Hi Dave,
Interesting that you brought this up as earlier today I was looking at this while developing standards for our OnBase configuration. My information comes from the Auto Fill Keyword Set MRG - it sounds familiar with what I have heard in the past, but as a customer, I am like you going by the information I can find, which may contradict one of the videos.
Any keyword that is set to Mixed Case Value is stored twice - once using the keyword value as entered and then again using uppercase values. So yes, configuring keywords as mixed case does require extra database space
While I am concerned about storage and minimizing the required space, my bigger concern is searching. If Mixed Case Values are selected, when searching, "Adams" is different than "ADAMS" which is different than "adams" in some databases and could affect results. There are some caveats when SQL Server is your database that should be kept in mind:
- Mixed Case Value keywords cannot be added to a Cascading Data Set. While we don't use many Cascading Data Sets, preventing access to functionality is, for me, an immediate no-no.
- the Case Sensitive searching option is not available with SQL Server, meaning that searches are always performed on the uppercase values only. I cant see the value in searching based on case sensitive values, so i would not check this box anyways, but there may be business cases where this is important.
Something else to keep in mind is that less options are often a very good thing for users. I am one of those people who might re-index a document because I saw a state abbreviation of Ca rather than the correct CA. I am not proud of myself, but I know who I am! Seeing all keywords in upper case is a bit jarring at first (for my fellow anal retentive's out there) but it is something that even I got used to quickly. That upper case is supported by technical best practices if you get push back makes upper case an easy win.
I have permanently removed Mixed Case Values from acceptable keyword configuration choices. Maybe it is not the right choice for others, but I am happy that I have the documentation for why it is a good practice now.
Let us know what your decision is - I am always interested in other configuration standards!
Thanks,
