[olug] Unicode in Databases

[Mike Hostetler]

This is a bit off topic (not Linux-related) but I've been scratching my
head over this one . . . though it reveals my lack of knowledge in
databases

I'm researching how to internationalize one of our database-based
products and am wondering about the size of the records.

Example:  let's say I have a database that can save UTF-8 data.  One of
the tables has a record defined with char(20).  That record then can
contain 20 characters, regardless if it's ASCII (1 byte/char), Greek (2
bytes/char), or Chinese (3 bytes/char).  My question is: how many bytes
does that char(20) take up?  Is it 20, 40, or 60?  Or does it depend on
what DB we are using?

It seems like an elementry question, but Google shows no elightenment --
just confusion on this matter. 

-- mikeh
 
--