A link to times: time of day
Andy Turner
What options are there for links for specific times of day? For
example 9AM GMT might be along the lines of:
http://data.gov.uk/time/09/00/00/
I don't want a link for a particular date, just for the time of day.
So for instance the following is a link for 9AM GMT this morning:
http://www.timeanddate.com/worldclock/fixedtime.html?iso=20110919T0900
Which alternatively might be provided as:
http://data.gov.uk/time/date/2011/09/19/09/00/00/
This pattern above is extendible to fractions of a second if for some
reason this was needed...
AFAICT, the worldclock service
(http://www.timeanddate.com/worldclock/) is not configured to allow
wildcards, so for instance the following does not work:
http://www.timeanddate.com/worldclock/fixedtime.html?iso=********T0900
I hope you can help. Sorry if this has been discussed before and there
is a solution I should have found but missed.
Thanks,
Andy
Melvin Carvalho
I could be wrong but ...
I think you need xsd^^time ( http://www.w3.org/2001/XMLSchema#time )
For a time in GMT/UTC, simply add a "Z" following the time:
09:00:00Z.
>
> Thanks,
>
> Andy
>
Melvin Carvalho
The full literal should read:
"09:00:00Z"^^xsd:time
There's a great example in good relations:
http://www.heppnetz.de/ontologies/goodrelations/v1#opens
>
>>
>> Thanks,
>>
>> Andy
>>
>
Tim Hodson
If you dereference the following link, you will get a whole heap of info including hours that are part of that day.
However you won't get a generic time as uri, because reference.data.gov.uk is modelled for time as a real thing and not as an abstract thing.
http://reference.data.gov.uk/id/day/2011-09-19
I think I got the date format right there. Typing uris on a phone is not my favourite thing.
Tim
Tim Hodson
Technical Consultant
http://consulting.talis.com
Stuart Williams
http://www.epimorphics.com/web/wiki/using-interval-set-uris-statistical-data
BR
Stuart Williams
--
Epimorphics Ltd www.epimorphics.com
Court Lodge, 105 High Street, Portishead, Bristol BS20 6PT
Tel: 01275 399069
Epimorphics Ltd. is a limited company registered in England (number 7016688)
Registered address: Court Lodge, 105 High Street, Portishead, Bristol BS20 6PT, UK
Andy Turner
looking for, but then I find thinking about time a little tricky in a
linked data way.
I appreciate that there has to be some granularity for links to specific times.
Initially I was asking for links for an on the hour (o'clock) hour of
a day and of course, there are only 24 of these. However I really
wanted this on the granularity of minutes. There are of course 60
minutes in any hour and 24*60=1440 in any day. Departure times and
opening/closing times are often regular and given as times with minute
granularity.
I can't think of a good example yet, but this also extends to seconds
for which of course there are 60 in any minute, 60*60=3600 in any hour
and 24*60*60=86400 in a day.
At least for minutes, it is not many links for an authority to provide.
I am happy to use Wikipedia to link to days of the week...
Any further thoughts/recommendations for this before I create my own links?
I am thinking geographically and that if there were such links it
would be easier to link coincidence in space and time...
Cheers,
Andy
Stuart Williams
> Thanks for your messages and links. I still haven't found what I'm
> looking for, but then I find thinking about time a little tricky in a
> linked data way.
>
tell you.
The original concept was to provide reference URI for reporting intervals like
years, half-years, quarters, weeks etc. where you might want to be aligning data
for nominally the same intervals - and where the fine-grained accuracy of when
and intervals starts/ends is not too much of an issue. However, I was asked to
go below the week level to days, hours mins and seconds (as intervals).
You'll find that there are also instants down to 1 second granularity.
I also made a deliberate decision *not* to time zone the intervals - and to
'duck' on the question of DST or not - so times are in a sense 'local'.
Given the intent as providing interval dimensions for cube like stats/reporting
data that seemed like the right choice - 2011-Q2 doesn't seem to me to require
the precision of time-zones or DST handling.
> I appreciate that there has to be some granularity for links to specific times.
>
> Initially I was asking for links for an on the hour (o'clock) hour of
> a day and of course, there are only 24 of these.
