Transliteration – a challenge for data providers, geocoders and mapping portals


Transliterate – verb (transliterated, transliterating) “to replace (the characters of a word, name, text, etc) with the nearest equivalent characters of another alphabet.” – Source Chambers Online

I have just got back from a holiday in Tel Aviv. To my dismay I discovered that none of the big mapping portals (GYM etc) have any decent coverage for Israel. Initially, I thought that might be because of the lack of availability of street level data but eventually I found a demo site for Atlas CT which has some reasonable mapping from MAPA – Mapping and Publishing – at the time of writing the site seems to be down but it worked very well while we were in Israel.

The only problem is that the site uses an English language interface which whilst great for me (who doesn’t write Hebrew very well) creates some challenging problems in transliterating the Hebrew road names. There is no precise way of switching from Hebrew to English and similar challenges exist for Arabic, Chinese, Japanese and other languages that do not use our 26 character alphabet. Unfortunately soundex just wont solve these problems and the parser in the site was having all sorts of difficulty with my attempts to input street names. In the major cities of Israel Street names are shown in Hebrew, Arabic and English so it should have been fairly easy to adopt standard spellings but apparently not.

Having worked on the whereonearth global geocoder in the past, I know how difficult it is to find a solution to this challenge. I wonder whether that is why GYM have not included coverage of Israel?

Over half the worlds population does not use the western alphabet, global mapping portals and LBS need a consistent way of dealing with transliteration.