Dear Joe, Thank you for your thoughtful response. First, let me say that philosophy is actually one of my specialties, so getting philosophical would be just fine.
I fully agree with everything you said regarding JSoftware's challenge and trade-offs regarding trying to run a business. I spent years of my own time and money on the one package I mentioned alone. It solved several significant problems, especially at the time. After investing years of time (I quit my job and worked on it full time), and unloading what was left of my savings on advertising, I made almost nothing. After 30+ years in the business, I've learned that a great solution to a significant problem promoted and supported in a professional manner will get you nothing. Intangible things like marketing budget, industry connections, charisma, some of the right, unrelated, advocates, and a lot of luck will get you much further. (Of course you still have to try!) Somewhat fortunately, in my case, I found a single company that invested enough into my product to make the whole project minimally worth my time. I eventually released it as open-source. I did this first in pieces, and then the whole thing. Except for that single company, I have no idea if anyone is using it. In the past, failed proprietary software simply disappeared. (What ever happened to "Actor", "InterLisp", etc.?). IMO, it is a travesty for the world to have lost such elegant pieces of software - especially because many of them failed for the intangible reasons, and not because they weren't a significant, elegant, well designed, promoted, and supported solution. These days, I suppose software vendors end up deciding to release their source code rather then bury it. Sometimes someone can eek out a living with it, other times a group takes it on out of appreciation of its elegance, sometimes hoping it'll take off. What this all means for J, I do not know. Unlike many on the list, I am not a mathematician, and I don't do mathematical models. My interest in J stems from my nostalgic past with APL. Those were the days when I wrote a complete application in a week. A month long project was massive. Applications had no GUI in those days. I certainly wrote more application in 1980-1983 than the combined remainder of my life. These days I spend a year on the GUI and two weeks on the logic. Attempting to write GUI applications for the Web these days stretches the very limits of my abilities. But all this is perhaps a different issue... JSoftware is playing a very delicate balancing act. On one hand they need to release enough source to attract people like myself, in the hopes that some of us will connect them to paying customers. On the flip side, they need to retain enough control to assure that paying customer actually pay them. I know they are in it for a combination of love for the language, the need to survive, and the desire to prosper. J is very elegant and a great solution, but a very delicate balancing act indeed. I surely wish them luck! My interest in J will be highly related to how much source they release. Whether my interest in J will ultimately assist in their success, no one can tell - not even me. Regardless, I like J, appreciate what they have done, and wish them well. Thanks. Blake On Tue, Apr 1, 2014 at 5:02 AM, Joe Bogner <[email protected]> wrote: > Blake, > > Welcome to the list. My opinion was similar to yours when I started using J > about 9 months ago. I will post a short reply as the full reply is more > philosophical and probably belongs on the [email protected] forum ( > jsoftware.com/forums.htm) > > I think of J as having several components of the the package or language. > The core library, the IDEs, proprietary extensions, and the package > library. The open source question needs to be contemplated at each of the > components. > > As far as I know, the latest version j8 is using a j7 core. The j7 core is > what is provided in the gpl source. I don't know if it's the exact same > core or if there has been changes between the gpl source and what is being > distributed today. I would assume there are few, if any deviations between > the binary library that is distributed in current versions and the gpl > source version. > > Buliding the gpl source is not exactly trivial on all platforms. I've built > it on windows using both mingw and visual c++ and on linux using gcc. It's > not that difficult on windows, but it's not a single command. I intend to > publish my findings on using mingw at some point because it's useful to > play with J source on windows and not require a 2 gig install of Visual > C++. It's also nice to build with other free (as in speech) software. > > The IDEs have a limited source release, although the latest j8 qt version > is provided: http://www.jsoftware.com/download/jqt/. I have not tried to > compile it. > > There may or may not be proprietary extensions. > > The package library seems to be largely open source and is maintained on > the web svn. > > My personal opinion is that J Software is a business and deserves to find > ways of being compensated for their work. Proprietary extensions and even > custom IDEs are well within the bounds in my opinion. I think it's > incredible that the gpl source is available. There is a large leap in terms > of effort and skill from providing