I still play the games. Testing GCompris is a massive time sink. You get caught up in all the activities.

You can also turn things on its head, like

"Krita supports a wide range of tablets and drawing devices out of the box, so you won't miss expensive closed proprietary alternatives like Photoshop one bit".

👆 improvised, but you get the idea. You get to reference something the user may know (and this helps you out giving them a clear idea of what you are talking about), and you cast "the alternative" (Photoshop) in a less positive light than the free/libre software at the same time.

JSON is really, really stupidly simple though.

I think "free" is okay. If the software does come at no upfront cost, then fine, why not add that as an incentive to get people on board. They will figure out how to "pay back" sooner or later.

I can tell you a word I do avoid, and that is "alternative". It makes FLOSS items sound like cheap knockoffs, always playing catch-up with their supposed proprietary and closed equivalents, always seeking feature parity, but never really getting as good as the original. This is not the case. Most software projects, once they reach maturity, more often than not, evolve into their own thing.