Gripes: The type system seems way too complex -- why should I care in the typical case, what type of value is used to represent the numeric value 1? Why can't it give me consistent operations? For example, I do not like that booleans do not include the value 0 and 1 -- this not only perpetuates a mistaken understanding of George Boole's work, but it perpetuates the type complexity.
I am also dubious about some of the comments in the documentation about some of the features relating to "efficiency" -- given the descriptions of some of these I am dubious that these claims are based on benchmarks which compare the described implementation from a more straightforward implementation. These kinds of mechanisms have overhead and only pay back in very specific cases. I can live with their implementation of scalar extension (a dimension of 1 is automatically extended to match). I am also a bit dubious about some of the mechanisms (for example, why does dot product sum along the second to last axis of the second argument, rather than the first axis of the second argument?) Anyways, these are just grumbles -- and I doubt I will ever use numpy as a replacement for J. Still, I might find myself translating J code to numpy rather than regular python in some cases. -- Raul On Mon, Dec 10, 2012 at 4:51 PM, Edward Mokurai Cherlin <[email protected]> wrote: > On Mon, December 10, 2012 1:26 pm, Raul Miller wrote: >> I would be interested in NumPy. >> >> I have to use python in some contexts, but I have not had the time to >> study all the options, and on my own it would take me years to get up >> to speed on something like NumPy. > > Start here. > > http://docs.scipy.org/doc/ > Numpy and Scipy Documentation > > Welcome! This is the documentation for Numpy and Scipy . > > Links to User Manual and Reference Manual. > >> As for the rest of your message: I will agree with you on the need for >> a variety of perspectives on a subject, to make the subject clear. >> >> Thanks, >> >> -- >> Raul >> >> On Mon, Dec 10, 2012 at 1:10 PM, Edward Mokurai Cherlin >> <[email protected]> wrote: >>> If we are going to argue J v Python, we need to at least look at NumPy >>> (based in part on APL) and SciPy. If anybody is interested, I can >>> start those topics. > > -- > Edward Mokurai (默雷/निशब्दगर्ज/نشبدگرج) Cherlin > Silent Thunder is my name, and Children are my nation. > The Cosmos is my dwelling place, the Truth my destination. > http://wiki.sugarlabs.org/go/Replacing_Textbooks > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
