I give a pretty complete example here: http://www.jsoftware.com/jwiki/Community/Conference2012/Talks/ParallelSimulationInJ.
On Wed, Jul 31, 2013 at 6:28 PM, Skip Cave <[email protected]> wrote: > Devon, > > Your techniques for parallelizing algorithms using multiple sessions would > make for an interesting wiki article. I would encourage you to share the > details of how you implement the system in such an article. > > Skip > > On Wed, Jul 31, 2013 at 11:34 AM, Devon McCormick <[email protected] > >wrote: > > > I continue to be amazed that the simple, consistent, generalized way APL > > incorporates arrays into the basic fabric of the language is so widely > > unknown. Of course, it took me maybe 15 minutes to understand this when > I > > was 13 years old, so maybe it's too advanced for much of the programming > > world </sarcasm>. > > > > BTW, I run my parallelized J code every single day - no threads required. > > > > > > On Wed, Jul 31, 2013 at 12:24 PM, Raul Miller <[email protected]> > > wrote: > > > > > Of course. > > > > > > I don't think anyone would say that the existence of languages like APL > > (or > > > any of its derivatives) means that all of the rest of achievement in > the > > > context of computers will go away, or should go away. > > > > > > Ok, maybe some would say such things, but I see no reason to express > > > agreement with them. > > > > > > Lessons to learn from though? > > > > > > That, I think is the point. > > > > > > Thanks, > > > > > > -- > > > Raul > > > > > > > > > On Wed, Jul 31, 2013 at 11:30 AM, CL Jason <[email protected]> wrote: > > > > > > > No, not-supporting-something is not a shame, nor a glory excuse. > > > > Concurrency/parallel and dbms have their own merits and are also > > > > interesting. > > > > > > > > > > > > On Wed, Jul 31, 2013 at 10:44 PM, Raul Miller <[email protected] > > > > > > wrote: > > > > > > > > > 0. I'd be interested in some additional perspective on why you > > believe > > > > the > > > > > contents are irrelevant. (thanks!) > > > > > > > > > > 1. kona, also - but since his general gist was about architectural > > > > issues, > > > > > rather than focussing on a particular implementation, I am inclined > > to > > > > > forgive him for the slightly confused presentation. > > > > > > > > > > 2. J is indeed single threaded, but you can run multiple J > processes. > > > > And, > > > > > in fact, jhs gives you a server implementation which (with a > > relatively > > > > > small amount of work - trivial compared to the amount of work > people > > > put > > > > > into serious programming efforts) can give you multiple J processes > > > under > > > > > the control of a single client. > > > > > > > > > > 3. Here, I think you are drawing a contrast between high volume > > > > transaction > > > > > processing (such as Amazon might need for its shopping cart > > > > implementation) > > > > > and analytics work (where someone tries to correlate information). > I > > > am > > > > > not sure that I'd use a K/Q rdbms implementation at Amazon - I > expect > > > the > > > > > hardware costs would be too high. Then again, I'm not working for > > > Amazon > > > > so > > > > > I'm not sure that I'll care a lot about this issue. [More > generally: > > a > > > > tool > > > > > being useful never means that other tools are not useful for other > > > > things.] > > > > > > > > > > And, as an aside, perl can be fun... (but I've not read that /. > page, > > > > yet) > > > > > > > > > > Anyways... I feel that the point you are trying to drive at is that > > no > > > > one > > > > > has been writing much about using J in multiprocess contexts, yet? > > > > > > > > > > If so, I'll just remind you that that's more a cultural > observation - > > > > about > > > > > what we have felt like doing and talking about - than anything > else. > > > > > > > > > > Thanks again, > > > > > > > > > > -- > > > > > Raul > > > > > > > > > > On Wed, Jul 31, 2013 at 10:22 AM, CL Jason <[email protected]> > > wrote: > > > > > > > > > > > 0. Yes, the pictures are taken of many APL/J masters, but the > > > contents > > > > > are > > > > > > much irrelavent. > > > > > > 1. arguments based on k/q: please check all the concrete examples > > > > against > > > > > > nosql apps, if those paragraphs of k/q were removed, there are > only > > > > empty > > > > > > assertions > > > > > > 2. up to now, J is single-threaded, k/q has some supports of > > > > > > multi-threading, which is one form of concurrency/parallel > > > programming > > > > > > 3. DBMS is more than just a query interface, and supporting the > > > subset > > > > of > > > > > > SQL doesn't mean the other parts of SQL are shit. > > > > > > > > > > > > there is even /. advertisement: > > > > > > > > > > > > > > > > > > > > > > > > > > > http://developers.slashdot.org/story/13/07/30/2348212/remember-the-computer-science-past-or-be-condemned-to-repeat-it > > > > > > where > > > > > > you'll even see ads for Perl... > > > > > > > > > ---------------------------------------------------------------------- > > > > > > 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 > > > > > > > > > > > -- > > Devon McCormick, CFA > > ^me^ at acm. > > org is my > > preferred e-mail > > ---------------------------------------------------------------------- > > For information about J forums see http://www.jsoftware.com/forums.htm > > > > > > -- > Skip Cave > Cave Consulting LLC > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > -- Devon McCormick, CFA ^me^ at acm. org is my preferred e-mail ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
