why jugged the arrays, what is a point?

JavaScript translation

Python translation

The number triangle in the description is not properly formatted (it's collasped in one line). Either put it inside a code block or a <pre> block, or use non-breaking spaces.

the numbers '13 15 17 19' in the description appear to be a hyperlink, bur my phone thinks it's a phone number... uh, why?

Following test cases weren't specified in the description:
FixedTest1 and FixedTest2 : exponent is -1 and return value -1
RandomTest : exponent is 0 and return value is -1

Would you like to describe this edge case in kata?

"Random tests" isn't actually random while "Fixed Test 6" is actually random. WTF?

Reference solution gives wrong answer on decimal division:

a: 05.123 (DivisionTests$1Dog)
b: 0.5522646805787415 (java.lang.Double)

expected:
<9.2763491495262[624912285926180520481378582682271633102530402411313378516100503618775823206663540009879055683457036367196464443458169537372547743780499229459382742426612523024617238182244037605799144233462757653674025989951012356863678439786182108564917216567584621199336668983839808696359093142326182315479139441045480527731958407529642428852274599973755708633673002458451623983978597550795116724495694770308228501339497948003622252785125516902479729777640709735169967688091339939101762936873549428230120349063243781975428168026689551171004547656245531912447745376809747119409482095554235260426229157822585251117156469519479863914597790838726449232799534733926260409662425910973329130675204751917593311577628615009250648927042132205621294753692457762164641463734324579258531914344959156083127124960503524002989374249999701884667370963656352591974872233413213531622753825592654556726588548309254045229788515369204622883551053212205470921105634840634811816033993194806418420325668254640029303293618098066]>
but was:
<9.2763491495262[593629551836755055068284693305388123262562652858480097696459495724051168282967219252846528368931217730549405502077810291288653395875629418652745304669460671027593738311281935500784199571431765586983079990813520450912408437849540499432808480613873895967429422857471067444206115482052892494851356270314052665986243568949805766214554923469556593332826594740170487310869892888996978558510305481748186467464897341506004177617882214084667555776549700351190232904631231936513419402838361643319689256612785847709890480987497549540148370402413244842584818315622136781931031920455354287382252250295934658152238401411052249819353836571935818150823997471040323396408252429030213152223334099444018922750264099125672910814900150930561889817433582832463162629132865359544089986923324600141031535689516116597834770831251073266111291023284075067498503560508911780553722234698528907021939792631247088810573786741414118318401026557105032117573041344464762245419074783023959691683594598328285156261651127567]>

5.123 / expected value is around 0.5522646805787413137591753927, not 0.5522646805787415.

Reference solution also gives some absurd results, like 5.123 / 1.0 = 5.123000000000001.

you should be able to handle ANY type of Object I throw at you and assume that it has a String representation that is numeric

Why are you so fond of string-typed programming? This is a common theme in your katas.

When a value is an instance of Number, Number.doubleValue should be preferred over parsing the string representation. This is the common sense. Please don't ask people to do insane things with you.

The point is to look at the test cases and figure it out

Note that you have not provided all the specs in the sample tests, and we cannot read the actual tests (and why does your kata need people to bang against the actual tests to figure out the specs? Your kata isn't a snowflake, so aren't you). Hence for all practical purposes this point is moot.

Notice how punctuation (commas, periods, etc.) and case are ignored for determining whether a word has been used, but they still show up in the final String.

If punctuation is ignored, why are they replaced as well? e.g: here, becomes _____ instead of ____,, which is nonsensical.

Notice how punctuation (commas, periods, etc.) and case are ignored for determining whether a word has been used, but they still show up in the final String.

Apostrophe is a punctuation, but it is considered to be part of a word, hence he'll and hell is considered the same word.

expected:<...g()
	{
		return "S E[ ]E: "+sEE+"\n"
			+"M...> but was:<...g()
	{
		return "S E[]E: "+sEE+"\n"
			+"M...>

It is not mentioned how consecutive letters are handled during splitting. Normally it is split continuously, not after every upper case letter.

Given an array of Strings representing parts of a java class, generate a String containing code for said class. If there are an invalid number of arguments, throw a RuntimeException.

It is not explained at all what the arguments means, except an example is given but it doesn't show what "invalid number of arguments" means. Example is not definition.

Unless you can actually justify your last answer, that's definitely an issue, and you need to actually resolve the problem before closing the issue.

Course of action:

1. justify
2. if we finally agree you can close

note the specified order, please.