For many countries where there are different scripts being used and where no single Romanization system/standard is defined, a master list can help guide users towards making use of the same spellings to identify or describe the same locations. For example, in Laos, the UNESCO World Heritage city of ຫຼວງພະບາງ can be spelled as Luang Prabang, Louangphabang, Luangprabang, and so on, because of a lack of a single Romanization standard for the Lao script.
Because of this, when users search for e.g. health facility or village names in the Explorer, they may spell Geo-Object names slightly differently. However, the Explorer currently seems to work only with exact matches. I’d like to suggest adjusting the search box to return approximate matches in the search results panel.
Currently you can search partial words as long as the characters entered match how the desired text starts. For example, searching “Centr” will return “Central Health Facility” in the results. This may be helpful in some situations.
I think there’s a more formal approach that I’d like your opinion on. In the the use case you describe those variations could be registered as synonyms. That way they are formally set as variations. There are two challenges with making this a complete solution.
There needs to be a way to formally manage synonyms (like deleting). Currently (as of version 1.0.0) you can only add synonyms through an upload.
The Geo-Object search needs to be expanded to search on synonyms. It does not currently search synonyms.
Do you think this approach of formally managing synonyms is practical from a data curation standpoint?
Hi @jlewis, thanks for the in-depth research and detailed explanation of your proposed solution! I see your point that variations could best be managed as synonyms to facilitate searches, however, for some languages, there are infinite (informal) ways people may be used to spelling names, such as in the Lao language. Managing synonyms in this case may become a daunting task, and in my opinion, would require a disproportionate amount of effort.
Is there any possibility to apply some sort of fuzzy matching in the search box?