Why don't you consider having user-inputted information rather than relying on freely available sources (which are outdated and possibly incorrect)? I know you couldn't possibly have every single verb ever that way, but really the tool is mainly used for the more commonly used verbs? Or a mixture of both? (eg. using the free resources to populate the database, but then having someone verify the information/fix mistakes prior to it being published).
Regarding the highlighting, I just realised that I can in fact highlight them, but it may be that if there's customised highlight colour (through CSS) that its not showing correctly in my browser and its not defaulting to the browser default. Or possibly the colour is too similar to the background to be seen on some screens? I'm using Firefox in Windows 7 currently if that helps.
I just noticed in the screenshot you provided with the drop down that there was a participle listed. Did I miss that originally, or something else? lol
I'm just being picky here, but I also find the hover-over bubble that says (for example) "The future forms of the verb skriva in Swedish" to be a bit redundant and obstructive. I mean, there's already a heading for the section that says "Future tense", and it's clearly Swedish, so why the need to repeat this in a pop up bubble?
Regarding "lär", yes "learn" can be used like that in English, but I feel it'd be more thorough to include the transitive forms also, since these don't always mean the same thing. Eg. "ge" means "to give", but "ge sig" means to "give up/surrender". (That and my main point about "lär" was that it was missing the final -a for its infinitive form - probably another symptom of using free sources
). And just to complicate it more, there are prepositional verb combinations like "ge sig av" (which means "leave, depart, go away from").
Just curious, how are the various conjugations stored? Are they static in a database (ie. you have "skriva" and all its forms saved), or does it conjugate them on the fly based on known patterns, but with irregular ones statically defined? I'm not sure what your plans are with other languages, but consider for example Japanese, where there are significantly more conjugation possibilities, but where they are 99% regular. You could save yourself a considerable amount of work if they are produced dynamically.