Bob, Thanks for the tip. It would be nice if there was a packages table of contents on the Jsoftware.com website, with links to each package as well as links to all the defined verbs in each package. Then one could track down useful verbs using the Jsoftware.com website search bar. Perhaps having a separate search engine for packages and defined verbs on the website would be even a better approach. When working on a project, I invariably find myself searching to see if the function I need has already been developed, or whether I will need to build it myself. If I build it myself, other more experienced forum members usually can improve on it, making it faster, more space-efficient, and often more general.
Windows' explorer search is excruciatingly slow, even on my new i7 machine with SSD. It searched for the keyword rand11 on my system for several minutes, but never found anything. On a related topic, it could be useful if there was a way for J forum members to submit new or improved verbs to be considered for inclusion in existing packages, or even to propose new packages that members could contribute to, such as machine learning, or NLP. I have seen many useful verbs discussed and polished in the forum, but those verbs then get buried in old forum posts, undocumented, and mostly forgotten. The partition verb discussion currently going on in the forum is but one of many such examples. Perhaps members could vote on submissions to be included in packages, or whether to start a new package. Contributors would be acknowledged in the verb definitions. I use Google forms (free) for voting schemes like this (https://goo.gl/Wgr9mx). It auto-summarizes votes in a spreadsheet. New mathematical approaches are being developed every day, but the J package library has been fairly static for some time. Particularly the fields of machine learning and natural language processing are discovering new approaches and methods at a rapid pace (neural nets, deep learning, word2vec, doc2vec etc.). J is particularly suited to concisely handle the dominantly array-centric algorithms in these fields. However, Python seems to be the choice for developing algorithms in these fields, even though J has much more concise and flexible array primitives than Python. Skip Skip Cave Cave Consulting LLC On Sun, Oct 29, 2017 at 10:34 AM, robert therriault <bobtherria...@mac.com> wrote: > Hi Skip, > > I did a search with the finder on my mac for the text "rand11=" and those > were two of the files that were shown to contain it. It is the trick that I > use to find the origins of verbs that don't show up in the usual ways. > > Cheers, bob > > > On Oct 29, 2017, at 12:23 AM, 'Skip Cave' via Programming < > programm...@jsoftware.com> wrote: > > > > Bob, thanks for the link. > > I searched the jsoftware.com website for the rand11 verb, and none of > the > > hits showed the code, or what package the function might be in. I just > > found a brief description, with no indication what package the verb might > > be in: > > > > rand11 random <http://www.jsoftware.com/docs/help602/user/script_ > random.htm> > > v generate y random numbers in interval (_1,1) > > Where does one find the list all the packages that are included in the J > > distribution, both standard & optional, along with the verbs in each > > package. Suppose I know the name of a verb in some package (usually from > > reading someone's code), how do I know which package that verb is in, and > > whether the package must be "loaded" (like stats), or "required" (like > > plot). > > > > Skip > > > > > > > > > > > > Skip Cave > > Cave Consulting LLC > > > > On Sun, Oct 29, 2017 at 1:20 AM, robert therriault < > bobtherria...@mac.com> > > wrote: > > > >> I found it in j602 in this path ~/j602/system/packages/stats/random.ijs > >> > >> setrl=: 9!:1 > >> > >> deal=: (# ? #) : (? #) { ] > >> dealx=: ] #: ?~ : ? & (*/) > >> toss=: ? @ (# #) { ] > >> tossx=: ] #: (?@# */) > >> rand01=: ?@$ 0: > >> rand11=: <: @: +: @: rand01 > >> > >> and it remains in the newest version of the j beta > >> ~/j64-806/addons/stats/base/random.ijs although with more documentation > >> > >> cocurrent 'z' > >> > >> setrl=: 9!:1 NB. set random link > >> > >> deal=: (# ? #) : (? #) { ] NB. deal x items from y (no repetition) > >> dealx=: ] #: ?~ : ? & (*/) NB. deal x items from shape y (no > repetition) > >> toss=: ? @ (# #) { ] NB. pick x items from y (with replacement) > >> tossx=: ] #: (?@# */) NB. pick x items from shape y (with > >> replacement) > >> rand01=: ?@$ 0: NB. generate y random numbers in interval > (0,1) > >> rand11=: <: @: +: @: rand01 NB. generate y random numbers in interval > >> (_1,1) > >> > >> > >> Cheers, bob > >> > >> > >> > >> > >>> On Oct 28, 2017, at 10:46 PM, 'Skip Cave' via Programming < > >> programm...@jsoftware.com> wrote: > >>> > >>> I found this verb description in the Jsoftware website" > >>> > >>> rand11 random <http://www.jsoftware.com/docs/help602/user/script_ > >> random.htm> > >>> v generate y random numbers in interval (_1,1) > >>> > >>> Bu I can't find the actual code implementing the rand11 function > >> anywhere > >>> on the Jsoftware site. Can anyone point me to the code? > >>> > >>> Skip > >>> ---------------------------------------------------------------------- > >>> 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
