RE: [w3c/sdw] [time] ISO week number (#726)

Thanks for the clarification Christoph.

The intention of this revision of OWL-Time is in fact to allow the good support for Allen’s interval algebra in OWL-Time to also be available for different temporal reference systems.
My own interest is in non-calendar systems (geologic time, etc) but we hoped that the :hasTRS property would provide a non-prescriptive way to link to whichever set of definitions you prefer.

Numbering of weeks within a year is clearly dependent on all kinds of things, including even the length of the week (the French revolutionary calendar had 10-day weeks!) The note on http://w3c.github.io/sdw/time/#time:week was only added very late, and reading again I see the error. Would this suit:

According to ISO-8601 the first week of the year shall include at least four days. Since Monday is defined as the first day of the week, in practice the first week of the year is the one containing the first Thursday in the year.

Otherwise,

(i)                 Perhaps you could supply a better set of words? Or a reference to where a more comprehensive explanation can be found, so we don’t mislead people with an incorrect summary.

(ii)               Maybe we should also mention other calendar standards to remove the impression that only ISO 8601 counts

(iii)             Else we could remove the note altogether.

Happy to be guided here. I hope this minor error (if it is one) does not dissuade you from enjoying the rest of the document, which mostly just provides a more systematic description of the ontology, the core of which is unchanged from the 2006 version.

Simon Cox

From: Christoph Päper [mailto:notifications@github.com]
Sent: Tuesday, 18 April, 2017 23:04
To: w3c/sdw <sdw@noreply.github.com>
Cc: Subscribed <subscribed@noreply.github.com>
Subject: [w3c/sdw] [time] ISO week number (#726)


I have not yet reviewed all of the Time Ontology in OWL draft, but this note for time:week<https://w3c.github.io/sdw/time/#time:week> is so blatantly wrong that I'm afraid there is still a lot of thorough work to do:

According to ISO-8601, the first Monday in the year is the start of week number 1.

This would only be correct if year referenced the ISO week year instead of the (Gregorian) month year, but then it would basically be a tautology.

  *   Calendar week number 1 contains the first Thursday of January and starts on the preceding Monday.
  *   Calendar week number 1 contains the fourth day of January and starts on this day if it is a Monday, else on the preceding Monday.
  *   The first Monday in January either starts calendar week number 1 or 2, according to ISO 8601.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub<https://github.com/w3c/sdw/issues/726>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AAlIL-aJPAEMXOhQRVAkXXn3pk9pw2cKks5rxLTJgaJpZM4NAUJX>.