Actually there are an infinite number of hour long intervals in a day - but only
24 that are what I've call calendar aligned... and even that's not true, there
can be up to 25 and as few as 23... and even that...
[feels a sense of deja-vu coming on and head done in'ed-ness.]
> However I really
> wanted this on the granularity of minutes. There are of course 60
> minutes in any hour and 24*60=1440 in any day.
Mostly but not always...
> Departure times and
> opening/closing times are often regular and given as times with minute
> granularity.
Now we're in space where time-zone and DST may matter - particularly where
journey legs cross zones.
You also raise the spectre of recurring intervals - like say "Monday". And in
practice you may want to capture when such recurrence both starts and ends (and
you may not know the latter until afterwards - and would you want knowing that
to affect the URI identity of an interval).
>
> I can't think of a good example yet, but this also extends to seconds
> for which of course there are 60 in any minute, 60*60=3600 in any hour
> and 24*60*60=86400 in a day.
Again mostly... but it is always the exceptions that bite (except when they
don't :-))
>
> At least for minutes, it is not many links for an authority to provide.
>
> I am happy to use Wikipedia to link to days of the week...
>
> Any further thoughts/recommendations for this before I create my own links?
I think that link-data works best when folks avoid proliferating multiple URI
names for the same thing. I think that it is inevitable that things will have
multiple URI names... but I would like for the natural forces of usage to be
such that some small number (say <3) of URI for any given thing come to dominate
usage and the remainder be come redundant and ultimately irrelevant.
So... *if* there are stable, persistent, sets of URI out there that meet your
need (and I'm not saying that the reference.data.gov.uk URI are such), I'd urge
re-use over re-invention.
I wouldn't wish sameAs link maintenance of interval and instant relate URI on
anyone :-).
>
> I am thinking geographically and that if there were such links it
> would be easier to link coincidence in space and time...
>
For zoned intervals, instants and geospatial points you might also look at Ian
Davis' http://placetime.com site.
> Cheers,
>
> Andy
Stuart
--
Andy Mabbett
> Now we're in space where time-zone and DST may matter
While not directly usable in the case in hand, such issues are also
addressed by 'Extended Date Time Format efforts':
http://www.loc.gov/standards/datetime/
Some of whose solutions may be transferable (or with which it may be
prudent to maintain compatibility).
--
Andy Mabbett
@pigsonthewing
http://pigsonthewing.org.uk
Andy Turner
I had not really considered the pain of non 24 hour days Stuart. It
does seem that to clearly specify time, it is necessary to specify
spatial location. Because of time zones though, I expect there are
only a finite number of official times in the world and these official
times are synchronized with one another (hours apart I expect).
I'm not sure though and now would not be surprised to learn about
exceptions and that there has been the weirdness of days with the same
time more than twice at the same location!
I will keep looking before creating my own set of links for times of
day. I still wonder though for example what is the best way to link to
the following time and whether I am right to even think about using a
link:
9:30am to 4:45pm weekdays (Monday to Friday).
Although the answer may be contained in this thread, I have not
realised it yet. Maybe one day I'll come back to this and find the
answer.
Bye for now,
Andy
Stuart Williams
Szemere Szemere
Szemere
Brian Hoadley
Brian Hoadley
entrepreneur. innovator. writer
Michael Saunby
Andy Turner
Sorry, I got a reply from Mark Linehan (with others copied in as per
this reply) and I just replied to Mark who is listed as an author of a
Date-Time Vocabulary (Date-Time) Object Management Group (OMG)
document which he shared with me and those others, but I don't think
it went to the uk-government-...@googlegroups.com address.
That OMG document is a draft all about date and time issues. However
when I skimmed it, I did not see it dealing with time interval
intersections or nesting of intervals as Stuart commented on back on
this thread. Maybe it should? I copy the gist of my email reply to
Mark below. In the latter part of my reply when talking about all
time, that should probably be considered as all time within the
context of a particular calendar.
Stuart raised the concept of nested intervals with regards my
particular use case. The example I gave does have an explicit nesting
i.e. within Monday to Friday I want to specify 9:30am until 4:45pm. (I
turned this around so that the broadest interval is first.) So, to
specify only for a set period in all time (Stuart's example was first
6 months in 2010) I can restrict to between two date intervals. Nested
intervals seems a good concept to me.
