Re: [Emc-users] Machinekit and LinuxCNC
Hi Dave; Thanks for the response; I don't think that its bleak at all. Like you, I see continual improvement, and a great future. Open Source: I did "manage" the largest and longest running open source project in the Canadian Government. Million++ of downloads, distributed for a while by Apple, used by NASA and CERN, etc, etc. But when it came time to present work at a HTML5 yearly meeting, I presented a colleagues' work (from Fraunhofer, in Germany) because that project made more sense for integration with HTML5. So, I fully expect that we will continue to see improvement and progress, and adoption of the best of the ideas. Regards - JohnS. -- ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Machinekit and LinuxCNC
Well.. I don't think it is quite this bleak. LinuxCNC has really not had any truly defined "direction or plan" for the years I have been using it (5 or 6?). There have been groups of people who have pushed in different directions for various reasons, but I'm not aware of any truly "defined plan". We have had occasional voting (even recently) on some key events which I think is a good thing. But the act of voting doesn't make for a "plan" or "overall direction". Having a plan infers that someone/s will be able to execute the "plan". I equate open source developers with cats (no offense - I like cats). They all have their own interests, they tend to be fiercely independent, and they wander in different directions. As such; managing volunteer open source developers would be akin to "herding cats". Which explains, in my mind, why there is no master plan/direction. Machinekit has its own direction which is also controlled by volunteer open source developers. (More cats.) Need I say more? :-) As users, we are much better off than we were 4-5 years ago. There has been a lot of progress; running on Non PC hardware (BBB), the new Planner, different user interfaces, the ability to run on different Linux Distributions, etc.These are huge improvements. I have Machinekit 2.7 running on one PC and I will be updating another PC soon to LinuxCNC 2.7..I just bought a Mesa Ethernet board since I want to try out the Ethernet driver which is supported in LinuxCNC 2.7 and I want to run it on a Fanless Intel Bay Trail quad core 2.0 ghz motherboard that costs less than $100 and draws very little power (less than 12 watts).These are things that were being dreamed about just a few years ago. Sure.. things could always be better, but still, these are exciting times. :-) Dave On 11/2/2014 12:01 PM, David Armstrong wrote: > or put simple a big mess rather than a concerted effort on one > my thoughts are that machinekit has gone the 3d printing route and forgot > about real machines doing real work > but with an attempt to bring things up to date like nml etc . > and linuxcnc folks have lost the plot in the real world , so we now have 2 > options and a mess in the middle > > with the real work of gui's being lost in the ether at the moment , and a > loss of direction , so yes John i feel > as if everything is dangling by a thread , but the direction or future is > unknown ... > > i'd rather leave everything , and concentrate on fixing whats broke , or > what should be their that is not > for me either way, theres holes for a good working system. at least for > real machines with real uses that work for their living > > but no doubt the mess will soon unravel and the mud will clear > > > > > > > > > On 2 November 2014 16:29, Jon Elson wrote: > >> On 11/02/2014 09:18 AM, John Alexander Stewart wrote: >>> Can someone tell me what the eventual plans for Machinekit and LinuxCNC >> is? >>> Are the two forks going to come together again? >>> >>> I really like LinuxCNC, and I really like the idea of small, embedded CNC >>> controllers, so I'm just wondering. (I have been loosely following the >>> machine kit google group; maybe the answer is obvious, but I've missed >> it) >>> >> There are not complete forks, never to share code again. A >> large part of the reason >> for the fork, as I understand it, was to break free from the >> constraints of NML and >> its dependence on shared memory, with the only alternative >> of massive overhead >> when sending the entire memory block across the net every >> servo period. >> >> The machinekit folks have their own system for doing this, >> enabling them to break >> parts of LinuxCNC to run on different hosts over the net. I >> think there is some >> disagreement on how this was done, and so the LinuxCNC group >> is doing this >> a different way, that they think will be more flexible. I >> REALLY do not know all >> the details of this. But, the machinekit folks have their >> scheme working, apparently, >> and the LinuxCNC way is not ready yet, I think. >> >> Also, the machinekit runs with a Xenomai kernel on the >> Beagle Bone, and with the >> assist from Charles Steinkuehler's PRU code for step >> generation, etc. it works >> quite well. You can't get much smaller than that! I have >> not run a machine with >> the beagle/machinekit, but build CRAMPS boards (Also >> Charles' design) and it >> all seems to work. I just ssh -X into the Beagle, so I have >> not explored the >> concept of running the GUI on a different CPU. If I was >> running a real machine, >> the GUI latency might be bothersome. >> >> Jon >> >> >> -- >> ___ >> Emc-users mailing list >> Emc-users@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/emc-users >> > ---
Re: [Emc-users] Machinekit and LinuxCNC
