Tyler Bell On Yahoo's Open Location API

blackbearnh writes "Yahoo! has been working for a while to promote a unified system for referring to places, through their Where On Earth IDs. Using a WOEID, you can query Yahoo's publicly available APIs to find out things like what cities are in a county, or what counties border each other. In an interview for O'Reilly Radar, Tyler Bell, the product lead for the Yahoo Geo Technology Group, talks about their Open Location program (not to be confused with openlocation.org, a different group altogether). He also talks about how privacy concerns interact with the increasing use of personal geotracking, and the troublesome problem of what to call places. 'I'm not even going to tell you about the problems we had when we accidentally called Constantinople Byzantium, just slipping back about 800 years there accidentally. That's a very sensitive issue. Any company dealing with geography is going to have to address it somehow. So I'll be very candid in how Yahoo addresses this. I mean first, our stated goal is to capture the world's geography as it is used by the world's people. We don't see ourselves as the definitive authority on how a place should be called.'"

Comments

Yahoo! locations?

[palegray.net]

I just engage in triangulation by yodeling repeatedly with my buddies to establish locations.

Re:Yahoo! locations?

[JustOK]

what about the echo?

Re:Yahoo! locations?

[palegray.net]

That's handled by minions strategically stationed every fifteen degrees around my person. I pay them in bandwidth by modulating their data into my yodels.

Istanbul (Not Constantinople)

[Anonymous Coward]

So take me back to Constantinople
No, you can't go back to Constantinople
Been a long time gone, Constantinople
Why did Constantinople get the works?
That's nobody's business but the Turks

Re:Istanbul (Not Constantinople)

[palegray.net]

Evr'y gal in Constantinople
Is a Miss-stanbul, not Constantinople
So if you've date in Constantinople
She'll be waiting in Istanbul

Re:Istanbul (Not Constantinople)

[memorycardfull]

This isn't just funny, it's true! If they are having a PR problem, it can't be because they are calling "Constantinople" Byzantium. That little switch happened about 1700 years ago. He goes on to say it's still a very sensitive issue, well I bet it is. He just pissed off the Turks all over again. Keep up the good work dealing with those sensitive PR issues...the song make a catchy little mnemonic if you need help keeping it straight.

Re:Istanbul (Not Constantinople)

[plover]

Even old New York
Was once New Amsterdam
Why they changed it, I can't say
People just liked it better that way!

Re:Istanbul (Not Constantinople)

[leamanc]

Is a Miss-stanbul, not Constantinople

Sorry, the line is:
Lives in Istanbul, not Constantinople.

Re:Istanbul (Not Constantinople)

[palegray.net]

That's in the They Might Be Giants version, but not in the (original) The Four Lads version ;).

Re:Istanbul (Not Constantinople)

[1u3hr]

we accidentally called Constantinople Byzantium, just slipping back about 800 years there accidentally

Byzantium was the name from ancient (abot 660 BC) times until about 330 AD, when it was named for Constantine I, as the capital of the Eastern Roman Empire.

And Constantinople became Istanbul in 1453 when it was conquered by the Ottoman Empire.

I don't know how "800 years" fits into any of that. He seems to have forgotten about modern Istanbul completely, if he ever knew. I hope his geography is better than his history or maths.

Re:Istanbul (Not Constantinople)

[Hognoxious]

I hope his geography is better than his history or maths.

Not really very likely, is it?

Re:Istanbul (Not Constantinople)

[AdamHaun]

Turks called the city Istanbul before the conquest, and many people still called it Constantinople afterward. See Wikipedia for more. Still, 1453 is a reasonable boundary. I bet the guy meant to say "600 years" but got confused.

Simple test

[Anonymous Coward]

Always a fun test of any geolocation system:

Taiwan.

Re:Simple test

[Anonymous Coward]

Well said. In Asia, it's a multinational political issue.

Re:Simple test

[jrumney]

Taiwan is an island in China. I don't think you'll find any disagreement about that.

Re:Simple test

[!coward]

*raises hand* Sorry, Taiwan is an island _in_ China? Umm, so Hawaii is an island _in_ the USA?