I see it that a major issue is with setting the time origin (time_0)
in each standard calendar default implementation. doing this should
allow us to get the time and date (and day of the week etc) as
specified in each calendar for a particular instant relative to
time_0. I expect this has all been thought through before, such as,
for the Java implementations of Calendar that I have used before. (The
Java source code contains documentation about the implementations
which might be useful to others.)
I presume that most computer operating systems have a clock and have
set their time_0 and that there is an official time, relative to
time_0, for representing now (time_now). For each operating system I
expect there is an official time for time_now that is available
on-line and which is used for computer clock synchronization and
informing users that have set their local what date and time it is.
(In my Java programming I sometimes make a system call to get the
current time in milliseconds, but because I am not modelling real time
things I really only use this for timing computational execution.)
Is there a default time interval? What I mean is that if there is no
interval set, do we assume a zero interval or the infinite interval? I
would have thought it easiest if in specifying intervals that if
nothing is specified, the default is for the infinite interval (all
periods). So, 9:30am until 4:45pm would be for every day there has
ever been. Monday to Friday 9:30am until 4:45pm would be for every
week there has ever been. Likewise, if I specify the first Tuesday of
every Month 9:30am until 4:45pm, this is for every month there has
ever been. Clearly we can restrict using a date interval or specifying
a start time and end time (each given relative to our standard time_0
reference point).
Maybe that helps, maybe not. Thanks in advance for any further feedback.
Andy
andy
Regarding timing specifications, can I suggest you have a look at some
pre-existing standards?
For example (ie one that I know), HL7/ISO have looked at these complex
timing issues already in great detail.
If you go to http://www.hl7.org/v3ballot2011may/html/welcome/environment/index.html
and navigate to
Foundation/Data Types: Abstract R2/Quantities/
there are detailed definitions for PointInTime,Calendar and
CalendarCycle which I'm sure will be of use.
The Foundation/Data Types: Abstract R2/Collections of Quantities/
section also discusses time intervals, periodicity and a general
timing specification.
Hope this helps.
Andy Harris
>
> >> I'm in one now. Bangalore is 4.5 hours ahead of the UK.
>
> >> > "hours apart" - no, some time zones are half an hour apart!
>
> >> > Szemere
>
>
> >> >> Thanks again.
>
> >> >> I had not really considered the pain of non 24 hour days Stuart. It
> >> >> does seem that to clearly specify time, it is necessary to specify
> >> >> spatial location. Because of time zones though, I expect there are
> >> >> only a finite number of official times in the world and these official
> >> >> times are synchronized with one another (hours apart I expect).
>
> >> >> I'm not sure though and now would not be surprised to learn about
> >> >> exceptions and that there has been the weirdness of days with the same
> >> >> time more than twice at the same location!
>
> >> >> I will keep looking before creating my own set of links for times of
> >> >> day. I still wonder though for example what is the best way to link to
> >> >> the following time and whether I am right to even think about using a
> >> >> link:
> >> >> 9:30am to 4:45pm weekdays (Monday to Friday).
>
> >> >> Although the answer may be contained in this thread, I have not
> >> >> realised it yet. Maybe one day I'll come back to this and find the
> >> >> answer.
>
> >> >> Bye for now,
>
> >> >> Andy
>
> >> >> On Wed, Sep 21, 2011 at 4:08 PM, Andy Mabbett
> >> >>> On 21 September 2011 14:58, Stuart Williams <skwli...@gmail.com>
>
> >> >>>> Now we're in space where time-zone and DST may matter
>
> >> >>> While not directly usable in the case in hand, such issues are also
> >> >>> addressed by 'Extended Date Time Format efforts':
>
> >> >>> http://www.loc.gov/standards/datetime/
>
> >> >>> Some of whose solutions may be transferable (or with which it may be
> >> >>> prudent to maintain compatibility).
>
> >> >>> --
> >> >>> Andy Mabbett
> >> >>> @pigsonthewing
> >> >>>http://pigsonthewing.org.uk
>
> >> --
> >> Brian Hoadley
> >> entrepreneur. innovator. writer
>
> >> 8 Briar Wood Close, Bromley, Kent BR2 8EJ
> >>www.brianhoadley.com|twitter. @brianhoadley