Time Zones

We had a workshop last week on gist (our minimalist upper ontology, more on that in another post maybe next week).  Anyway as part of the aftermath, I decided to get a bit more rigorous about some of the lowest level primitives. 

One of the basic ideas about gist is that you may not be able to express every distinction you might want to make, but at least what you do exchange through gist will be understood and unambiguous.  In the previous version of gist I had some low level concepts, like distance, which was a subtype of magnitude.  And there was a class distanceUnit which was a subclass of unitOfMeasure.  And unit of measure has a property that points to conversion factor (ie how to convert from one unit of measure to the base unit of that "dimension") 

But what occurred to me just after the workshop is that two applications or two organizations communicating through gist could still create problems by just picking a different base (ie is one said their base for distance was a meter and another foot they have a problem).  This was pretty easily solved by going to NIST, and getting the best thinking on what these dimensions should be and what the base unit of each dimension should be.  Looking at it, I don't think there ought to be much problem with people adopting these.

Emboldened, I thought I would do the same for time.  For starters universal time seems to the way to go.  However many applications record time in local time so we need some facility to recognize that and provide an offset. 

Here's where the problem came in and maybe you dear readers can help. After about an hour of searching the web the best I could find for a standard in this area is something called the tz database.  And while you can look up various cities I didn't see anything definative on what the geographical regions are that make up each of the time zones (and to make things worse the abbreviations for time zones are not unique, there is an EST in North America and one in Australia for instance).

So if anyone has a thought in this area, I'm all ears.

Comments

[make-money]

Love to read semantic-conference.blogs.com
make money with link

[CPA-network]

I come back to semantic-conference.blogs.com every single day!!
cpa network

[pnr status]

great post thanks

[website development]

Your blog keeps getting better and better! keep it up!

[Plano Alarms]

Nice Article, thank you for sharing, and so much for this information.

[Plano Alarms]

Very cool post. I just recently started following your blog, but I look forward to contributing more in the future .

[Atlanta Alarms]

Many thanks for this brilliant post! Many points have extremely useful. Hopefully you'll continue sharing your knowledge around.

[website design]

Hello ...cheers for the very informative post .I'm really happy I found it on big Keep up the wonderful work because I for sure will check it out for updates

[rugs]

You will need to really seem into the new upgrades offered for BE blogs. I believe yours could really benefit from it.

[Antivirus_man]

You write well will be waiting for your new publications.

[RamonGustav]

Have you been turned down by other lenders?

[brandon]

nasa map & info

[brandon]

lol cia map, maybe this could help Link text

[Nicicx]

I’m surprised because of the good enough topic connecting to this good post. You made it like the freelance writing jobs , hence, you deserve really prestige grant for this, I thougt!

[Professional PowerPoint Presentation]

Hi,
Thank you for sharing information in the blog. You are really doing a good work. I personally like this blog and appreciate your efforts.

Professional PowerPoint Presentation

[andy33]

Did you have a knowledge that we would provide with you excellent research . We promise that men buy custom essay papers and get the experienced custom essays writing companies to buy term papers at.
http://www.bestwritingservice.com

[Amit Gupta]

Your post was interesting precicely because of that reason: it highlighted that the Information emperor wears many accessories (technologies & processes de jure) but few undergarments...

The 'somebody' who makes up timezones is probably comprised of 195 bodies (i.e. the number of countries on earth; UN recognizes 192). Just as likely, there is no globally recognized body that registers/consolidates this data (other than likes of TZ / CIA / Microsoft etc.)

China had five timezones until 1949, but then changed to just one; the process that gave rise to this was the Chinese Civil War. In the summer of 2007, we applied time zone updates to our computers: The process that gave rize to that was the US Energy Policy Act of 2005 (lobbyists for this provision included Sporting Goods Manufacturers,Convenience Stores, and the National Retinitis Pigmentosa Foundation Fighting Blindness; lobbyists against included the U.S. Conference of Catholic Bishops, the United Synagogue of Conservative Judaism, the National Parent-Teacher Association, the Calendaring and Scheduling Consortium, the Edison Electric Institute, and the Air Transport Association).... So much for the process. :-)

Who is the first to know when a country changes it timezones? -- Probably Microsoft (via people in that country calling the helpline for updates). Geolocations for country boundaries are more fundamental and less fluid, but just as hard to find: e.g. where exactly does China end?

PS. Great blog; very thought provoking! Am still pondering over Categories & Classes post...

[Dave McComb]

Amit, yes, this is the type of information I was looking for. I'm out of my depth here (haven't spent much time working with or thinking about time zones) but was surprised that the consensus wasn't more obvious). What you suggest (extracting the coordinates of the time zone geography from the CIA map) is interesting, but begs the broader question: where did the CIA get their data from. They don't get to make up time zones (I hope). But somebody does. Those red lines on the CIA map describe agreed upon non overlapping and fully covering georegions on the surface of the earth (ie no matter where you are you are in exactly one time zone). Would like to know what process gives rise to that, and who has the geo coordinates.

PS I love maps, and I especially love looking at maps like this and thinking: why did the inhabitants of Chatham Island decide they needed their own time zone 12 3/4 hours off of GM, and were they trying to be on the other side of the date line, and did they experience Y2K before anyone else, and did they notice.

Anyway, thanks Amit.

[On behalf of Amit Gupta]

[for some reason this comment got blocked when Amit tried it initially, let me see if it works if I put it in. DMc]

Not sure if this is at the level of abstraction that your post is interested in, but here are some thoughts: The CIA time-zone map has a visual representation that maps time zones to longitudes/latitudes (a more readable image here)... if the CIA doesnt share the underlying data, it can be extracted by processing the image programmatically (might also be an interesting project for manual transcoding via Amazon Mechanical Turk). Wikipedia has a list that maps latitude and longitude to timezones, but unlike the CIA's image it points to the zone's central location only, not bounds. As for zone names/codes (the EST in NA vs EST in Australia issue): name ambiguity is a known issue... for example, Java 2 docs say "three-letter time zone IDs (such as PST, CTT, AST) are also supported... However, their use is deprecated because the same abbreviation is often used for multiple time zones (for example, CST could be U.S. Central Standard Time and China Standard Time), and the Java platform can then only recognize one of them." Finally, I think that the non-Windows standard for zone names/IDs is indeed the TZ database (e.g. see Google's timeZoneId enumeration, which seems to be using the same names as the Tz database; Linux does the same); Windows stores all time zone data in the registry under HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Time Zones; Windows' vs TZ DB timezone names are mapped here.