This commit is contained in:
Jan-Piet Mens 2015-10-21 10:46:55 +02:00
parent e0f04c6fd2
commit 42feb5ee65

View File

@ -329,7 +329,7 @@ We recommend you keep reverse-geo lookups enabled, this data (country code `cc`,
### Precision
The precision with which reverse-geo lookups are performed is controlled with the `--precison` option to _recorder_ (and with the `--precision` option to _ocat_ when you query for data). The default precision is compiled into the code (from `config.mk`). The higher the number, the more frequently lookups are performed; conversely, the lower the number, the fewer lookups are performed. For example, a precision of 1 means that points within an area of approximately 5000 km^2 would resolve to a single address, whereas a precision of 7 means that points within an area of approximately 150 m^2 resolve to one address. The _recorder_ obtains a location publish, extracts the latitude and longitude, and then calculates the [geohash](https://en.wikipedia.org/wiki/Geohash). If the calculated geohash width (i.e. precision or string length of the geohash) can be found in our local LMDB cache, we consider the point cached; otherwise an actual reverse geo lookup (via HTTP) is performed and the result is cached in LMDB at the key of the geohash.
The precision with which reverse-geo lookups are performed is controlled with the `--precison` option to _recorder_ (and with the `--precision` option to _ocat_ when you query for data). The default precision is compiled into the code (from `config.mk`). The higher the number, the more frequently lookups are performed; conversely, the lower the number, the fewer lookups are performed. For example, a precision of 1 means that points within an area of approximately 5000 km^2 would resolve to a single address, whereas a precision of 7 means that points within an area of approximately 150 m^2 resolve to one address. The _recorder_ obtains a location publish, extracts the latitude and longitude, and then calculates the [geohash](https://en.wikipedia.org/wiki/Geohash) string and truncates it to _precision_. If the calculated geohash string can be found in our local LMDB cache, we consider the point cached; otherwise an actual reverse geo lookup (via HTTP) is performed and the result is cached in LMDB at the key of the geohash.
As an example, let's assume Jane's device is at position (lat, lon) `48.879840, 2.323522`, which resolves to a geohash string of length 7 `u09whf7`. We can [visualize this](http://www.movable-type.co.uk/scripts/geohash.html) and show what this looks like. (See also: [visualizing geohash](http://www.bigdatamodeling.org/2013/01/intuitive-geohash.html).)