What I would love to see as an English speaker that do not use diacriticals is a change to the search parameter on the program so when I search Penelope Cruz, it will also find Penélope Cruz and vice versa.

Quoting northbloke:

Quote:
I have a few Polish films I've been unable to properly profile because I can't enter the information correctly, so yes I'd love to see Profiler updated to cater for all languages!
And not only do I agree that diacriticals should be ignored for searches, I think it would be very helpful if they were ignored for name linking too.

+1

I'm not that experienced in that area but I think that is very hard to implement, especially when you don't have the support of the underlying database.

You'd basically need a mapping table of all letters that could be another letter. It's not just e, é, è and ë, there are ç, c, c and all other kind of stuff even this forum can't handle. And that's just the central-european languages that use an extension of the latin alphabet.

So, while not dismissing the idea, I wouldn't get my hopes up.

Does anyone know how it works in other programs? Google seems to ignore diacriticals in its searches, I wonder how they get it to work.

Quoting northbloke:

Quote:
Does anyone know how it works in other programs? Google seems to ignore diacriticals in its searches, I wonder how they get it to work.

Diacritics are usually associated within the program's "character map" (so-to-speak), the method for which can vary depending on the purpose of the terms being used.

An example of character association are "case sensitive" letters.

Ever notice how some websites say that your username and password are case sensitive while others don't? That is a result of (or lack of) associating the letter "A" with the letter "a" and so on through the alphabet. A site (eBay for instance) registers my username "Blair" in their database, exactly how I typed it. Associating capital and lowercase letters has a two-fold benefit. It means I can go back and [may be able to] type my username in as "bLaiR" if I want, and I would still end up logged in as "Blair"... exactly how I typed it in during registration.  It also prevents anyone else from coming along and being able to create the username "blAir" or a similar variant. This association is also why/how when you type using a word processor or often in texting that the software automatically capitalizes the first letter of each sentence. ("Let's see, this is the first letter and he typed 'a', so I need to make it an "A.'")


For searches, diacritic association works in a similar but usually more complicated way. In the most basic style (Google is a very bad example because their system is far more complex) you type in the word "fiancee" (without the accent) and all at once the system looks for every diacritical variation of the word that it can [find in the database], returning them all. "Fiancee" will likely end up at the top of the list but further down you will eventually find "fiancée" (with the accent). The same works in reverse. Type "fiancée" and those results come up with "fiancee" somewhere in the list.

Dealing with diacritical variations in this manner means more complex searches and therefore more time to perform them.