@Szabolcs I heard that more and more in Mathematica is programmed in the WL (instead of their C derivatives, C++, etc
I guess that with the long promises of a functional compiler, this direction makes sense.
But what might happen is also the development of code that might not be very optimized now, as there’s a promise of automatic optimization for the near future. This can eventual cause a prioritization of functionality addition, instead of speed optimization. I’m, as always, just supposing... and also, not saying that this might be the justification for any particular case.
Just that it might be a possibility in some cases...
@JasonB. Oh now that's a fun idea. If you can get it working nicely it'd be nice for battle-testing some of my bond-guessing code. PubChem unfortunately had too much cruft to simply random sample from there.
Oh and since I'm on this topic are you planning on supporting z-matrix inputs? I've been updating my z-matrix importer lately so it's on the brain.
@VitaliyKaurov @halirutan It seems that the code characters is misreported ... I have 64 but it reports 230. The best solution is supposed to have 35 only, but just the part before the := takes up 30 characters as given! I don't think it's possible to solve it in 5 more.
Another problem is that the solutions are not always properly tested. I think if one wants to get the shortest solution, it'll pay off to try to game the system and find the holes in the tests. But that makes the whole thing not fun.