depends on what your applciation is, and what its going to run on.
In the past I have implemented single language per db, but more recently I prefer multi language in one global db. Lots of reasons
-maintainability (any change to the app happens once in once db, one version to roll out/support), bug tracking is easier....
-extensibility (if a customer wants a new language we/they 'just' provide the translations and the db is then extended, if a new language is required its an addition ot the existing product
-some sites/apps are multilingual so there is a deployment benefit in that the app doesn't care who is running it it can switch to the desired language (this can be triggered by user id or selecting icons (eg flags or from a combobox)
some downsides
you have to be very carefull with display sizes (esepcially with text boxes & command buttons, as some languages (mostly non English European languages can take a great deal more space than English)
you may have a problem with non alphabetic (A-Z) languages (eg Chinese, Japenese, Arabic etc...), some of which display bottom to top, right to left. Fortunately I haven't had to deal with this yet so have always avoided the issue.
takes more resurces and the apop my well be slower / more processor/network hungry (some of that can be minimised by utilising local tables to store the local language if your app permits it)
it depends on your development environment / front end. But most environments allow you to populate virtually any form or report at run time. Its pretty easy in Acccess,
VB & .NET. You need to either calucalte the maximum text width at load time to adjust the display size (either font/font size or display width)
if you are developing in
VB .net C++ then you could use resource tables to handle the application translation, but you still have the requirmeent to translate your data
Its going to take much longer to design and test the application, but I would honestly say if you are going to support more than one language then unless your customers have the budget I wouldn't go down the route of separate language per db.