GeoNames Forum

Hi Andy

These countries are available, only the geonameId is missing in the countryInfo. You could add the geonameId to your list above and I will add it then to the countryInfo.txt file.

Marc

PS: It is related to another thread : http://forum.geonames.org/gforum/posts/list/570.page

At the moment we don't have the resources to open the natural language geocoding code. Would you have time to help us with this? (weeks not hours)

Cheers

Marc

Thanks, I have updated the countryInfo doc.

You could help us sort out the missing or wrong geonameIds in the countryInfo file.

Cheers

Marc

Do you use any indices on your columns? You should also look at the geospatial features of mysql :
http://dev.mysql.com/doc/refman/5.0/en/spatial-extensions.html

We at GeoNames are using Postgres/postgis.

Cheers

Marc

I prefer not to have to add this feature to the result. It would make the service considerably more complex with only little gain.

Marc

Du kannst uns dabei helfen oder du kannst die codes auch bei ISO kaufen :
http://www.iso.org/iso/en/CatalogueDetailPage.CatalogueDetail?CSNUMBER=8349&ICS1=1&ICS2=140&ICS3=30

Gruss

Marc

Hallo Larry

Denkst Du dabei an etwas wie den Parameter 'name_equals' bei der allgemeinen Volltextsuche :

http://www.geonames.org/export/geonames-search.html

Gruss

Marc

Sure, parsing xml on the server will work. A lot of people to this.

Marc

We only use the search engine results for ordering. In Austria there are hundreds of places with the same postal code and if users search with a particular postal code they expect the largest city on top of the result.

If I look up the US zip code 98294 on the United States Postal Service website I get only one single result 'Sultan'. Exactly the same on GeoNames.

http://zip4.usps.com/zip4/citytown_zip.jsp

Do you think the USGS is wrong on this?

Cheers

Marc

Hallo Larry

Die Suchfunktionen sind eigentlich alle nach 'Relevanz' geordnet. Bei den Postleitzahlen verwenden wir dazu die Anzahl hits bei Yahoo für den Ortsnamen und die PLZ.

Grundsätzlich wäre eine rein alphabetische Sortierung möglich. Ich denke es ist eher ein 'nice-to-have' Feature und die GeoNames todo Liste ist schon beträchtlich lang.

Gibt es noch andere Benutzer die and einer Sortierung Interesse hätten?

Gruss

Marc

Lets say a user comes to your site www.yoursite.com than the browser of this user will refuse to load xml data from the server geonames.org. There is no such security restriction for JSON data.

Marc

I should have added that this service is not yet documented since it is impossible to performance optimize it for all combinations of our 650 feature codes.
If the above combination of codes is requested pretty often we will add a database index to speed it up.

Regards

Marc

Hi Eugene

Sorry, We were having performance problems. It shouldn't happen too often. You see it in the threads of this forum how often it does happen.

If you are running a missing critical application we have some commercial servers with more RAM and less load where we can make uptime guarantees. Contact me for details.

Regards

Marc

You cannot use our XML service from your javascript code, you will have to use XML via you server. This restriction comes from a browser 'security' feature. The browser does not allow XML calls to third party servers.
You can overcome this problem and use the JSON version of the same service.

You can look at our example for the search service to get an idea of JSON programming :
http://www.geonames.org/maps/json-googlemaps-example.html

Hope this helps to get you started

Marc

Hi Chris

We are still working on fully integrating the postal code data with the main GeoNames data. For the time being we cannot release a link to the geonameid for the postal codes and it is too early to release what we already have. For some countries (France, Spain, and some more ) we have started to add postal codes as alternate names to the main GeoNames entries.

Cheers

Marc

PPLL is pretty small 'an area similar to a locality but with a small group of dwellings or other buildings'.

Marc

Hi Eugene

The 'findNearbyPlaceName' returns all populated places. If you don't care about PPLX you have two possibilities. You can either get more than one row and filter it out yourself or you can use the (not yet documented) 'findNearby' service with the featurecodes you are interested in (PPLC,PPLA and PPL).

http://ws.geonames.org/findNearby?lat=48.865618158309374&lng=2.344207763671875&featureClass=P&featureCode=PPLA&featureCode=PPL&featureCode=PPLC

Regards

Marc

Hi Chris

The totalResultsCount is included in the search result because it is easy to implement and it has some meaning.
What would be the meaning of a total results count for a find nearby service? The total results count would always be the total number of records in the database. Wouldn't it?

Cheers

Marc

Hi Chris

You are right. I have added the parameter 'startRow'.

http://ws.geonames.org/postalCodeSearch?country=GB&style=MEDIUM&placename=london&startRow=200

Marc

Hi
The order of the PPL* features goes like : PPLC > PPLA[2-4] > PPL > PPLX > PPLL

More than 96% of the P features are PPL. The exact number are here : http://www.geonames.org/statistics/total.html

Features with PPLC, PPLG, PPLA and PPLR don't have a duplicate entry. We may use the tagging functionality in the future to add multiple tags to single features with more than one function. For the time being the 'higher' code takes precedence. (for instance PPLC instead of PPLA)

There are hardly any PPLS, so you can forget about them. It is a plural feature if a couple of villages are merged into one GeoNames entry.

Regards

Marc