GeoNames Forum

Sort of my fault that I provided material for the wrong direction by previously correcting those 19 names manually

Anyway, thanks! You're being very helpful, and this whole Geonames service is an amazing thing.

Dan

Hi marc,

this topic has already been discussed under "07/10/2006 07:47:28 Subject: Re:Železničné Stanice". You wrote: "19 records have just been updated with a replacement of ŭ by ů. All using your userid."

Well, I did not mention I found this number surprisingly small. Now I found evidence that some names in Czechia (at least two) escaped your batch. One of them is "Čápŭv Dvŭr".

I am not sure what happened. The wrong character is Unicode decimal chr(365), LATIN SMALL LETTER U WITH BREVE. The desired character is chr(367), LATIN SMALL LETTER U WITH RING ABOVE. Toponyms affected can be in any category.

Best,
Dan

Thank you, marc! Don't worry about "Řeka"/"Říčka". Frankly, I had not seen these two in the data, I only tried to come up with all possible terms I could think of. So your Ř is probably not the cause.

Dan

Great, thanks!

Your extra example makes me suggest another batch correction. "Golčŭv Jeníkov" should actually read "Golčův Jeníkov" and this is also something I have seen several times elsewhere. Seems like some of your sources consistently use "ŭ" instead of "ů". You can safely replace ŭ by ů anywhere in CZ provided it's in the main name and not in a foreign-language alternate name. The former is not in the Czech alphabet.

Some other thing I've been correcting repeatedly are wrongly capitalized words inside of a toponym. Czech does not routinely capitalize every word in a name like English. This is a little bit more tricky, since different conventions hold for populated places and different for natural features like streams or mountains. So, a mountain shall be "Krušná hora" but same-named settlement shall be "Krušná Hora". Even a second word in a name of a mountain will be sometimes capitalized if it is derived from another toponym and not a common noun. However, I feel pretty confident that at least the words "hora" and "vrch" in mountain/hill names, and the words "potok", "říčka", "řeka", "jezero", "nádrž" in stream/lake names shall be lowercase - as long as it's Czech and they are not the first word in the name.

Best,
Dan

I noticed couple of toponyms starting "Železničné Stanice". This means "railroad station" but the spelling is incorrect. Correct is "Železniční stanice" (note the accute i and the lowercase s). I corrected some occurrences I came across but as this seems to be a global problem, is there a way how to do it in a batch? All affected toponyms would be in Czechia and all would be in the railroad-station category.

Hi Bernard,

yes, my guess about how it's done was about the same, although it does not answer the question why sometimes only the country is computed. Maybe the subdivision info comes with batch imports from external sources.

Indeed, if I was able to make visible and edit the border polylines, I would occasionally correct them. But I understand this would require a considerable effort in changing the user interface: not only move points but also add them, and associate with the polyline...

This raises another interesting question (although not related to the original topic): There are many features of non-point nature. Many features are better described as lines or areas, and simply placing their names somewhere as a point may not be of much use. I noticed that most stream names are placed at the end of the streams, i.e. their mouth into a bigger stream. That results in a number of H flags along a river, which however bear names of inflows; finding the name of the river itself is very difficult. Would it be inappropriate to have more than one point with the same name for the same river?

Hi,

I realized that neither creating new names, nor editing the old ones allows me to change the encompassing administrative division and/or the country for a name.

Example: I realized that "Turnov" is put in "Czech Republic, (EZ61)". I don't know where the cryptic code EZ61 comes from, but I do know that Turnov lies in "Czech Republic, Liberecký kraj, Okres Semily". Some (or all?) manually created names don't list divisions of countries at all. Any way to change such things?

Sometimes this even affects countries. Not that countries are represented by cryptic codes but the border lines seem not to be always accurate. I remember to have created a name along the Czech-Polish border (but definitely in Czechia), and the system automatically assigned the "Poland" label to it. Can I influence that?

Great, thank you!

Hi, I recently made (logged as zeman) couple of additions, movements and spelling corrections in and around Praha, Czechia. Although I see the changes in the list of recent changes and in the history of the places, the map does not reflect them. Even after emptying the cache and reloading. Any ideas why this might happen?

Thanks,
Dan