I might do a mixin varient. I have also been thinking about if I wanted to do a variant aimed at high performance for some test case. But both of those imply more life and use for the approach than what I currently have implemented.
I think my first focus should be implementing something fun in webgl. I have been playing with sketches and designs, but need to dive deeper into documentation and get a bit experimental to connect where I am right now with where I hope to go. Thanks, -- Raul On Thu, Jan 23, 2014 at 4:34 PM, Joe Bogner <[email protected]> wrote: > Cool. It feels similar to underscore (http://underscorejs.org/) which > has probably 1/10 of the ubiquity of jQuery which is still quite a > bit. It's the first library that probably comes to mind for many > people when it comes to 'make javascript more functional'. It's very > approachable. I've also seen bilby.js (http://bilby.brianmckenna.org/) > thrown around - it's like haskell for javascript. It's a bit too > extreme for me. > > I like that yours has similar method naming to J. If you wanted to, > you could also extend underscore to do that > http://underscorejs.org/#mixin > > The chaining in yours is also implement in underscore (from the docs): > > var stooges = [{name: 'curly', age: 25}, {name: 'moe', age: 21}, > {name: 'larry', age: 23}]; > var youngest = _.chain(stooges) > .sortBy(function(stooge){ return stooge.age; }) > .map(function(stooge){ return stooge.name + ' is ' + stooge.age; }) > .first() > .value(); > => "moe is 21" > > Thanks for sharing > > On Thu, Jan 23, 2014 at 4:23 PM, Raul Miller <[email protected]> wrote: >> Here's a gist of the current draft: >> https://gist.github.com/anonymous/8587032 >> >> I must admit that I do not have very extensive tests yet, and I may >> mistreat some edge cases. I am not sure if that matters. I'll paste my >> tiny "test suite" at the bottom of this message. >> >> Thanks, >> >> -- >> Raul >> >> <html><head><title>tests for a.js</title></head> >> <body> >> <pre id="out"></pre> >> <script src="a.js"></script> >> <script> >> window.out= function(t) { >> var log= document.getElementById('out'); >> log.innerHTML+= t; >> return t; >> } >> function maintain(f) { >> var r; >> try { >> r= f(); >> } catch (e) { >> var t= f+' failed '+e; >> alert(out(t)); >> r= A.one; >> } >> if ((r instanceof A) && 1 === r.DATA[0]) {out('.\n');} else {out(f+' >> failed '+r+'.\n');} >> } >> window.v123= A(3).indices().plus(1); >> window.m1to9= A([3,3]).reshape(A(9).indices()); >> maintain(function(){return v123.times(2).match(A([2,4,6]))}); >> // maintain(function(){return v123.take(2).match(A([1,2]))}); >> maintain(function(){return >> m1to9.dot('plus','times')(m1to9).match(A([3,3],[15,18,21,42,54,66,69,90,111]))}); >> maintain(function(){return >> m1to9.dot('plus','times')(A([3,2],[1,2,1])).match(A([3,2],[5,3,17,15,29,27]))}); >> maintain(function(){return v123.dot('plus','times')(v123).match(A(14))}); >> maintain(function(){return v123.match(A([1,2,3]))}); >> maintain(function(){return v123.reduce('plus').match(A(6))}); >> </script> >> </body></html> >> >> >> On Thu, Jan 23, 2014 at 12:50 PM, Joe Bogner <[email protected]> wrote: >>> Raul, your library sounds interesting. Please share more if you can later. >>> >>> Coffeescript is a fun language. The lambda syntax and lack of braces >>> make it feel closer to the problem domain than native javascript. It >>> also makes it harder to shoot yourself in the foot. The debugging and >>> compilation has greatly improved over the years too. The compiler >>> throws useful errors with line numbers etc. >>> >>> I agree that toys are useful for learning. It's common in the land of >>> lisp to write a toy interpreter to better understand the language. I >>> did that too and found putting this one together a nice break >>> >>> On Thu, Jan 23, 2014 at 12:12 PM, Raul Miller <[email protected]> wrote: >>>> I've been building an vsapl-like javascript library. I want that so >>>> that