The ambiguous way to go, without actually offending anyone, is to say that Taiwan is a chinese island (which China is another matter), or an island in the Sea of China, or even that it is an island off the south-eastern coast of China.

Re:Simple test

[sakdoctor]

Your post advocates a

( ) technical ( ) legislative ( ) market-based ( ) vigilante (X) political correctness

approach to not offending China. Your idea will not work. Here is why it won't work...

Re:Simple test

[Smallpond]

"without actually offending anyone"

except for the non-Chinese aboriginal population (ob. you insensitive clod).

Re:Simple test

[mb0]

Yesterday i had the same problem. I read http://www.gwytb.gov.cn:8088/detail.asp?table=WhitePaper&title=White%20Papers%20On%20Taiwan%20Issue&m_id=4 which states the official Chinese view of the situation. For a better introduction see http://www.wufi.org.tw/eng/timovmnt.htm .

Re:Simple test

[anothy]

guy i worked with was doing an internal directory system for a major multinational telecom equipment manufacturer. one of the functions was giving information about various corporate locations in long-form text.
we had a manufacturing facility in taiwan. he anticipated the issue, so did what he thought was the best option: he used the official ISO standard on the subject. should be safe, right? nice international standard, and the name even gets picked by the entity represented. so who could argue with that?

well, the official ISO name for taiwan is "Taiwan, Province of China". he got no end of abusive and arguably threatening email on the subject. pointing to the international standard was no use; pointing out that taiwan themselves got to pick what went in the standard was no use. and people from both taiwan and the mainland were upset.
eager to reach a nicer solution, he went to what probably seemed like a reasonable authority - the taiwanese consulate in the US - and asked what probably seemed like a reasonable question - what do you claim is the name of your country?
it took weeks to get a response. and that wasn't time spent trying to get someone on the phone; he managed to talk to people pretty quickly. but they all responded with something to the effect of "um, i'm not sure. i'll check and get back to you." he eventually got an answer that irritated a different but substantially smaller set of people and went with that.

Re:Simple test

[jrumney]

Perhaps you should look for something that states the official ROC government's position rather than a fringe pro-independence group within Taiwan. Or look at the official positions of the governments of the world ("what they say" rather than "what they do") - all recognize either PRC or ROC, never both, even if in practical terms every country treats Taiwan as if it were an independent nation.

Re:Simple test

[mb0]

Your are right, and the word 'better' is misleading. But i still think both of the links are interesting, especially because of the obvious bias. And after reading both nobody would disagree with you and maybe even learn a bit why this issue exists.

Another Mapping service, with Historical facts?

[spiffydudex]

Interesting to see a personal geotracking API. I am just wondering how will this be any different than having Google maps installed on your cell phone. There are at least a million users with Google maps installed on their GPS enable cell phones. If yahoo intends for GPS tracking and plotting, a garmin esque service or really what there intended market is.

I can see the usefulness as a meta information gathering tool for a certain area. So perhaps it is more like a free TomTom service but mixed with historical information about a certain area.

I'm just wondering exactly what they are trying to accomplish here other than having another mapping service. While Google maps is almost a required app on any unlimited data plan cell phone.

Re:Another Mapping service, with Historical facts?

[Anonymous Coward]

It may be only that they're only trying to make another mapping service. While that seems pointless given Google's lead (in time and talent), Yahoo may see it as crucial to their own relevance to advertisers to expand into geographically sophisticated results to queries. They may believe that the appoaching mobile-device dominance means not being able to relate things spatially will seem like an internet with a 'sense' cut off. Also not all of their effort will duplicate Google's. A few innovations, patents, and good sized databases will make them a more attractive purchase to Google, even though Yahoo has no ultimate hope of matching them in the market.

Re:Another Mapping service, with Historical facts?

[spiffydudex]

True, a lucrative market share for yahoo to get into. Good Point.

Re:Another Mapping service, with Historical facts?

[plover]

It's not just "yet another mapping service." He's saying that the service will provide spatially relevant information, that the API will add value to the information (from both Yahoo and user-contributed sources), and how overcoming the difficulties isn't as simple as scratching out a few requirements.

One example might be if I searched for "ATM" and I was on the freeway when I made the request, it would search for ATMs around the nearest exit ramp instead of the nearby store on the other side of the fence. Or maybe it would incorporate police reports about snatch and grabs in the area, so I'd choose a different ATM.

