I published recently in JOSS and would agree with your thoughts. JOSS is exactly the right kind of journal for your software I guess - like me you probably don't want to waste time writing a long formal paper when you have already documented the hell out of everything. Many academics are still stuck in a traditional way of thinking and presume JOSS is not a "real journal" because it takes an unconventional approach. JOSS is also relatively new on the scene, hence some of the negative opinions here, but my experience was a thorough review process that actually digs into the code. Sure, I could have probably aimed for JSS or something like that, but frankly I'm also fed up with a lot of the bullshit that goes along with traditional academia, having been in the game for some years! I would try to talk your supervisor around, honestly.