Talk:Chowla numbers: Difference between revisions

: Do we really need to agree on this kind of thing ? Optional seems more than good enough [[User:Hout|Hout]] ([[User talk:Hout|talk]]) 12:24, 11 March 2019 (UTC)

:: +1 for that. See below. If the language supports locales for numbers then OK. Writing a subroutine putting commas every 3 digits in an Anglophobe interpretation of appropriate seems worse than pointless almost stupid.--[[User:Nigel Galloway|Nigel Galloway]] ([[User talk:Nigel Galloway|talk]]) 15:50, 11 March 2019 (UTC)

:::: I agree that the excessive specification on the formatting of task output takes the focus away from how different languages best solve a task. It makes perfect sense to me to move those formatting specifications to their own separate task. --[[User:Tikkanz|Tikkanz]] ([[User talk:Tikkanz|talk]]) 20:16, 14 March 2019 (UTC)

 

::Unfortunately I later learned that the sympy library holding divisors is pure Python, but it is well regarded so I used it rather than creating my own.

::Hi Gerard, I mainly program in interpreted Python but '''don't''' want the present limits to be lowered.

 

::: Does Python use a non-interpreted "language" for its BIFs?   (In particular,   '''divisors'''.)     The reason I ask is that I wrote a home-grown   '''divisors'''   REXX function that instead of finding the divisors, and then adding them (say, with a   '''sum'''   function), I modified a version of the   '''divisors'''   code to instead of creating a list of proper divisors, I had the function just add the divisors on-the-fly (eliminating the stand-alone summation part).   I then further modified the function to be aware if the target is odd or even, and adjusted the   '''do'''   loop accordingly (along with the   '''do'''   loop increment);   that doubled the speed (or halved the computation time,   pot-tay-toe, pot-tah-toe).   Essentially, I coded a   '''sigma_proper_divisors'''   function with the subtraction of unity as being built-in by starting the summation with zero instead of unity).       Do you happen to know if Python's   '''divisors'''   BIF does that?   I would suspect that it does.   In any case, this is why I included several formulas/algorithms to calculate the   '''chowla'''   function so that programmers could choose the fastest (most efficient) algorithm.     -- [[User:Gerard Schildberger|Gerard Schildberger]] ([[User talk:Gerard Schildberger|talk]]) 23:00, 12 March 2019 (UTC)

 

::: I'm really impressed by the large perfect numbers computed with   '''Visual Basic .NET'''   using the Chowla function   (2^nd section, under   '''more cowbell''').   That's going the extra mile, by gum.   Going from roughly '''33 million''' to over '''8 billion''',   and then to over '''137 billion''',   and then to over '''2 quintillion'''.   Was the computer smoking or losing its magic smoke?     Kudos.     -- [[User:Gerard Schildberger|Gerard Schildberger]] ([[User talk:Gerard Schildberger|talk]]) 02:05, 13 March 2019 (UTC)

::::No magic cigarette, VB.NET benefits from .NET's JIT. It's not your grandma's VB6. [[User:Eoraptor|Eoraptor]] ([[User talk:Eoraptor|talk]]) 07:24, 13 March 2019 (UTC)

 

== Mistakenly posted discussion on the task page ==