#emc-devel | Logs for 2008-02-28

[01:20:41] <cradek> jmkasunich: does easycad do polylines?
[02:13:43] <jmkasunich> it does polygons
[02:13:55] <jmkasunich> I'll boot it up and see if does polylines
[02:16:23] <cradek> I'm finally writing dxf-realize
[02:16:52] <jmkasunich> yay
[02:17:43] <jmkasunich> easycad has "paths"
[02:17:57] <cradek> is that a polyline made of lines and arcs?
[02:17:58] <jmkasunich> but they seem to consist only of lines, not lines and arcs
[02:18:05] <cradek> oh, hmm
[02:18:46] <cradek> I fear dxf-realize is gonna have to do the sucktastic guessing of what things are supposed to be a path
[02:19:16] <jmkasunich> searching for polyline in the help system finds spline only, the text is "CUBIC B-SPLINE (smooth polyline in AutoCADĀ® terms) draws an open smooth curve"
[02:19:29] <cradek> darn
[02:19:54] <jmkasunich> the sucktastic guessing takes it from being something simple to something not
[02:20:02] <cradek> well if this is going to work with (the free) qcad, I have to bite that bullet anyway
[02:20:11] <cradek> yes I agree
[02:20:15] <jmkasunich> qcad doesn't have polylines either?
[02:20:20] <cradek> not the free version
[02:20:42] <cradek> and it doesn't have 3d (a polyline can't be at z=-1)
[02:20:51] <cradek> so that scheme is out the window too
[02:21:03] <jmkasunich> I thought realize used layers for Z
[02:21:15] <cradek> no, that's also sucktastic
[02:21:25] <cradek> layers for *tools*
[02:21:30] <jmkasunich> oh
[02:22:33] <jmkasunich> easycad doesn't do 3d either - fastcad does (thats the deluxe version)
[02:23:01] <cradek> I'm now remembering why I never started this
[02:23:10] <cradek> but, I remembered too late, as it's well-started
[02:23:14] <jmkasunich> oh
[02:23:26] <jmkasunich> I was about to say don't feel like you have to start it on my account
[02:23:31] <cradek> :-)
[02:24:01] <jmkasunich> it seems like it will be extremely hard to make it work right
[02:24:28] <jmkasunich> very easy to have a badly behaved drawing, and no real way to check the result short of loading it into axis
[02:24:54] <cradek> the whole way it works is polylines are tool paths, and depths are depths, and layers are tools
[02:25:18] <jmkasunich> without polylines and depths, you are in a bit of a pickle
[02:25:19] <cradek> nearly all of that has to be thrown out to work with these lesser cad programs
[02:26:13] <cradek> no offense, that's not how I meant to sound
[02:26:19] <jmkasunich> seems like you aren't really talking about rewriting realize
[02:26:28] <jmkasunich> we're talking about a different kind of cam program
[02:26:33] <cradek> yeah
[02:27:04] <cradek> I was thinking minor hacks might make it work, but it would have to be pretty much reamed
[02:27:12] <jmkasunich> yep
[02:27:21] <jmkasunich> I'm thinking about a different structure completely
[02:27:22] <cradek> I could have it reading autocad's dxfs in an hour, but fat lot of good that does anyone
[02:27:47] <jmkasunich> list of lines and arcs, mission is simply to make paths that include all of them
[02:28:28] <cradek> you mean unordered list right?
[02:28:31] <cradek> mishmash
[02:28:38] <jmkasunich> pick one line or arc arbitrarily, find another with its endpoint within some epsilon of the first one, continue till you run out of things within epsilon
[02:28:39] <jmkasunich> yep
[02:28:50] <cradek> yes that's what it would have to do
[02:28:56] <cradek> probably not even that hard
[02:29:02] <cradek> but where do you get depths?
[02:29:26] <jmkasunich> well, for what I've been doing to date I don't need depths
[02:29:32] <jmkasunich> but thats not a very general solution
[02:29:55] <jmkasunich> if its 2d cad, you will have 2d paths
[02:30:31] <jmkasunich> its the users job to set the depths - the program just removes the drugery of connecting lines and arcs, and calculating I and J values, etc
[02:31:15] <jmkasunich> the epsilon algorithm would get confused if you had a t-junction anywhere
[02:31:28] <jmkasunich> which makes sense, cause you can't machine that
[02:31:50] <cradek> why not?
[02:32:30] <jmkasunich> which one is the toolpath, and which one is the mistake?
