HHC 2017 RPL Programming contest information and results thread

[Gene]

Winning RPL program at HHC 2017 by Bill Butler
71.5 BYTES #4BF0h

Gene: Oops. Left the "." off after the 12 below. Corrected.

'EGP'

<< { } SWAP DO DUP INV CEIL ROT OVER + UNROT UNTIL INV - DUP -9. ALOG < END 12. ALOG * SWAP >>

[3298]

Odd, I can only reproduce that size when my calculator is in exact mode - when it's in approx mode as the rules say, the integer 12 is compiled to the real number 12. instead, increasing the program size to 75.5 bytes. On the other hand, when it's compiled in exact mode to get the integer 12, it works just fine even in approx mode. My 6. ALOG SQ in place of 12 ALOG would have been even shorter, by the way.
Other than that, it's remarkably similar to what I and a few other participants came up with (just without both "evil judge protections" to shed some bytes (analysis in my previous post, one of them actually improved the score), and with a few commands reordered in a way I didn't find which reduced the size by another 2.5 bytes). It seems the majority of us gravitated toward this solution, with varying levels of success.

Again, thanks to Claudio for going against the trend with his unique and very effective approach. I'll treat that as a lesson for future contests to think harder on the less straightforward paths towards a solution. I think I want to take that solution and improve on it just for fun. But not right now, I need to get some sleep in the next 6 hours or so.

[brickviking]

My poor old HP50G ... didn't like one number I fed Bill's solution...Pi.

Oh. And Claudio's really long program is sstill chewing on it about 50 minutes after I started. How would I fix that?

<evil_grin>

EDIT: corrected from 2 minutes to 5 to 50 minutes.

EDIT2: It eventually coughed up a two-line answer that looks like it has a lot of parentheses (weird for a RPN calculation) scattered with CEIL and pi symbol and 1/ and - and ... well, lots of stuff. Yeesh. And I put my 50G through that for nothing.

Regards, DrSmokey
(Post 91)

[Claudio L.]

(09-23-2017 10:58 PM)brickviking Wrote:  My poor old HP50G ... didn't like one number I fed Bill's solution...Pi.

Oh. And Claudio's really long program is sstill chewing on it about 50 minutes after I started. How would I fix that?

<evil_grin>

Easy: install newRPL. Each answer took roughly one second.
<end of shamelessly plugging newRPL to another unsuspecting victim>

(09-23-2017 10:58 PM)brickviking Wrote:  EDIT: corrected from 2 minutes to 5 to 50 minutes.

EDIT2: It eventually coughed up a two-line answer that looks like it has a lot of parentheses (weird for a RPN calculation) scattered with CEIL and pi symbol and 1/ and - and ... well, lots of stuff. Yeesh. And I put my 50G through that for nothing.

Regards, DrSmokey
(Post 91)

If you got symbolic output it means at some point you had symbolic input. Perhaps you ran it in exact mode? Perhaps you compiled it in exact mode?
It's supposed to take a while, but not that much, and it's 100% numeric, nothing symbolic anywhere.

[Gene]

(09-23-2017 10:58 PM)brickviking Wrote:  My poor old HP50G ... didn't like one number I fed Bill's solution...Pi.

Note: All numbers are to be between 0 and 1, not inclusive. No 3.14... numbers to be fed. See if Bill's solution works for 0.1415... :-)