the source to running an open source > project. As a business, J Software needs to weigh where to make investments > and devoting resources to running J as an open source project/community may > not yield the best return. So, I agree with Raul that it's up to the > community. > > The community faces the same challenge in terms of time and skill. I think > many of us are demotivated because J is so mature and has so few bugs that > it may not have a significant impact to be a source maintainer. A language > this mature has tough choices to make on whether to include or exclude a > feature. Part of J's beauty comes from its simplicity and saying no. > > I remember early on that I was disappointed that J didn't have dictionaries > or hash tables. I figured that I might add it to my version of the code. > I'm glad that I didn't. I researched some options and was ready to > implement. I stopped as I read and got more familiar with J. I've since > learned that the J way has made some deliberate choices that often better > than my ad hoc, project based needs. Most solutions do not require a hash > table. I haven't actually needed it yet. > > There has been one main source contribution since I've been following the > list, the functional extensions. These extensions are probably 10 years > ahead of my skill level in J. I haven't seen any real mentions of bugs - > most of it's been interpretation. Are there warts? Sure, but every language > has them and they may not be bugs. > > Anyhow, I said this was going to be short and is clearly not. My objective > was to let you know that, like you, J's source availability was a big > factor in deciding to invest my life into it. Other than tinkering around > in the source to better learn, I haven't needed to extend it. I don't think > many others have either, otherwise we probably would have made more of an > effort to build a process for maintaining it. > > Regards, > Joe > > > On Tue, Apr 1, 2014 at 2:02 AM, Blake McBride <[email protected]> wrote: > > > Dear Raul, > > > > I understand what you are saying, but I don't think that was my question. > > > > While I am technically capable of maintaining J, there are many, many > > things I am also capable of maintaining (like the Linux kernel, GCC, my > > video drivers, etc.). Given the practical limitations of time, I was > > hoping to leverage off of the wonderful and deeply appreciated > up-to-date, > > ongoing work done by the JSoftware team while others could leverage off > of > > the work I have done (i.e. https://github.com/blakemcbride/Dynace ). > > Certainly, the JSofwrae team is under utterly no obligation, moral or > > otherwise, to share their work. I was simply asking the question. > > > > Given time limitations, although I am sure it would be great fun, I am > > highly unlikely to ever "maintain" J. On the flip side, it is highly > > likely that I would contribute build or porting fixes, and document and > > report bugs. Given the fact that full source is available to many fine > > languages, I am reluctant to devote significant time to a single source > > (from one vendor) proprietary language - just a personal sentiment. > > > > With respect and appreciation, > > > > Blake McBride > > > > P.S. In order to build J701b on current 64 bit Linux machines, I had to > > change bin/jconfig > > From: SOLINK=" -shared -W1,soname,libj.so -lm -ldl -o " > > To: SOLINK=" -shared -Wl,-soname=libj.so -lm -ldl -o " > > (Haven't checked yet but I sense that jconfig is generated and I need to > > propagate that change elsewhere. > > > > Also, after building on this 64 bit machine, the majority of the tests > pass > > but several failed. I won't bother to report this since the code is so > old > > anyway. > > > > > > > > > > > > On Tue, Apr 1, 2014 at 12:20 AM, Raul Miller <[email protected]> > > wrote: > > > > > The official release has not been changed since 2011. > > > > > > I think we (the J community) need to step up and take start addressing > > > the bug list: > > > http://www.jsoftware.com/jwiki/System/Requests%20%26%20Bug%20Reports > > > > > > That's a direct consequence of it being open source now. It's up to us. > > > > > > (Personally, I've got a machine that I intend to use for J source > > > debugging and development, but I can't get it on the internet right > > > now. I think I'll need to arrange for the aging and overloaded router > > > that's immediately upstream of me to be replaced before I can get much > > > of anywhere.) > > > > > > Thanks > > > > > > -- > > > Raul > > > > > > On Tue, Apr 1, 2014 at 1:05 AM, Blake McBride <[email protected]> > > wrote: > > > > Greetings, > > > > > > > > I was under the impression that J source code (to the engine) was > > > > available. I understand there is an old snapshot of 701b, but I have > > > been > > > > keeping up-to-date with your SVN repository thinking I was keeping up > > > with > > > > the source. Wanting to build it now, I realized it isn't there. > > > > > > > > Having source code is an important factor to me. Is it available > > > > somewhere, or are there plans for additional source code releases? > > > > > > > > Thanks. > > > > > > > > Blake McBride > > > > > ---------------------------------------------------------------------- > > > > 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