[02:32:52] <cradek> neither - they're both toolpaths
[02:33:03] <cradek> one is two segments, one is one segment
[02:33:29] <jmkasunich> well, I don't see that making a meaningfull part
[02:33:38] <jmkasunich> one or the other path will be cutting stuff thats already been cut
[02:33:50] <cradek> I cut a T-shaped path in max
[02:34:03] <cradek> it's true some of it was a recut
[02:34:10] <cradek> but if you want a T, you have to do that
[02:34:34] <jmkasunich> you can't make a T surface in 2D
[02:34:42] <jmkasunich> I'm not talking about a T-shaped slot
[02:34:53] <cradek> what are you talking about surface?
[02:34:55] <jmkasunich> I'm talking about a T shaped surface
[02:35:07] <cradek> these are tool paths, not surfaces
[02:35:27] <jmkasunich> all the realize things you did for me were paths that I used with radius comp, they were the actual part shapes
[02:35:53] <jmkasunich> hi ray
[02:35:57] <cradek> hello
[02:36:09] <rayh> Hi Guys
[02:36:15] <rayh> I got my microwave
[02:36:19] <jmkasunich> cool
[02:36:26] <cradek> telnetww is fast?
[02:36:34] <cradek> slick
[02:36:35] <rayh> And just got a Ubuntu server setup to route to my local network.
[02:36:40] <jmkasunich> got the bits marching in rows and columns now, instead of crawling up the anchor line?
[02:36:55] <rayh> Something like that.
[02:39:53] <rayh> Fast-air has a speedometer. Something around 2m is average though I've seen it near 3.
[02:40:20] <jmkasunich> nice
[02:40:49] <jmkasunich> hey SWPLinux
[02:40:54] <SWPLinux> hi jmkasunich
[02:41:01] <rayh> Any thoughts on the timeline for 2.3
[02:41:08] <jmkasunich> have you pulled the photos off your camera yet?
[02:41:18] <SWPLinux> sometime after 8.04 is released, I think
[02:41:19] <jmkasunich> I want to see the one with max and the big rougher
[02:41:22] <SWPLinux> jmkasunich: no, but I can
[02:41:49] <SWPLinux> hmmm. I left the camera in the room
[02:41:58] <SWPLinux> no wifi there, unfortunately
[02:42:08] <jmkasunich> oh, you're on the road again?
[02:42:13] <SWPLinux> yep
[02:42:15] <jmkasunich> is this the bollywood trip?
[02:42:20] <SWPLinux> Sunny Mishicot, Wi
[02:42:22] <SWPLinux> nope
[03:54:56] <SWPLinux> jmkasunich: were you involved with EMC when support for RTAI was first added?
[03:55:14] <SWPLinux> (it llooks like RTLinux was first, then RTAI a year or two later)
[03:55:18] <jmkasunich> no
[03:55:21] <SWPLinux> ok
[03:55:44] <jmkasunich> when I started, there were lots and lots of compile time #ifdef, etc stuff to switch between support for different OSs
[03:55:47] <SWPLinux> I've procrastinated a lot on getting that email out to the EMC movers and shakers
[03:55:56] <jmkasunich> I got involved with RTAPI to get rid of that
[03:55:59] <SWPLinux> yep, I'm seeing a lot of that
[03:56:11] <SWPLinux> ok, RTAPI was around 2002 or so maybe?
[03:56:17] <jmkasunich> there was even stuff in there for windows and windows CE
[03:56:28] <jmkasunich> 2003 I think, but I might be getting mixed up
[03:56:36] <SWPLinux> yep, and whatever the DEC Alpha UNIX was at the time :)
[03:56:59] <jmkasunich> my coding started after a meeting held a the NAMES show - not an emcfest or anything, just a meeting in a room
[03:57:12] <jmkasunich> but I'm no longer sure which year it was
[03:57:27] <SWPLinux> hmmm. was that the one where the board was first discussed?
