Devon - thanks for the link, I'll take a look at Beau's work! Eric - I've explored that route a bit, but a more robust solution, perhaps one without necessitating serialization for cross-task communication, would be ideal. I'm actively exploring the language Rust for a more general solution to this set of problems, and have been pretty impressed so far. That said, J's definitely a tool I'm keeping around, if for naught other than for prototyping and research.
Thanks! -- Ian On Tue, Jul 9, 2013 at 1:32 PM, Eric Iverson <[email protected]>wrote: > Not sure what direction you should explore. I am interested, but > unfortunately am too busy with other stuff. > > A quick suggestion is for a brute force threading by having multiple J > sessions and having different browser tabs/windows using different J > sessions. This would allow, for example, your slow data collection > session to be completely separate from your interactive user session. > Having multiple J tasks serving different ports is quite easy and > opens up some possibilities. > > On Tue, Jul 9, 2013 at 12:42 PM, I.T. Daniher <[email protected]> > wrote: > > JHS is a really neat concept for doing interactive data exploration, > > especially with tools like https://github.com/jordantirrell/D3-for-J and > > 'graphics/plots' canvas backend. > > > > Most of the data with which I work is grabbed live, off of hardware, via > > various bindings I've been writing. ( > > https://github.com/itdaniher/JNotebook/blob/master/libusb.ijs, > > https://github.com/itdaniher/JNotebook/blob/master/cee.ijs, > > https://github.com/itdaniher/JNotebook/blob/master/librtlsdr.ijs) > > > > Unfortunately, the collection of data isn't always a low-overhead task, > so > > I've largely been stuck with vim and jconsole, where my prototyping > > workflow consists of writing an experiment, running the script, and > viewing > > / processing the resulting data. This is perhaps more a petition for > better > > threading / async support in J in general, but I've found JHS and the > other > > J IDEs to be ill suited to a more realtime/interactive paradigm due to UI > > requests, and in some cases, interaction with the text editor window > > itself, being blocked by "long-running" data collection. > > > > For an example of the sort of experiments I'm working to facilitate, > check > > out http://www.nonolithlabs.com/blog/2012/09-19-software-features, an > old > > post from my now largely defunct startup, which was working to provide > > makers and learners with tools to poke and prod the world at the level of > > voltage and current, and interpret the results to build better > > understanding of electronics and physics. We worked extensively to build > a > > very tight and low-latency connection between physical hardware and > > graphical representations of information streams. J would be a tool to > > expand that paradigm to involve mathematical stream processing, but I > > haven't been able to figure out how to realize the necessary concurrency > > bit with J, as-is. > > > > I started learning J about two months ago, inspired by a friend who > worked > > with IBM to standardize early drafts of APL. I love the language, but > it's > > clearly not a systems language, and the lack of async / threading > paradigms > > has made it difficult to fully realize some of the concepts I was hoping > to > > explore. I'd love advice on how to proceed! > > > > Thanks, > > -- > > Ian > > > > On Tue, Jul 9, 2013 at 12:22 PM, Eric Iverson <[email protected] > >wrote: > > > >> JHS is single threaded. The javascript code / ajax request / J code > >> all run in the same thread. This was easy and convenient for the > >> original JHS proof of concept. > >> > >> This would be easy to relax in many different ways with multiple > >> threads in javascript. Did you have particular requirements in mind? > >> There are so many possibilties that it would help to have a concrete > >> project to think about. > >> > >> On Tue, Jul 9, 2013 at 11:55 AM, I.T. Daniher <[email protected]> > >> wrote: > >> > Question I've been meaning to ask about JHS - AFAIK, the code written > >> runs > >> > in the same thread that handles the AJAX requests. If this is indeed > the > >> > case, is there any plan to change the situation? > >> > > >> > Thanks! > >> > -- > >> > Ian > >> > > >> > On Tue, Jul 9, 2013 at 10:59 AM, Eric Iverson < > [email protected] > >> >wrote: > >> > > >> >> I :) use JHS all the time and think it has a bright and secure > future. > >> >> > >> >> I'm happy with the codemirror text editor currently available in JHS. > >> >> It is fast and has undo/redo, find/replace, ctrl key shortcuts for > >> >> save and run, syntax coloring, etc. > >> >> > >> >> Thanks for the ymacs reference. It looks interesting and would be a > >> >> nice addition to JHS. Hooking codemirror to the JHS framework was > less > >> >> that a days work. I hope someone interested in ymacs will try it out > >> >> in JHS. I'd be happy to answer questions and give pointers, but think > >> >> the codemirror example might be all that is needed. > >> >> > >> >> > >> >> On Tue, Jul 9, 2013 at 6:33 AM, Raul Miller <[email protected]> > >> wrote: > >> >> > jhs might use http://www.ymacs.org/ > >> >> > > >> >> > Does anyone still use jhs? > >> >> > > >> >> > -- > >> >> > 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 > >> ---------------------------------------------------------------------- > >> 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
