#linuxcnc-devel | Logs for 2015-10-16

[00:22:39] <KGB-linuxcnc> 03Sebastian Kuzminsky 05seb/2.6/motion-logger 6b357d6 06linuxcnc 10src/emc/motion-logger/motion-logger.c motion-logger: remove Motion commands that are not available in 2.6 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=6b357d6
[09:34:30] <skunkworks> http://www.cnczone.com/forums/tormach-personal-cnc-mill/260722-cnc-tormach-forum-15.html#post1775410
[09:34:44] <skunkworks> light bulbs are going off...
[09:36:00] <cradek> I wonder how long they can get away with using ubuntu10
[09:36:13] <cradek> ... it's not even getting basic security fixes anymore
[09:36:40] <cradek> I guess until it gets hard to find hardware that will boot it
[09:41:37] <skunkworks> sure
[09:42:15] <skunkworks> it is a little funny that people are all for pathpilot - when you could have gotten 90% of that from linuxcnc
[09:46:32] <jepler> you can't lead a horse to water, but marketing might help
[09:53:21] <malcom2073> It's the last 10% that makes all the difference in the world. Ask apple
[09:58:27] <pcw_home> "we are building a religion"
[10:01:11] <malcom2073> Community pcw... it's called a community ;)
[12:10:25] <cradek> andypugh: I wish the guy would just say what he wants
[12:14:14] <ssi> "This didn't take long. You can't be honest and pay Tormach for their hard work? This sure sounds like trying to pirate their software. I'll bet cnczone won't let allow this Hacking on their site.
[12:14:18] <ssi> "Hacking"
[12:28:33] <cradek> ssi: arggghblarggh I don't understand things but I am mad
[12:28:56] <ssi> sounds about right
[12:29:19] <ssi> he's not personally being wronged in any way, real or perceived, but he thinks someone else is based on something he doesn't understand
[12:29:23] <ssi> so lets namecall
[12:31:03] <cradek> (I don't know the context, but I assume someone is doing something the GPL allows)
[12:31:31] <ssi> in that thread skunkworks linked above about people trying to get a copy of pathpilot
[12:31:46] <cradek> pathpilot is freely distributable
[12:31:54] <ssi> yeah, via HACKING!
[12:31:56] <ssi> :P
[12:32:34] <cradek> heh, that's silly
[12:32:39] <ssi> indeed
[12:33:05] <ssi> I guess, some folks on there are saying they paid tormach $10 for a copy of the "upgrade", and tormach canceled their orders because they weren't previous machine customers
[12:33:17] <ssi> so that's what made some of those guys decide that distributing pathpilot must be stealing
[12:33:20] <cradek> tormach can do that, just fine
[12:33:49] <ssi> can they though? I thought they had to make their software available for no more than the cost of distribution
[12:33:52] <cradek> yeah, they're simply wrong, which is easy to be :-)
[12:34:08] <ssi> yeah, it's even easier to be wrong but mad about it :)
[12:34:34] <cradek> no, they do not have to provide it to anyone who asks. they DO have to provide the source to people they provide the binaries to, and those people have all the rights the GPL gives them.
[12:34:45] <ssi> ahh ok
[12:34:51] <jepler> It is not a GPL violation to decline to sell a binary of GPL software.
[12:34:59] <cradek> right
[12:35:23] <cradek> fwiw, the gpl2 faq is really easy reading, as is the gpl2 itself
[12:35:35] <ssi> so they are obligated to provide the source to people who have machines with pathpilot, and they're not allowed to stop any of those people from distributing said source
[12:35:56] <jepler> 3. You may copy and distribute the Program (or a work based on it,
[12:35:56] <jepler> under Section 2) in object code or executable form under the terms of
[12:35:56] <jepler> Sections 1 and 2 above provided that you also do one of the following:
[12:36:03] <jepler> b) Accompany it with a written offer, valid for at least three
[12:36:03] <jepler> years, to give any third party, for a charge no more than your
[12:36:03] <jepler> cost of physically performing source distribution, a complete
[12:36:03] <jepler> machine-readable copy of the corresponding source code, to be
[12:36:03] <jepler> distributed under the terms of Sections 1 and 2 above on a medium
[12:36:06] <jepler> customarily used for software interchange; or,
[12:36:23] <jepler> I assume that Tormach is doing 3b when they distribute GPL software in binary form, but I have not verfied the fact
[12:36:32] <jepler> note the specification of "any third party" there
[12:37:39] <cradek> does that mean a binary-haver could request that tormach sends the source to me, a binary non-haver?
