Choosing the right English Name for a POI

Why not consider this simpler and clearer rule:

  • If a Wikipedia entry exists, use the corresponding entry name.
  • Otherwise, transliterate the Thai name using the RTGS system.
  • Alternative names found on signs can be stored in alt_name:en.

This approach can encourage mappers to use Wikipedia, which is often integrated into third-party applications.

However, the challenge remains in informing/educating mappers to use the official RTGS system, as there is no straightforward way to produce an accurate transliterated term for OSM.

The tltk library/service from Chulalongkorn University does not translate generic terms, nor offer capitalization. Ideally, we would need a one-to-one transliteration process that can be directly copied and pasted into OSM and other sites, such as:

เกาะเสม็ด > Ko Samet
สวนสาธารณะหนองบวกหาด > Nong Buak Hat Public Park
จังหวัดภูเก็ต > Phuket Province

In the absence of a such transliteration service, users will likely continue relying on visible sign names or translation services like Google, leading to conflicting variations.

1 Like

What concerns me is that Wikipedia is not always correct, but since we can edit the wiki ourselves and the Wikipedia community appears to be larger than our OSM ones, using Wikipedia may be fine and may help correcting the wiki. In addition, Wikidata is another good source.

Currently, our wiki stated that:

  1. Collect it from signs when collecting data. Map users will compare to signs.
  2. Wikipedia, other sources on the internet or for names of administrative organisations
  3. Use the program for Windows [Chulalongkorn University: Thai Romanization Program]
  4. Transcribe yourself according to [wikipedia: Royal Thai General System of Transcription]

I’d like to suggest:

  1. If it is a private organization/business with an English name that is intentionally used in their sign, document, or media, using their written name.
  2. Otherwise, use Wikipedia/Wikidata (if available).
  3. If Wikipedia/Wikidata do not exist, use RTGS.
  4. Any found variations can be stored in alt_name:en.

This is a challenging problem to solve. Some Thai words can be pronounced differently, despite having the same spelling, so we have to know the pronunciation to ensure correct transliteration. We can sometimes get around this by typing the direct pronunciation into the tltk. It should be 100% correct, except for spaces, which require the vocabulary database to add them correctly. If we talk about the perfect software with a vast vocabulary database, I believe it is possible, but the developer has to put in a lot of effort.

This, by the way, is not limited to Thai to Latin transliteration. Transliterating an English word into Thai script brings similar challenges because the same alphabet can be pronounced differently. (This kind of transliteration standard is also documented by RTGS.)

2 Likes

Just another remark regarding business signs: Do type the name, do not copy typography into the name tags. Means: Tesco, not TESCO. This one likely to not be much of a problem any more regarding rebranding to Lotus’s. But you get the idea.

Here again Wikipedia often provides hints to the right spelling. And for brand names it is usually “brand” and not “name”. You could optionally place it in both for compatibility. This might bring other challenges.

I still like the idea of having a mini dictionary with common translations. I would like to point to the discussion we had a while ago regarding “$1 Subdistrict Health Promoting Hospital” for “โรงพยาบาลส่งเสริมสุขภาพตำบล (.+)” (Rongphayaban Songsoem Sukkhaphap Tambon)

2 Likes

I’m also on board with the idea. How can we go about compiling a list?

I think we can create a list in our wiki for places that we often found.

However, I’m not sure how many of them would be included in the wiki. If there are too many, the wiki may get messy.

If it possible, I think these places should have an English Wikipedia page.

1 Like

FYI: I have created a Google spreadsheet containing some initial terms. You are welcome to extend the list by adding terms that you commonly use. You can access and edit the spreadsheet via this link:

Additionally, I have reached out to K. Wirote at Chulalongkorn University to solicit feedback and input on this matter.

1 Like

I disagree with “district”. It should be sub-district, also based on earlier discussion in the forum.

I added some more.

1 Like

Thanks. It seems that I may have made an error while copying the district terms.

Here is the feedback provided to me by K. Wirote at Chula University:

1 Like