Voileexperiments
commented
7 years ago Note that this not a good idea at the moment because CCC is pretty outdated, and many information has either become obsolete or just not right. For example:
- you don't use
arguments.callee.caller.caller.caller.toString();to get user code anymore - floating point test: there is really no "golden rule" for the amount of precision required. For example, for numerical computations (e.g ODE and well-known mathematical functions) even 1e-4 might be impossible. This largely depends on the problem at hand and requires extensive testing (which many kata authors don't do)
- Haskell submission sometimes just timeouts for no particular reasons (seems to depend on server condition?)
My set of rules of thumb are, instead, these:
- if you're writing a kata/doing a translation, make sure you know the language you're writing well before doing/trying anything! (Especially translations, we all know how many translations are broken)
- read the docs (speaking of which, we should really finish the new documentation project)
- get other people to peer-review when in doubt, e.g you can't really catch bugs in your author solution easily if only you have solved it and it's a very complicated one
- note that different languages exhibits different class of problems. Languages like Haskell are facing first-world problems: because people writing stuff in Haskell has to be quite skillful in the first place the problem are usually minor and accidental. However, the biggest languages around (I'm looking at you, JS) exhibits third-world problems: quite often, people writing JS katas are not skilled enough, don't know about common CS concepts (e.g what is a float? And did you know every number in JS is represented by a float?), don't know about duplicates, don't know about basic language features (a common one is that
Function.prototype.toStringcan be overridden easily. And then of course, those ES6 syntax), don't know about other commonly used stuff like regex. so on and so on... If we're writing a guide for all users we need to account for the average proficiency of people writing and translating katas.
bkaestner
I know this is a rather bizarre suggestion but I'd like to recommend that a pop-up notification be made for Codewars users who are authoring their first ever Kata on kata authoring best practices (e.g. following PEP8 / language conventions, including edge cases and random tests etc.) so they are aware of what constitutes a good, high-quality Kata and what constitutes a crappy Kata which is bound to receive Issue Reports, reason being that it is quite tiring and sometimes frustrating when us Power Users / Moderators strive daily to ensure that the Kata quality on Codewars is top-notch (more so for JavaScript than any other language which by far is the most popular) by authoring high-quality Kata and authoring top-notch Kata Translations, only to have our efforts undermined by wave upon wave of new, inexperienced users on Codewars who may author like 10 completely worthless Kata daily (OK, maybe I'm exaggerating but you get the point :wink:). And while we can raise Issues with detailed explanations on what needs to be improved, there's always a chance that some Codewars user who has newly unlocked the privilege of resolving external Issues but isn't yet familiar with the mode of work here comes in and resolves such issues prematurely without fixing it or providing a valid reason to do so which in some cases may lead to another Moderator accidentally approving it having overlooked the (falsely resolved) Issue. In particular, I would recommend that the Conjured Codewars Codex be included in some way in this Notification. Although the author of this Codex (@bkaestner) admits that the rules aren't formal, this article is very well written and I believe that every Kata author should at least be aware of it and hopefully consider it when authoring any Kata which would greatly reduce the proportion of worthless (JS) kata authored so fewer issues need to be raised and the community would be more peaceful as a result.
┆Issue is synchronized with this Clickup by Unito