[12:37:57] <ssi> I also like how so many of the people on cnczone looked at tormach's promotional info about pathpilot and assumed that all the features they call out are ones that tormach added to linuxcnc
[12:38:40] <cradek> yeah tormach has contributed a lot to the project, but it's probably not 1% of the work we've all done over so many years
[12:38:48] <jepler> cradek: you mean, is "any third party" referring to who may *receive* it or who may *request* it?
[12:38:53] <ssi> "
[12:38:54] <ssi> During the webinar, however, Tormach mentioned they optimized the kernel for motion control. In other words, you can do whatever you want and motion control will always take precedence over everything else. This is definitely one of the most awesome features PP has to offer. In fact, this IS the feature that I am most interested in!"
[12:39:00] <cradek> jepler: yes, I can't quite parse it for sure
[12:39:34] <ssi> jepler: to me it reads that any third party who requests it must receive it
[12:39:46] <jepler> yes, I have understood it more or less like ssi says
[12:39:58] <pcw_home> I do think they should dump Ubuntu 10.04 and switch to something modern on Preempt-RT
[12:40:08] <jepler> but I had never considered that for some reason I might want to cause Tormach to send an 8-track tape with a tar archive on it to cradek's house, possibly as some sort of joke
[12:40:27] <jepler> er, however many tracks those tapes had :-P
[12:40:31] <cradek> heh they're called 9-track
[12:40:32] <ssi> 9 track I think
[12:40:39] <cradek> but 8-track would be funnier
[12:40:44] <ssi> I don't think the gpl allows you to force them to distribute it on a given medium :)
[12:40:46] <pcw_home> or 7
[12:41:01] <ssi> if I were tormach and wanted to troll, I might distribute it on paper tape
[12:41:10] <ssi> that'd be a lot of damn paper tape
[12:41:40] <jepler> unfortunately units doesn't convert bytes to papertapefeet or punchardfeet
[12:41:58] <cradek> the written offer "accompanies" the binary, so I figured it would only give rights to the person "possessing" the offer
[12:42:07] <pcw_home> 10 BPI?
[12:42:08] <ssi> well if I had to guess I'd say that 8 bit per row tape is probably about 3/16" per byte
[12:42:23] <cradek> yeah it's 10 punches/inch
[12:42:45] <jepler> > Tape for punching was 0.00394 inches (0.1 mm) thick. The two most common widths were 11/16 inch (17.46 mm) for five bit codes, and 1 inch (25.4 mm) for tapes with six or more bits. Hole spacing was 0.1 inch (2.54 mm) in both directions
[12:42:46] <cradek> 7 bits wide, I think
[12:43:24] * pcw_home admits to loading and storing programs on PT
[12:43:38] <ssi> I really wanted to get a tape punch and write some programs for the HNC before I converted it
[12:43:44] <ssi> but punches weren't nearly cheap enough for those particular lulz
[12:43:59] <ssi> the diacro turret punch press I have still has a paper tape control on it
[12:44:07] <cradek> ssi: I punched and tried tapes on mine, but couldn't get it to read the tape
[12:44:28] <ssi> cradek: the guy I bought mine from had four of them running in production less than a year before I bought it
[12:44:46] <ssi> but he didn't give me any programs with it
[12:45:33] <jepler> wikipedia gives the thickness of IBM punchards as 143 per inch. but I'm not sure if I should use 80byte or 12*80bits as the capacity.