I can code webgl without having to round-trip to jhs for minor >>>> array manipulation. It's pretty simple since I only support one >>>> numeric type (javascript numbers) in the arrays. So basically it's a >>>> functional object library with a shape/data pair representing an array >>>> - so for webgl arrays I basically can just extract the data and use it >>>> "as-is". (I say "functional object library" because I use a >>>> javascript object to represent each array, and monadic functions are >>>> niladic methods ("this" is the right argument) which return a new >>>> array while dyadic functions are single argument methods which also >>>> return a new array ("this" is the left argument). Or at least, that's >>>> how I currently have it wired up.) >>>> >>>> I'm not using git for my versioning, yet, because it's so trivial I >>>> work better at this stage with a file for each version. I should >>>> probably just stop working on it, and get back to coding webgl, but >>>> coffeescript looks fun and I would like to compare code size and >>>> simplicity between a native javascript implementation and a >>>> coffeescript implementatin). >>>> >>>> I am using the vsapl semantics because they seem simpler to implement >>>> for the subset of functionality I need (numeric operations on small >>>> arrays, when using javascript) than the full J semantics. Also, >>>> browsers have intentionally crippled functionality so any significant >>>> use of the library would want jhs for heavy lifting. >>>> >>>> Anyways, toys are good for learning, and I've been struggling with >>>> concepts of what I should be doing that's worthwhile for other people, >>>> and if nothing else, this is a sometimes relaxing effort. >>>> >>>> Thanks, >>>> >>>> -- >>>> Raul >>>> >>>> >>>> On Thu, Jan 23, 2014 at 11:56 AM, Joe Bogner <[email protected]> wrote: >>>>> Thanks for the feedback. I just added in the insert adverb and pushed >>>>> the code. It makes it feel more like APL now since I can do the >>>>> prototypical +/1,2,3,4. >>>>> >>>>> Implementing forks might be up next (fun). Then I can do the J "hello >>>>> world" of (+/%#) 5,10,15 >>>>> >>>>> On Thu, Jan 23, 2014 at 11:41 AM, Raul Miller <[email protected]> >>>>> wrote: >>>>>> I like this, and I think it is time I learn coffeescript. This relates >>>>>> to some other efforts of mine. >>>>>> >>>>>> Thank you, >>>>>> >>>>>> -- >>>>>> Raul >>>>>> >>>>>> >>>>>> On Thu, Jan 23, 2014 at 11:20 AM, John Baker <[email protected]> wrote: >>>>>>> Interesting. I don't know cofffescript but it's not hard to figure out >>>>>>> what's going on here. >>>>>>> >>>>>>> >>>>>>> On Wed, Jan 22, 2014 at 9:22 PM, Joe Bogner <[email protected]> wrote: >>>>>>> >>>>>>>> I ported Arthur Whitney's one page APL interpreter fragment[1] to >>>>>>>> coffeescript (which enables it to run in javascript) as learning >>>>>>>> exercise. >>>>>>>> >>>>>>>> http://csilo.com/dev/apljs.html >>>>>>>> >>>>>>>> code is here: https://github.com/joebo/apl-js/blob/master/apl.coffee >>>>>>>> >>>>>>>> This could be extended to create in-browser games/learning exercises. >>>>>>>> Of course, it would also make sense to change the symbols to J. I kept >>>>>>>> the original ones for historical reasons for now. >>>>>>>> >>>>>>>> [1] - http://www.jsoftware.com/jwiki/Essays/Incunabulum >>>>>>>> [2] - http://www.jsoftware.com/papers/AIOJ/AIOJ.htm >>>>>>>> ---------------------------------------------------------------------- >>>>>>>> For information about J forums see http://www.jsoftware.com/forums.htm >>>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> -- >>>>>>> John D. Baker >>>>>>> [email protected] >>>>>>> ---------------------------------------------------------------------- >>>>>>> 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 >>>> ---------------------------------------------------------------------- >>>> 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 > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
