L2/10-210

Proposal: Add informative text to UTS#10 documenting the use of asymmetric searching
From: Mark Davis, Peter Edberg
Date: 2010-06-10

Users often find asymmetric searching to be a useful option, and it would be good to describe that in UTS#10. Here is draft text:

Users often find asymmetric searching to be a useful option. With asymmetric searching an unmarked character in the query will match a marked or unmarked character in the target, but a marked character in the query will only match the same marked character in the target. This is easiest seen with the following examples, where a lowercase query character matches that character or the uppercase version of that character, and an unaccented character matches that character or any accented version of that character.

Query     Target Matches
resume    resume, Resume, RESUME, résumé, Résumé, RÉSUMÉ, rèsumè, ...
Resume    Resume, RESUME, Résumé, RÉSUMÉ, ...
résumé    résumé, Résumé, RÉSUMÉ, ...
Résumé    Résumé, RÉSUMÉ, ...

This can be implemented by masking off the secondary and/or tertiary weights in the target (depending on the strength setting), wherever the corresponding collation elements in the query have the default values.