Hmmm. Reminds me of a company I worked for; two groups who could not talk. One group (mine) was spending a pot of money (about a million bucks, 1980s money) to keep an old computing system going, the other wanted to spend that same pot of money on the next generation of computers. I asked for, and got a desk in Group II's building; right in the midst of Group II. First question I got asked by a very nervous Group II was "what the h*ll are you doing here?" My answer was "Group 1 is trying to keep the engineers delivering code to keep the lights on, and you are trying to progress the tools so that things work better in the future". Put that way, Group II realized that, "hey, that makes sense - we can't stop production for a year", and my old group knew that the old way was increasingly not really cutting it, but we had to keep the lights on... Within ONE DAY the conflict was over, and we all worked towards a common goal. Maybe not relevant, but there are a lot of really smart people here, working away to make better environments, and each and every one has something to contribute to the success of the code base. Off to cook dinner - JohnS. -- ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Machinekit and LinuxCNC
or put simple a big mess rather than a concerted effort on one my thoughts are that machinekit has gone the 3d printing route and forgot about real machines doing real work but with an attempt to bring things up to date like nml etc . and linuxcnc folks have lost the plot in the real world , so we now have 2 options and a mess in the middle with the real work of gui's being lost in the ether at the moment , and a loss of direction , so yes John i feel as if everything is dangling by a thread , but the direction or future is unknown ... i'd rather leave everything , and concentrate on fixing whats broke , or what should be their that is not for me either way, theres holes for a good working system. at least for real machines with real uses that work for their living but no doubt the mess will soon unravel and the mud will clear On 2 November 2014 16:29, Jon Elson wrote: > On 11/02/2014 09:18 AM, John Alexander Stewart wrote: > > Can someone tell me what the eventual plans for Machinekit and LinuxCNC > is? > > > > Are the two forks going to come together again? > > > > I really like LinuxCNC, and I really like the idea of small, embedded CNC > > controllers, so I'm just wondering. (I have been loosely following the > > machine kit google group; maybe the answer is obvious, but I've missed > it) > > > > > There are not complete forks, never to share code again. A > large part of the reason > for the fork, as I understand it, was to break free from the > constraints of NML and > its dependence on shared memory, with the only alternative > of massive overhead > when sending the entire memory block across the net every > servo period. > > The machinekit folks have their own system for doing this, > enabling them to break > parts of LinuxCNC to run on different hosts over the net. I > think there is some > disagreement on how this was done, and so the LinuxCNC group > is doing this > a different way, that they think will be more flexible. I > REALLY do not know all > the details of this. But, the machinekit folks have their > scheme working, apparently, > and the LinuxCNC way is not ready yet, I think. > > Also, the machinekit runs with a Xenomai kernel on the > Beagle Bone, and with the > assist from Charles Steinkuehler's PRU code for step > generation, etc. it works > quite well. You can't get much smaller than that! I have > not run a machine with > the beagle/machinekit, but build CRAMPS boards (Also > Charles' design) and it > all seems to work. I just ssh -X into the Beagle, so I have > not explored the > concept of running the GUI on a different CPU. If I was > running a real machine, > the GUI latency might be bothersome. > > Jon > > > -- > ___ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users > -- ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Machinekit and LinuxCNC
On 11/02/2014 09:18 AM, John Alexander Stewart wrote: > Can someone tell me what the eventual plans for Machinekit and LinuxCNC is? > > Are the two forks going to come together again? > > I really like LinuxCNC, and I really like the idea of small, embedded CNC > controllers, so I'm just wondering. (I have been loosely following the > machine kit google group; maybe the answer is obvious, but I've missed it) > > There are not complete forks, never to share code again. A large part of the reason for the fork, as I understand it, was to break free from the constraints of NML and its dependence on shared memory, with the only alternative of massive overhead when sending the entire memory block across the net every servo period. The machinekit folks have their own system for doing this, enabling them to break parts of LinuxCNC to run on different hosts over the net. I think there is some disagreement on how this was done, and so the LinuxCNC group is doing this a different way, that they think will be more flexible. I REALLY do not know all the details of this. But, the machinekit folks have their scheme working, apparently, and the LinuxCNC way is not ready yet, I think. Also, the machinekit runs with a Xenomai kernel on the Beagle Bone, and with the assist from Charles Steinkuehler's PRU code for step generation, etc. it works quite well. You can't get much smaller than that! I have not run a machine with the beagle/machinekit, but build CRAMPS boards (Also Charles' design) and it all seems to work. I just ssh -X into the Beagle, so I have not explored the concept of running the GUI on a different CPU. If I was running a real machine, the GUI latency might be bothersome. Jon -- ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
[Emc-users] Machinekit and LinuxCNC
Can someone tell me what the eventual plans for Machinekit and LinuxCNC is? Are the two forks going to come together again? I really like LinuxCNC, and I really like the idea of small, embedded CNC controllers, so I'm just wondering. (I have been loosely following the machine kit google group; maybe the answer is obvious, but I've missed it) Thanks - JohnS. -- ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
