RE: OWL-Time extensions for Era

Thanks Elie - 

In fact this is generally the approach I used in developing an RDF implementation of the geological timescale - see
https://github.com/CGI-IUGS/timescale-ont with documentation at https://raw.githack.com/CGI-IUGS/timescale-ont/master/html/gts.html and https://raw.githack.com/CGI-IUGS/timescale-ont/master/html/thors.html 
The last one is a general-purpose scheme for timescale defined using 'eras'. 

Geological data at http://resource.geosciml.org/classifier/ics/ischart/Precambrian etc 

This has been online at the latter URIs for many years now. The alignment with OWL-Time was done about 3 years ago. 

Simon 

-----Original Message-----
From: Élie Roux <elie.roux@telecom-bretagne.eu> https://raw.githack.com/CGI-IUGS/timescale-ont/master/html/gts.html

Sent: Thursday, 16 July, 2020 01:14
To: public-sdwig@w3.org
Cc: Cox, Simon (L&W, Clayton) <Simon.Cox@csiro.au>; Ted Guild <ted@w3.org>
Subject: Re: OWL-Time extensions for Era

Dear Chris,

Thanks a lot for your answer!

> I was slightly worried by the definition of Common Era as an unbound  interval. The original Allen Algebra of temporal intervals was only for finite intervals, but I think there has been substantial mathematical work to show that, generally, extension to allow left half-infinite, right half-infinite and fully infinite intervals would not cause any problems in practice.

That was a core issue in the design, thanks for raising it!

My understanding of the common use of the word "era" is that it denotes a period of time (a temporal interval) that can be used as a reference to assign numbers to years (year x or era Y). So if we follow that, an era is both:
- a time:Interval (something that actually exists, and which boundaries are known)
- a reference frame used to assign numbers to years

and the two concepts are conflated in one name (era). My initial idea was to be prudent and to differentiate these two concepts, defining time:Era only as the latter. So in my initial idea was that a resource could be both a time:Era and a time:Interval.

Now, I think we could have something like:

time:Era rdfs:subClassOf time:Interval .

so that time:Eras are always intervals. Were you thinking of something like that?

The reason I was uncertain about that is the following one:

I'm sure there are dates indicated in ancient documents that relate semi-imaginary events in eras that we're not sure actually happened...
so how ok is it to encode an imaginary date or interval using OWL-Time? To take an extreme example in a fictional world (Game of Thrones), what about:

ex:date1 a time:GeneralDateTimeDescription
               time:year "0003" ;
               time:era [ a time:Era ;
                               time:reignOf ex:AerysIITargaryen ;
                             ] .

?

> Ted Guild could advise, but I think that the proposal is currently on your GitHub. Perhaps it should be in the W3C GitHub so that issues can be raised, discussed and resolved?

That would work for me

Best,
--
Elie

Received on Wednesday, 15 July 2020 21:17:17 UTC