Anonymous user
Talk:24 game/Solve: Difference between revisions
→Use 3 3 8 8 in example runs?: added some comments.
No edit summary |
(→Use 3 3 8 8 in example runs?: added some comments.) |
||
| (7 intermediate revisions by 4 users not shown) | |||
Line 42:
Hi, After finding out that the digits 1 1 2 and 7 ''need'' the use of two sets of brackets for its (infix) solution (see above), maybe all examples might like to show how they handle this case? --[[User:Paddy3118|Paddy3118]] 06:22, 6 January 2010 (UTC)
: The OCaml solution fails on this example as well as 2 2 5 7. --[[User:Andrew Gacek|Andrew Gacek]] 05:32, 27 July 2010 (UTC)
== Negating the first number? ==
I was thinking about making a solution for this in ABAP, and I was wondering if a negated first number is valid?
E.g.
-6 + 6 * 6 - 6
I realise 6 + 6 + 6 + 6 is an alternative solution for this, but I'm curious nonetheless.
--[[User:Rjf89|Rjf89]] 00:28, 6 January 2011 (UTC)
: Sounds fine to me, although thats leads to the question: "Is there a set of digits where a leading minus is necessary"? I don't think it is necessary - but do think it is allowed. --[[User:Paddy3118|Paddy3118]] 06:04, 26 January 2011 (UTC)
:: This leads to another question: what operations may be considered part of an [http://www.webopedia.com/TERM/A/arithmetic_expression.html arithmetic expression]?
::: wp mentions ''"Addition, subtraction, multiplication, or division, and sometimes other operations"'' so there are probably variants that use a different mix of operators, but our game specifically mentions just those first four. --[[User:Paddy3118|Paddy3118]] 21:23, 27 January 2011 (UTC)
== Use 3 3 8 8 in example runs? ==
I have just updated the Python example to use precise rational arithmatic whenever a division is involved as the solution with the digits 3 3 8 and 8 of 8 / (3 - 8 / 3) does not evaluate to 24 when using floating point. You might have to switch too (although the task description does not mandate this)! --[[User:Paddy3118|Paddy3118]] 06:00, 26 January 2011 (UTC)
: I just noticed that my REXX program also suffered the same weakness. I have corrected the comparison logic and added support to find solutions for such numbers (that needed rational arithmetic or something similar). I also agree that '''3388''' should be one of ''those'' numbers to be tested for program validation. -- [[User:Gerard Schildberger|Gerard Schildberger]] ([[User talk:Gerard Schildberger|talk]]) 09:22, 1 January 2019 (UTC)
| |||