Deduplicate - Summation

[hobovsky]

From #39

1. Grasshopper - Summation

• 8kyu
• Allows O(n) solutions
• Satisfaction 91% with 180k+ completions
• 1 pending issue
• 52 languages, no pending translations
• Published Aug 2015, author inactive

2. Gauß needs help! (Sums of a lot of numbers).

• 7kyu
• Allows O(n) solutions. Feeds all type of rubbish as input.
• Satisfaction 84% with 13k+ completions
• 5 pending issues
• 7 languages + 1 pending translation
• Published Feb 2015, author inactive

3. Triangular Treasure

• 7kyu
• Task mentions "triangular numbers" rather than "sum of numbers from 1 to n". Seems to require O(1) solutions. Return 0 for negative inputs.
• Satisfaction 85% with 24k completions
• 2 pending issues
• 17 languages + 1 pending translation
• Published Oct 2013, author active.

---

Conclusion

• Keep 1. Grasshopper - Summation
• Keep 3. Triangular Treasure. Create a follow-up ticket about hardening its test to allow only O(1) solutions.
• Retire 2. Gauß needs help! (Sums of a lot of numbers).

[hobovsky]

I think that kata 1. is good to keep because it has a large coverage of languages and good difficulty.

I think kata 3. would be good to keep if it enforced O(1) solutions. The "sum numbers 1 to n" and "triangular numbers" might be not obviously related for total newbies, and I think that spoilers posted into kata 1. are not that relevant on this level. My proposal is to review all languages, make sure that all require O(1) solutions, harden or mark for removal these which do not, and keep the kata.

Kata 2. is the worst of the three: it's a blatant duplicate with a lot of annoying things like invalid inputs. Its description might be considered interesting, but even with the description, I think the kata is not worth keeping.

[monadius]

Languages of kata 3 which allow O(n) solutions:

• C
• COBOL
• CoffeeScript
• C# (no random tests)
• Dart
• Haskell
• Java
• JavaScript
• Julia
• NASM
• Ruby
• Rust
• SQL (O(n) solutions are more difficult than O(1) solutions)
• Swift (no random tests)

It will be easy to fix all these languages by increasing the upper bound of tested numbers.

[JulianNicholls]

Obviously, I'm biased since I am the author of Kata 3, but I will throw my vote in for it anyway.

When I wrote it (10 years ago), it was not possible to use an O(n) solution in the three original languages, but computers have moved on and an O(n) solution finishes before the timeout. As @monadius says, it would be simple to increase the upper bound to exclude the O(n) solutions.

[hobovsky]

Hi, and thanks for sharing your point of view!

The question now would be, would you be willing to go through languages of your kata and change tests in a way only O(1) solutions are accepted? Or, if its too much work for you, would you let users provide appropriate forks, and would you approve them?

[CiprianAmza]

I would say to keep 1 + 3. Having only O(1) solutions for the 3rd kata would be also interesting (right now O(n) passes the tests).

[hobovsky]

Created a follow-up issue to harden tests of Triangular Treasure: #194