W3C home > Mailing lists > Public > public-sdwig@w3.org > July 2020

RE: OWL-Time extensions for Era

From: Little, Chris <chris.little@metoffice.gov.uk>
Date: Tue, 21 Jul 2020 12:14:00 +0000
To: "Cox, Simon (L&W, Clayton)" <Simon.Cox@csiro.au>, Élie Roux <elie.roux@telecom-bretagne.eu>, "Ronald Henry Tse (tse@ribose.com)" <tse@ribose.com>
CC: "public-sdwig@w3.org" <public-sdwig@w3.org>
Message-ID: <CWXP265MB00535DD9D0EDA9ACF5E532D1A7780@CWXP265MB0053.GBRP265.PROD.OUTLOOK.COM>
Hi Simon,

A few years ago (well, 1988, ?) ISO8601 used to allow the notation of intervals (which they call periods) that started or ended with 'now'. This option was removed in later versions of the standard. 

At the time, lots of people (well, some programmers) were annoyed. Do you know why and how that decision was taken? I think it may be relevant to this discussion and perhaps indicates caution. I've copied Ron in case he knows the history.

Chris (now living in a temporal regime of seasons only with a free-running body clock ;-)

PS Recently, and rather less controversially, the term 'midnight' for time 00:00 UTC was also removed from ISO8601!


-----Original Message-----
From: Cox, Simon (L&W, Clayton) <Simon.Cox@csiro.au> 
Sent: 21 July 2020 06:15
To: Élie Roux <elie.roux@telecom-bretagne.eu>
Cc: public-sdwig@w3.org
Subject: RE: OWL-Time extensions for Era

If 'unknown' and 'now' are constructed as temporal positions, not instants, then you could go with 

```turtle
time:Indeterminate a time:TemporalPosition . 
time:Now a time:TemporalPosition .

ex:interval1 a time:Interval ;
	time:hasBeginning [ a time:Instant ;
		time:inTemporalPosition time:Now ] ;
	time:hasEnd [ a time:Instant ; 
		time:inTemporalPosition time:Indeterminate ] ;.
```
Etc. I think those two will do. 

> -----Original Message-----
> From: Élie Roux <elie.roux@telecom-bretagne.eu>
> Sent: Monday, 20 July, 2020 17:07
> To: Cox, Simon (L&W, Clayton) <Simon.Cox@csiro.au>
> Cc: public-sdwig@w3.org
> Subject: Re: OWL-Time extensions for Era
> 
> > I'm suggesting that four canonical individual time:Instants be 
> > defined as
> part of OWL-Time.
> > Their URIs can then be used as the limits of specific intervals (and 
> > eras)
> when required.
> 
> Ah ok I understand, thanks! As part of defining a RDF vocabulary I 
> encountered a problem with unknown entities. Basically if you have:
> 
> ex:interval1 time:end time:Unknown .
> ex:interval2 time:end time:Unknown .
> 
> RDF logic says that ex:interval1 and ex:interval2 end at the same 
> instant, which is probably not right. On the other hand this kind of 
> pattern works
> better:
> 
> ex:interval1 time:end [ time:after ex:year2020 ] .
> ex:interval2 time:end [ time:after ex:year2020 ] .
> 
> that way RDF won't infer that the two instants are the same...
> 
> What do you think?
> 
> Best,
> --
> Elie
Received on Tuesday, 21 July 2020 12:14:16 UTC

This archive was generated by hypermail 2.4.0 : Tuesday, 21 July 2020 12:14:18 UTC