Or perhaps if I'm in Minneapolis and I search for Miami hotels, it'll look in south Florida, but if I'm in Oxford, Ohio, it might find one near the university.

As for their added value, perhaps they'd put ads for restaurants near the ATM.

Re:Another Mapping service, with Historical facts?

[Hognoxious]

I am just wondering how will this be any different than having Google maps installed on your cell phone.

Maybe it won't suck?

I think allot of people are in the same boat here

[linguizic]

So, this is interesting and all, but I have NOTHING to say about it (except this). If the article were about something new with the API's that Yahoo offers that would be something to talk about.

ummm already...

[johnjones]

ok I actualy really like yahoo maps...

google maps look pretty but they are not that fast compared to yahoo flash (.flv) maps for certain things

and frankly their fire eagle stuff blows the socks off the rest

BUT this feels much like the http://www.geonames.org/ and frabkly I prefer to use geonames I would have thought it would be much better that yahoo actually did something like this allowed people to download the data and provided a web service...

I prefer geonames because its open and well frankly works

it might be solving a differant problem but really this article does not point that out to me

regards

John Jones
http://www.johnjones.me.uk

Re:ummm already...

[FooAtWFU]

Yahoo! has also done a fairly good job of supporting initiatives like Open Street Map (the Wikipedia of online maps). Google... just buys map data from the usual suspects. Support free information. :)

Re:ummm already...

[xwizbt]

I kind of wished umm already that like you used grammar or what english is to be understood? It's just that other lines, and yes you're in the minds of the moderators, already and so on. But run on and something sentences are like in the understanding my milkshake brings all the boys to the yard. And you are my boy, oh yes... taste my milkshake, for I have already won.

Mmm, Taco Bell!

[slash.duncan]

Taco Bell. That sounds good right about now!

Wha? /Tyler/ Bell? Who wants to eat floor tile, or for that matter, the floor tiler? Taco Bell has floor tiles...

Honestly, it looked like Taco Bell to me when I first glanced at the feed and I wondered what that had to do with Yahoo. Yes, I /am/ hungry. How'd you know?

Why?

[slashtivus]

Finding what cities are in a county? Finding bordering counties?

This has already been done. I have 6 year old software that could do most of this. And in that software it was already old-hat. It's called GIS.

TIGER data (free from the Federal government / census) has it, as well as many other (non-free) sources.

Re-creating all of this from scratch seems a lot like re-inventing the wheel.

Re:Why?

[linguizic]

Yes, but can you simply type "what county borders Contra Costa County, CA to the south?"?* I think this kind of functionality is what they are striving for.

*Yes I know my punctuation's off, this is what makes the most sense to me damn it!

Re:Why?

[Anonymous Coward]

Not really inventing the wheel -- they build it on the where-on-earth acquisition the did a few years back where they got this database that were developed and enhance for over 20 years to include all sorts of data -- far better than just tiger data -- they probably got the best data in the world.

Re:Why?

[slashtivus]

A simple parser can interpret "CA", "Contra Costa" the keyword(s) [County] and [border]. A parser is needed, but the very sleep inducing article otherwise brings nothing to the table.

The summary indicates that it is a new 'tagging' system, but in reality it is mostly about a better language parser (There are already some very good ones). The summary indicates that they are trying to to a "re-naming" of all this. A parser results in: "California", "Counties", [Contra Costa] => [borders].

This already exists many times over. There are many systems where everything is already numbered, Geo-coded, cross-correlated, and tagged with names.

Re:Why?

[Quixote]

Wow! If only they'd talked to you first.

BTW... can you point me to which TIGER/ZIP file contains the information for Istanbul? Or how about Casablanca, Morocco? And which county (district) contains the town of Aleppy, Kerala, India?

Thanks! Much appreciate it.

You know, America isn't the center of the known world.

Re:Why?

[Fred_A]

You know, America isn't the center of the known world.

I typed "here be dragons" in their thing and got nothing back, so it doesn't work anyway.

Re:Why?

