P: Case insensitive search with diacritic in uppercase letters

[DenisPac]

Hello Community,

Following my recent post in the related thread here, I'm creating this new one dedicated to the issue "Case insensitive search with diacritic in uppercase letters".

Actually I've noticed that the case insensitive search in LRC (latest 11.0.1) still doesn't work when there is an accent on the uppercase letter.

I mean "église" doesn't return "Église" (and possibly vice-versa) where it should be expected since the Lightroom search is not case sensitive.

Thanks,

Denis

[allen-c]

At the moment a letter carrying an accent (in a keyword, a photo title, etc) is treated as being distinct from a letter without one - "théâtre" and "café" are not equivalent, are not implicit synonyms for "theatre" and "cafe". Moreover an uppercase accented letter is not considered equivalent to a lowercase one: Ô and ô are not the same. This seems to me to be totally counter-intuitive, especially when we are used to services like Google which do treat these forms as being equivalent. If I have a photo of "Le Théâtre de la Comédie Française" and another of "The Roundhouse Theatre" I'd expect to able to find both using a filter on either "theatre" or "théâtre" out-of-the-box with at most a configuration option to tick. Admittedly that wouldn't solve the problem of American English spellings (that I, personally, would only use in proper names anyway). It should also be possible to have a search for "osaka", "Ôsaka", and "ôsaka" produce identical results out-of-the-box (if you don't use accents "Osaka" and "osaka" does). "Ôsaka", "ôsaka", when entered as keywords should be treated as being strictly identical.

[dorin_nicolaescu]

I think a more technically correct term for this issue is "diacritic sensitivity/insensitivity", similar to "case sensitivity/insensitivity".

So, it might help to give the topic a more relevant title, such as "Diacritic insensitive search".

[allen-c]

Diacritic insensitive search

Sorry, I only came across your remark today - the e-mail I received ended up in the same place as all those that tell me that I've won a googolplex of $ on the Milky Way interstellar lottery, I'm afraid...

That seems indeed to be much better. I don't now how to change the title now, though. Is it possible, or do I have to resubmit?

Thanks for your remark.

[dorin_nicolaescu]

I changed it. Also made a feature request.

[John R Ellis]

I think there's a broader bug here -- when searching, LR 3 and 4 don't ignore case of non-ASCII characters. For example, enter the Cyrllic "Б" as the caption of an image, and then search for the lower-case "б".

A clue to the cause of the bug: The SDK's function LrStringUtils.lower() is to supposed to return the lower-case version of a string. But LrStringUtils.lower ("Б") returns "Б", not "б".

[dorin_nicolaescu]

Well, actually there are two related, but different issues in this post:
a) case sensitivity — Ô vs. ô
b) diacritic sensitivity — Ô vs O.

[John R Ellis]

See this related topic:

http://feedback.photoshop.com/photosh...

[mustermark]

Halfway between "problem" and "idea" here.

If I search my keywords for "Brașov" it returns all instances of "Brașov" and "Brasov", but if I search for "Brasov" it returns "Brasov" only.

This seems inconsistent and is confusing, which is why I have reported this as a problem.

Personally, I would prefer e, say, to return all instances of è, é, ê, ë, ē, ė, ę, etc. and possibly vice versa. At the moment it is just vice versa, which I think is less useful.

[DenisPac]

Hello Community,

I've noticed recently that the case insensitive search in LRC (latest 11.0.1) still doesn't work when there is an accent on the uppercase letter.

I mean église doesn't return Église (and possibly vice-versa).

But since this is a different issue to the one that the current subject relates, as noted by @dorin_nicolaescu, I'm going to create a dedicated thread.

On the other end, specific to the current subject, I would love to have at least an option to be able to make a diacritic insensitive search, and this one is clearly a feature request).

Thanks,

Denis