Go to file
Kevin Dungs 1311a32577 Merge pull request #14 from betatim/google-ngram-wordlist
New word list based on google ngram from 2008
2014-11-21 10:06:08 +01:00
static Smarter landing page 2014-11-18 15:40:05 +01:00
views Submit button works again 2014-11-18 16:47:47 +01:00
words New word list based on google ngram from 2008 2014-11-21 09:23:07 +01:00
.gitignore Add lines for Vim swapfiles and OS X Ds store. 2014-11-18 14:35:04 +01:00
LICENSE Updated install instructions and added new collaborator 2014-11-17 14:55:12 +01:00
Procfile Add Gunicorn to requirements. 2014-11-17 16:00:32 +01:00
README.md Yet another wordlist, based on WordNet this time 2014-11-17 18:25:37 +01:00
app.py More tidying up from Tims rebase screw up 2014-11-18 15:46:56 +01:00
requirements.txt Add Gunicorn to requirements. 2014-11-17 16:00:32 +01:00
thesethreewords.py Undo temporary filename change 2014-11-21 09:32:23 +01:00

README.md

These3Words

Address any 3meter x 3meter square on earth with a unique three word name.

Inspired by http://what3words.com/

example

>>> import thesethreewords as these

# the home of particle physics
>>> CERN = (46.232355, 6.055419)

>>> three = these.three_words(CERN)
>>> print three
'spitting-ripple-fontanel' 
>>> these.decode(three)
(46.232335567474365, 6.055419445037842)

Check out where this is on google maps.

requirements

You need to install the geohash and bottle libraries:

$ pip install geohash
$ pip install bottle

six words

There are a lot of 3x3m squares on the earth's surface. To encode them in only three words requires a long wordlist, as a result some fairly obscure words get on it. If you can live with having to remember six words the wordlist is much shorter. The six word wordlist comes from the amazing humanhash library. Words were chosen to maximise clarity in human communication, they should be more familiar than the words on the three wordlist:

>>> six = these.six_words(CERN)
>>> print six
'spaghetti-carolina-kentucky-oscar-iowa-table'
>>> these.decode(six)
(46.232335567474365, 6.055419445037842)

how it works

Each latitude/longitude pair is converted to a nine character geohash. This provides about 3meter resolution at all latitudes. The geohash is then converted to an integer which is encoded as a string of words.

The wordlist used to encode the geohash into just three words uses your local computers dictionary. Some attempts are made to remove really obscure words but it could be better. You need to use the same wordlist when encoding and decoding a these-3-words hash.

The these-3-words hash shares the property of a geohash that nearby locations share have similar these-3-words hashes

>>> other_CERN_site = (46.256811, 6.056792)
>>> six = these.six_words(other_CERN_site)
>>> print six
'spaghetti-carolina-kentucky-utah-seventeen-neptune'
>>> these.decode(six)
(46.256797313690186, 6.056792736053467)

The other CERN site is here.

webservice

The file server.py provides a tiny webservice that allows to display a location given by three words on a Google Maps map.

The server requires bottle.py to be installed. It can be run locally by typing ./server.py or python server.py respectively.

brought to you by @betatim and @kdungs productions