Seems to me that a more important conversation is why the authors of Julia did not consider J either as a technical computing platform or as a performance comparison.
Julia has a lot of very good design put into it. — Sent from Mailbox for iPhone On Tue, May 7, 2013 at 2:25 PM, R.E. Boss <[email protected]> wrote: > I'm glad you agreed on the fact that 1000 milliseconds = 1 second. > (and not "1000 seconds * 1 millisecond" which would give "a dimensionless 1", > which seems nonsense to me in any interpretable way.) > That is what I was stating, but obviously not clear enough (for you) to > interpret it right. > And it was also in line with the remark made by Sherlock, who was assuming > that Wang expected the result of 6!:2 in milliseconds. > I don't expect to understand ever any explanation of how "multiplying > milliseconds by 1000 would give microseconds", as you wrote. > R.E. Boss >> -----Oorspronkelijk bericht----- >> Van: [email protected] >> [mailto:[email protected]] Namens Raul Miller >> Verzonden: dinsdag 7 mei 2013 19:36 >> Aan: Chat forum >> Onderwerp: Re: [Jchat] [Jprogramming] J on Julia benchmark >> >> On Tue, May 7, 2013 at 11:43 AM, bob therriault <[email protected]> >> wrote: >> > So, 0.1 seconds * (1000 milliseconds % 1 second) would give you 100 >> > milliseconds and the original seconds in the numerator and > the >> conversion seconds in the denominator cancel out leaving milliseconds (and >> confirming that this is the result that you want). >> >> Put differently, 1000 milli- (or 1000 milliseconds % 1 second) is >> equivalent to the numerical value 1. >> >> On Tue, May 7, 2013 at 12:05 PM, R.E. Boss <[email protected]> wrote: >> > Milli is a prefix which stands for %1000. >> >> Agreed. >> >> > So if I multiply 1 millisecond with 1000 I get (*%)1000 seconds. >> >> I still have a problem with this notation. >> >> 1 millisecond is 0.001 seconds. >> >> ((*%) 1000) is 1000 * % 1000 so I would expect that (*%) 1000 seconds >> to be 1000 seconds * % 1000 seconds - in other words, that looks to me >> like a dimensionless 1. >> >> Meanwhile, it's indeed the case that 1000 * 1 millisecond would be 1 >> second. Here, though we are not changing the units being used to >> express the original value. Instead, we are finding a new value. >> >> But that does not seem, to me, relevant to the original post, where >> the units provided by 6!:2 are seconds and the context involved >> comparing timing from 6!:2 with other times which were apparently >> expressed in milliseconds. >> >> It's the difference between "x = 1 second, what is x in milliseconds" >> and "x = 1 second, how long would a thousand repetitions of x take". >> Both can be valid questions, but only one seems valid in the context >> of 1000 * 6!:2 >> >> > If I divide it by 1000 I get (%~%)1000 which is a micro second. >> > That's what I wrote and (IMO obvious) how it should be interpreted. >> >> It may be obvious to you, but I am having trouble fitting these >> observations into the original context of this thread. That's what's >> confusing me. >> >> -- >> Raul >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
