Deduplicate - Find first non-repeated characters in the string

[hobovsky]

1. All Star Code Challenge #16

   • 7kyu
   • Accepts O(N^2) solutions. Only lowercase letters, non-repeated characters is always found.
   • Satisfaction 92%, ~1000 solutions.
   • No pending issues.
   • 6 languages, no pending translations.
   • Published Dec 2016, author inactive.

2. The First Non Repeated Character In A String

   • 7kyu
   • Accepts O(N^2) solutions. Only lowercase letters, non-repeated characters are not guaranteed.
   • Satisfaction 87%, ~1500 solutions.
   • No pending issues.
   • 3 languages, no pending translations. Exclusive Java translation.
   • Published Apr 2016, author inactive

3. First non-repeating character

   • 5kyu
   • Accepts O(n^2) solutions. Upper + lowercase letters with tricky case handling, non-repeated characters are not guaranteed.
   • Satisfaction 89%, ~35k solutions
   • 9 pending issues, mostly regarding insufficient tests.
   • 7 languages, 4 exclusive + 9 pending translations
   • Authored Dec 2013 by OverZealous (active)

[hobovsky]

I have totally no idea what to pick here.

I like 3. because of its requirements on case handling, and large selection of languages. But, OTOH, it's overranked and tests need improvement. I would vote to keep it and retire others if it could be done harder (accept only O(n) solutions), and would have issues fixed, but it would mean a significant effort (unless the author would step in ;) )

If you think that 3. cannot be saved, I have no good candidate to keep. Each of the kata is missing something: edge cases, performance reqs, languages... Maybe retire all and replace with one, more complete version? But then we could just fix the 3. I dont know.

Any ideas?

[Siddharth114]

which is easier from mod/editor perspective? retiring all and replacing with one complete version, or fix the third one also 3 has an active author so ig that helps?

[ejini6969]

3 is an advanced version of 1. & 2. in which both of them are duplicate to each other! So, I vote to keep 3.,

• Fix all pending issues (not the hardest since 9 languages only) with many of those issues being invalid or repeating the same thing (not enough edge cases) ~~
• I also suggest to send it back to beta as well for re-ranking (if that's doable and executable)
• handle the massive amount of translations (reject some of them with massive framework changes and lack of random tests; or fork them if fixable).
• Description needs rework too (add example tests and fix the return None part, not sure why not just return "")

Also: I don't like the idea of creating new kata with similar requirements instead of fixing existing ones due to i) author of original kata will feel disgraced as if his / her efforts are down to the drain (since he / she is active) ii) It requires massive work to add all existing languages to that new kata ( + finding people to review them) iii) New kata still needs to get pass beta approvals which might take a long time (might even not exit beta due to issues raised about duplicates with no proof whatsoever too often)

[hobovsky]

• I also suggest to send it back to beta as well for re-ranking (if that's doable and executable)

Reranking kata with this many solutions is something what kazk once mentioned as "[...] I wouldn't want to start reranking old kata with tens of thousands [of solutions] and more." so I guess it can be difficult, but we can always ask explicitly.

[akar-0]

3 is outrageously overranked and its varations / constraints are uninteresting and disconnected with its rank. 1 and 2 seem equivalent. I would favour to keep 1 because the greatest number of translation will make the transfer easier.

[Blind4Basics]

about 3 in python: I'm not sure if it's possible to forbid O(n²) solutions that are regex based against a O(n) solution working in plain python.

[EloiseRosen]

maybe we can ask if 3 can be reranked, and if yes we keep 3, and if no we keep 1 or 2?

[wtlgo]

I can't go into deeper details because I don't understand them, but by solving all of them I enjoyed the second one the most. It still makes you take into account edge cases but doesn't throw a random requirement about upper/lowercase in your face. Also, the 3rd one is IMHO overranked, because I was able to solve it with the same one-liner I used to solve 2, by just adding one statement to it.

[Kacarott]

I vote to keep 2. Enforcing O(n) in 3 seems difficult, and even then its an easy 6 at best. So between the other 2 I much prefer 2. This is also a fairly simple kata concept so I think moving/adding translations should not be difficult.