be two different datatypes, the value mapping of one gets normalized to UTC and the other doesn't. In the latter case the user has to worry about what is and isn't comparable, but users are well aware of that issue. 2) There is a problem with the approach Deb described for inferring mappings to UTC. When we fall back from summer time (daylight saving time in the US) to standard time, the same local time will repeat, thus the pair [location , Local DateTime] doesn't uniquely determine a time on the UTC ISO8601:Time Scale. If we can only have one of these options, I would vastly prefer requiring a timezone offset per 8601/XSD, that way unambiguous timestamps can be used where the Time of Day component is in local time. This would be user friendly. Implied in Uli's comment is that neither option is supported and all times must be in 'owl-standard-time-zone-time' where I assume that the Time of Day will correspond to UTC, so I guess only the "Z" form would be supported. I would prefer that to having no native time support at all, but then it seems that you have done the hard part in DateTime which is handling the irregular calendar intervals, while punting on an easier feature that would make the language considerably friendlier to users. -Evan
Received on Friday, 25 July 2008 19:45:01 UTC