Our data is released under various Creative Commons licenses.더 많은 정보
해당 콘텐츠를 좋아한다면 기부도 고려 부탁해요.
메뉴
Bug (?)
Searching for "påse" in "any language" yields, besides Swedish, some unexpected results in Turkish.
https://tatoeba.org/eng/sentenc...rom=und&to=und
http://i.imgur.com/mWYmHkM.png
http://i.imgur.com/GtXMpy1.png
That's interesting behavior, but it shouldn't get in anyone's way. People can get reasonable results by restricting the search as follows:
- From: Swedish; Search: påse
- From: Any language; Search: *påse*
- From: Any language; Search: =påse
People searching in Turkish would presumably not use the å character.
Apparently, the Turkish stemmer (software that tries to condense multiple forms of a particular word) converts the "på" part of the string into something that matches "bunu", "bindim", and so on, and the "se" part of the string into something that matches "senden", "senin", and so on. Note that the stemmer was not developed by us, or even by the developers of Sphinx, the search engine that we use. It was contributed by someone named Evren Çilden to a project called Snowball that Sphinx uses.
In case anyone is curious, here's a link to the stemmer:
https://raw.githubusercontent.c...em_Unicode.sbl
Thank you, that’s interesting.