[clickety6]

I also typed in "here be dragons" and got back Indonesia ;-)

http://en.wikipedia.org/wiki/Komodo_dragon

Re:Why?

[Anonymous Coward]

You know, America isn't the center of the known world.

Sometimes it seems to be the centre of population density, though.

Re:Why?

[witherstaff]

The US National Geospace data files have a variety of information for basically anywhere in the world. It includes some of the info you're after.

Re:Why?

[Anonymous Coward]

ROTFLMFAO.

Hello USA, Rest of the world here...

Re:Why?

[slashtivus]

Meh. Yahoo! and Slashdot are both American companies. Also: http://slashdot.org/faq/editorial.shtml#ed850/. Saving time by using available data is still a valid post, even if it does not apply world-wide.

Re:Why?

[FindItByMe]

Although TIGER/ZIP does not contain information outside the US, the National Geospatial Intelligence Agency does maintain a database of place names which are available for search/download at http://earth-info.nga.mil/gns/html/namefiles.htm. The USGS also publishes a database available for download for place names for the US. It is located at http://geonames.usgs.gov/domestic/download_data.htm. To create a format that would be meaningful to any viewer from any country, I believe you are faced with the "esperanto" concept. Esperanto was supposed to be a universal language that everyone would learn in addition to their native tongue so we could all communicate. However, objections arose when it was discovered that the language was heavily related to the romance language group and was written with the Roman alphabet. So it is mainly a curiosity now rather than useful. In a geospatial context, if you are Chinese, it would be hard to try to figure out the Pinyin equivalent of the whatever place name on the map so you could translate it back into the your native character set. The same for other scripts as well. So maybe a monster wiki is needed for each cultural/linguistic context, but it would never be authoritative in the classic sense of the word. And then there is the question of regional boundaries which can shift from time to time and how would you specify the time/historical parameters? I think a foundation should be built which can accomodate any context in terms of region/culture/language/historiography so that it can be useful to anyone who views it.

Ummm

[Dodder]

Isn't this also known as latitude and longitude? Or is that too 20th Century?

Re:Ummm

[linguizic]

Or is that too 20th Century?

I don't know what the 20th century has to do with it, I couldn't tell you my current latitude and longitude back in 1999. However, I could tell you that I'm currently in San Ramon, California, which is just east of Castro Valley and south of Danville. This would probably be a hell of a lot more meaningful to you. Even if you didn't know any of the places I just told you about, you could easily look them up.

Personally, I think it would be really useful for me to enter the search terms "town north-east of San Jose, CA" and be taken to a map that shows Milpitas, CA on it.

Re:Ummm

[Anonymous Coward]

Long-lat will only get you position not the name or the contained geographical names -- what they do is actually pretty cool -- they will tell you that Rome is in province of Lazio which is in Italy which is in Europe, and they will tell you that down to zip level and tell you what the names is of all the surrounding areas on each level.

Re:Ummm

[Anonymous Coward]

How is that a remotely difficult problem? It sounds like a first year CS project.

Re:Ummm

[omuls are tasty]

Dunno where you live, but your CS students must have a load of time on their hands if they're able to enter the said data for every settlement in the entire world.

Re:Ummm

[Hognoxious]

Sounds more like junior school geography.

Re:Ummm

[Anonymous Coward]

Personally, I think it would be really useful for me to enter the search terms "town north-east of San Jose, CA" and be taken to a map that shows Milpitas, CA on it.

Here's the rub.

How far northeast? How do we distinguish Milpitas from Rancho Cordova from Montreal from your query? If you go far enough, Saratoga is northeast of San Jose. Some 25,000 miles northeast.

Google has a clever idea along these lines, but it won't get you Milpitas. Maybe Livermore, and probably Sacramento.

Re:Ummm

[dino213b]

This wouldn't be a question of data entry. In most cases, the data is already there. All you need is a way to connect it. For example, the U.S. Census Bureau has a list of neighborhoods and their polygonal borders.

Now, the CS students in your reference would simply apply a PIP (point-in-polygon) algorithm to see which neighborhood a particular lat, lon coordinate belongs to.

And since that's computationally intensive, the CS "students" would have to know a bit about optimization - including quad trees, and so forth. Bla.

Even old New York was once New Amsterdam.

