a/2 reducing to 0.5a is surprising

[dlfivefifty]

I would expect it to symbolically represent rationals.

[MasonProtter]

Hmm. That’s a good point. One thing I was thinking is that if one wants rationals they can just use // but that is a bit uglier syntactically and I’d have to make a new Rational type that works with SymExprs anyways, so maybe it’s best to just default to a symbolic rational.

As you can see, I haven’t worked on this package in many moons, but I’m planning to dive back in soon with some major overhauls, so I’ll add this to my list.