Problem with diacritics in Search/Create bar
Completed
When I type “ma” in the Search/Create bar, the “a” is selected and thus deleted when I continue typing, so typing “m”, “a”, “l”, “o” leads to “mlo”, and typing “m”, “a”, “a”, “a” only leads to “ma”.
I think the reason is that I have a note beginning with “mål”. My guess is that nvUltra
- identifies “ma” with “må”
- thinks I’ve typed only one character (maybe the counting goes wrong because it decomposes the diacriticised letter)
- selects all characters except the first one (“m”) instead of correctly selecting all characters except the first two ones (“ma”)
I should add that in my opinion, (1), while probably intentional, is not a good solution, at least for Danish “å”, which is a distinct letter from “a”. No big deal, though.
Thomas
-
I'll take a look -- I had fixed this issue previously, but had to rewrite some of the search bar code, and may have left this out.
As for your opinion about (1), I recognize that diacritics in some/many instances technically create a new letter, but for search purposes it doesn't make sense in many instances. For instance, searching for "crepe", I would expect to find documents that include "crepe" or "crêpe". Especially since some documents may have been imported from sources that are less "fastidious" about diacritics, or documents that were in plain ASCII.
When indexing files, the unicode characters are decomposed and stripped of diacritics, as are the search terms. This slightly reduces the size of the index files, but makes searching more inclusive and forgiving of error in the source documents.
0 -
The issue with autocompletion has been fixed for the next release.
Thanks!
0 -
OK, thanks a lot!
Yes, I agree it’s probably a better solution overall to ignore diacritics.
Thomas
0
Please sign in to leave a comment.
Comments
3 comments