[Kuroji]

Why they changed it, I can't say. People just liked it better that way.

Re:Even old New York was once New Amsterdam.

[MrMista_B]

The Might Be Giants, for the win.

Re:Even old New York was once New Amsterdam.

[julesh]

so, why'd the first two get funny mods, but this one got interesting?

Re:Even old New York was once New Amsterdam.

[Hognoxious]

The song is much older. Probably dates from before they were born.

Re:Even old New York was once New Amsterdam.

[SlipperHat]

AFAIK, New Amsterdam was a Dutch colony, named after the Dutch capital. When the British took it over (through conquest I believe) they changed the name to something more in line with their own geography. The fact that York was not the British capital was meant to be a snub e.g. a town like York (at the time) was better than Amsterdam.

Istanbul/Constantinople/Byzantium

[cborg]

Is he saying he referred to Constantinople as Byzantium, or is he referring to Istanbul as Constantinople within the Byzantine empire? Either way I'm sure will irk the Turks.

Re:Istanbul/Constantinople/Byzantium

[Fred_A]

Either way I'm sure will irk the Turks.

Advanced Diplomacy end of year exam : "Find a topic that will not irk the Turks".
You have four hours.

Re:Istanbul/Constantinople/Byzantium

[Anonymous Coward]

Advanced Diplomacy end of year exam : "Find a topic that will not irk the Turks".
You have four hours.

That's simple, bathing.

Re:Istanbul/Constantinople/Byzantium

[Fred_A]

Advanced Diplomacy end of year exam : "Find a topic that will not irk the Turks".
You have four hours.

That's simple, bathing.

Heh, good one :)

The first rule

[actionbastard]

About Yahoo's Open Location API, is not to talk about the embarrassing mistakes that you made because you failed both Geography AND History.

Re:The first rule

[Anonymous Coward]

Tyler Bell says, the first rule of Yahoo's Open Location API is DO NOT TALK about Yahoo's Open Location API.

Oh wait. I'm thinking of Tyler Durden. Carry on.

Yo quiero Tinker Bell

[tepples]

Honestly, it looked like Taco Bell to me when I first glanced at the feed

I saw Tinker, not Taco. But then I babysit a single-digit-year-old girl.

Re:Yo quiero Tinker Bell

[Anonymous Coward]

babysit

Is that what they call it now?

A false start...why do I think so?

[bogaboga]

Is it me alone? But why do I always think that Yahoo's ventures including this particular one, are always false start...or better put..."dead on arrival?"

Re:A false start...why do I think so?

[linguizic]

Because you've never played with BOSS or YQL.

More geo?

[gmuslera]

How this mix with GeoNames or MaxMind?

All are diffenent, but some way to standarize data?

There's a joke in here somewhere...

[Anonymous Coward]

> WOEID With a name like that, the product is doomed to eternal... oh wait.

You'll inspire students

[troll8901]

Darn, you are making me excited about the future possibilities of smart map applications! (Keyword: Smart)

You should be a lecturer. You'll inspire students greatly!

WOEID?

[Dodder]

So you know your WOEID? I could have sworn the article was talking about using a unique identifier to identify places that had nothing to do with their names because the name could potentially change.

And pretty sure long/lat is a little more precise than zip code, pretty universally easy to look up for any location and more amenable to wildcard searches and approximation.

"Yahoo! has been working for a while to promote a unified system for referring to places, through their Where On Earth IDs. Using a WOEID."

Where on Earth indeed.

[novakreo]

We don't see ourselves as the definitive authority on how a place should be called.

I should hope so. As far as Australia goes, geotagging photos on Flickr is a frustrating process. When it doesn't get the suburb right, and you click "See other nearby options", you'll be presented with a random list of other suburbs and municipalities, none of which are anywhere near the actual location. It's atrocious.

Cyprus

[herbertchapman]

The road maps in Cyprus don't acknowledge the fact that half the island is effectively another country. There's just a vague wording saying "Area inaccessible due to Turkish Occupation". How do you plan a route round that without upsetting anyone ?

The GDR is still alive: on Googlemaps

[Anonymous Coward]

Deutsche Demokratische Republik

OK, it doesn't beat Byzantium.