[12:45:42] <cradek> https://www.gnu.org/licenses/old-licenses/gpl-2.0-faq.html#WhatDoesWrittenOfferValid
[12:46:11] <cradek> it's pretty clear from this that your interpretation is right
[12:46:24] <jepler> using 143*80byte/inch, a gigabyte is about 1.4 punchcard miles
[12:46:29] <cradek> so I was wrong - anyone can request the source from tormach and expect to get it for cost of distribution.
[12:46:40] <cradek> (but ... not the binaries?)
[12:46:47] <ssi> jepler: that's a stack of cards 1.4 miles tall, yes?
[12:46:50] <jepler> ssi: right
[12:46:53] <ssi> nice
[12:47:14] <pcw_home> I still have the sounds of my 2702? ASCII Flexowriter punching a binary dump burned in my brain
[12:47:20] <ssi> Section 2 says that modified versions you distribute must be licensed to all third parties under the GPL. “All third parties” means absolutely everyone—but this does not require you to do anything physically for them. It only means they have a license from you, under the GPL, for your version.
[12:47:34] <jepler> defining punchtape at 7bit/.1in , a megabyte is about 1.8 punchtape miles
[12:47:45] <ssi> so regardless of whether or not you can get it from tormach, the isos that are floating around of pathpilot are 100% legal to distribute and use, and anyone with it has license from tormach to use it under the gpl
[12:47:57] <cradek> ssi: yes, absolutely
[12:48:15] <ssi> that's how I've always understood it
[12:48:46] <pcw_home> Does that apply to their GUI or can that have a separate license?
[12:49:19] <ssi> I guess it kinda depends how their gui is built
[12:49:24] <cradek> it applies to the whole combined work
[12:49:27] <jepler> I don't know the contents of the tormach iso. It could contain some elements that may not be freely copied because they are not GPL licensed but their arrangement on the CD is "mere aggregation". For instance, something as simple as Tormach original copyrighted artwork could mean that the ISO as a whole cannot be freely copied
[12:49:48] <ssi> jepler: true
[12:50:52] <cradek> sounds like if you want to use or redistribute the GPL parts, it would be better to make the source request and start there
[12:51:05] <ssi> getting deeper into this thread, it looks like tormach started providing it to anyone that asks
[12:51:09] <ssi> initially they weren't
[12:51:19] <ssi> now it's just fun watching all these machtards trying to build pathpilot
[12:52:21] <jepler> I wouldn't personally call them that...
[12:52:34] <cradek> yeah, gross
[16:18:48] <andypugh> Has anyone pointed to the chap on our forum running PathPilot with the Parallel Port? That might make them froth :-)
[16:23:34] <jepler> froth? is that the language where entering + 2 3 prints 5?
[16:24:39] <andypugh> I used a game development system once that was based on FORTH. I really liked it.
[16:26:10] <jepler> The idea of forth is really keen. But I've never written anything of consequence in it.
[16:26:25] <jepler> pcw should give us a hostmot2 with a forth interpreter built in
[16:29:04] <andypugh> I am starting to think that the fact that Tormach has switched to LinuxCNC is making a number of other Mach3 folks consider moving. Including Hoss.
[16:33:26] <skunksleep> andypugh: link to boss?
[16:33:38] <andypugh> It was in that thread.
[16:33:42] <skunksleep> *hoss
[16:34:15] <skunksleep> Really?.. Maybe I should read the whole thing
[16:34:57] <andypugh> Post 20
[16:35:39] <skunksleep> Hoss has played around with mach4... I didn't see much on that in his threads
[16:36:20] <PCW> Its probably not that hard to make a small 32 bit 100 MIPs or so Forth oriented CPU in a Spartan6
[17:00:17] <mozmck> logger[psha]
[17:00:17] <logger[psha]> mozmck: Log stored at http://psha.org.ru/irc/%23linuxcnc-devel/2015-10-16.html
[17:13:43] <jepler> http://www.excamera.com/sphinx/fpga-j1.html has always looked interesting to me
[17:14:40] <jepler> > J1 is a small (200 lines of Verilog) stack-based CPU, intended for FPGAs. A complete J1 with 16Kbytes of RAM fits easily on a small Xilinx FPGA.
[17:15:26] <mozmck> that does look neat