[03:57:42] <jmkasunich> no, that was at the emcfest in Ann Arbor (smithy)
[03:57:46] <jmkasunich> at least one, maybe 2 years later
[03:57:48] <SWPLinux> ok
[03:58:51] <jmkasunich> ohh, history: http://www.metalworking.net/EMC_news_history/Meeting2003.html
[03:59:08] <SWPLinux> kewl
[03:59:30] <SWPLinux> I've got half a dozen of those open :)
[03:59:38] <SWPLinux> some from the defunct NIST site
[04:00:28] <SWPLinux> it looks like RTAI was released some time after RTLinux, but I think before FSM Labs took it
[04:00:34] <SWPLinux> (it being RLTinux)
[04:00:40] <SWPLinux> RTLinux
[04:01:24] <SWPLinux> I didn't realize that RTAI was a university project at the beginning (just like RTLinux)
[04:10:16] <jmkasunich> judging from those meeting minutes, RTAPI started shortly after the 2003 NAMES show
[04:10:34] <jmkasunich> that was the first coding I ever did for EMC
[04:11:48] <jmkasunich> my first introduction to emc was at the 2002 names show, I met Jon Elson, Ray, and Paul there - over the next few months I started reading the mailing lists etc
[04:12:11] <SWPLinux> cool. amazing how much happens in a few years, isn't it
[04:15:27] <jmkasunich> yep
[04:17:14] <jmkasunich> http://emc.sourceforge.net/NIST-archive/msg04097.html
[04:17:26] <jmkasunich> fred, reporting that he set up an emc mailing list
[04:17:29] <jmkasunich> 1999
[04:17:37] <jmkasunich> it had 4 subscribers
[04:17:44] <SWPLinux> heh
[04:18:09] <SWPLinux> hmmm. I Think I have an archive of that list traffic, but I'm not sure how far back it goes
[04:19:30] <jmkasunich> I think this is all of it: http://emc.sourceforge.net/NIST-archive/threads.html
[04:19:51] <SWPLinux> hey, thanks
[04:19:51] <jmkasunich> the emc-users list was created shortly after that 2003 meeting, IIRC, and thats when we phased out the list at nist
[04:20:24] <SWPLinux> strange - I think I got involved in 2003 or 2004
[04:20:59] <SWPLinux> 2004, must have been
[04:21:08] <SWPLinux> the first fest I went to was 2005, at NIST
[04:25:41] <jmkasunich> http://emc.sourceforge.net/NIST-archive/msg06216.html HAL rev 0.01 ;-)
[04:26:28] <SWPLinux> heh. I'm just reading through a long message from Ray on how to go about developing HAL (new CVS module, etc)
[04:27:06] <SWPLinux> heh - funny that you mention the G2002 in that post
[04:31:08] <jmkasunich> damn, somehow it went from 10:00 to 11:30 in five minutes
[04:31:18] <SWPLinux> oops. sorry
[04:31:33] <SWPLinux> the dog must be ready to explode by now
[04:31:34] <jmkasunich> wasn't your fault
[04:31:55] <jmkasunich> oh, he's not driven by bodily functions - took care of those a bit ago
[04:32:05] <SWPLinux> phew! :)
[04:32:19] <jmkasunich> he's driven by the urge to sniff every tree, bush, leaf, and snowflake in the neighborhood
[04:32:21] <jmkasunich> most of them twice
[04:32:35] <SWPLinux> must get a cold nose this time of year
[04:34:44] <jmkasunich> always
[04:34:51] <jmkasunich> he seems immune to cold tho
[04:35:03] <jmkasunich> earlier I had him out back, where the snow is chest deep
[04:35:17] <jmkasunich> he'd run two steps, then plunge his entire head into the snow to sniff something
[04:35:25] <SWPLinux> heh
[04:35:29] <SWPLinux> sensitive schnozz
[04:35:34] <jmkasunich> a few seconds later he'd snort, come up for air, run two steps, and plunge the head in again
[04:35:39] <jmkasunich> up to his hears in the snow
[04:35:42] <jmkasunich> ears
[04:36:53] <jmkasunich> ok, this time I'm really off
[04:36:55] <jmkasunich> goodnight
[04:36:59] <SWPLinux> see ya
[09:12:39] <alex_joni> cradek: http://www.christian-kohloeffel.homepage.t-online.de/dxf2gocde.html
[09:12:50] <alex_joni> if you need translation services, ping me ;)
[10:00:39] <fenn> hey neat it actually works too
[10:16:25] <fenn> emc is a little too picky about its arc endpoints :\
[13:06:59] <alex_jon1> alex_jon1 is now known as alex_joni
[14:28:44] <alex_joni> cradek: seen the dxf2gcode ?
[14:29:03] <cradek> I haven't run it, but I looked briefly at the code
[14:29:42] <cradek> if it's good, we should try to get the guy to help us set it up for distribution with emc
[14:31:04] <alex_joni> afaik it already works as an AXIS filter
[14:31:14] <cradek> brb
[14:36:44] <alex_joni> we should look at packaging it for the LiveCD
[16:13:14] <alex_joni> hi ray
[16:13:24] <alex_joni> how's the microwave?
[16:28:20] <rayh> hi alex
[16:28:28] <rayh> Great
[16:28:46] <rayh> I'm getting the network worked around to what I want