#emc | Logs for 2005-03-28

[08:41:28] <alex_joni> encoder on the first -> command to the second one
[08:42:00] <Imperator_> that is no solution for me
[08:42:15] <alex_joni> bummer
[08:42:18] <alex_joni> why not?
[08:42:43] <Imperator_> Don't know the english word but in german i call thet " Pfusch"
[08:42:47] <Imperator_> that
[08:43:05] <alex_joni> lol
[08:43:11] <alex_joni> even Pfutsch
[08:43:25] <Imperator_> jo
[08:43:35] <alex_joni> und dann.. Futsch
[08:44:09] <Imperator_> real gantry support is not a big thing but i need to comunicate with the user
[08:44:20] <alex_joni> right
[08:44:27] <alex_joni> then it should get into task
[08:44:43] <A-L-P-H-A> morning folks.
[08:44:49] <Imperator_> morning
[08:45:20] <Imperator_> and NML is to complicatet to add some messages
[08:45:32] <Imperator_> it is too complicated at all
[08:45:43] <Imperator_> we need a simple message layer
[08:46:14] <Imperator_> EMC developers are no software guys like you ALex
[08:46:26] <Imperator_> even John can't change it
[08:46:33] <Imperator_> maybe Paul
[08:49:22] <alex_joni> I'm no software guy ;)
[08:49:29] <alex_joni> at least that's what I like to think
[08:49:43] <Imperator_> :-)
[08:53:01] <Imperator_> * Imperator_ back again
[08:53:24] <alex_joni> wb
[09:38:55] <alex_joni> hello paul_c
[09:40:40] <paul_c> Morning Alex
[09:40:48] <alex_joni> what's up?
[09:41:11] <alex_joni> do you have some tcl knowledge?
[09:41:32] <alex_joni> I'll bug rayh when he's around.. but till then I thought I'd ask
[09:43:17] <paul_c> what's the problem ?
[09:43:45] <alex_joni> I think the construct:
[09:44:15] <alex_joni> set types {
[09:44:15] <alex_joni> {[msgcat::mc "All files"] *}
[09:44:15] <alex_joni> {[msgcat::mc "Text files"] {.txt}}
[09:44:15] <alex_joni> {[msgcat::mc "NC files"] {.nc .ngc}}
[09:44:15] <alex_joni> }
[09:44:20] <alex_joni> is causing problems
[09:44:32] <alex_joni> last night I checked without locale set
[09:44:54] <alex_joni> and tk complained about having "]" instead of " " in the list construct
[09:46:59] <alex_joni> the second thing I want to talk about is NML, CL, IO
[09:47:03] <alex_joni> but that can wait..
[09:47:03] <paul_c> did it give you a line number ?
[09:47:19] <alex_joni> it's in proc fileDialog {} {
[09:47:36] <alex_joni> if you have a running emc2
[09:47:43] <alex_joni> try pushing Open.. and see if it works
[09:53:01] <paul_c> lemme compile
[09:55:51] <paul_c> where are you putting the msgcat ?
[09:56:10] <alex_joni> msgcat is built in
[09:56:16] <alex_joni> in tcl/tk
[09:56:41] <alex_joni> I just put the [msgcat::mc "All files"] instead of "All files"
[09:57:07] <alex_joni> the function msgcat::mc should return a string with the translation, or the same string if no translation is found
[09:57:23] <alex_joni> but I think tk is performing a syntax check which fails
[10:11:00] <CIA-4> 03paul_c * 10emc2/tcl/tkemc.tcl: Small change to the bracing in an attempt to kill a small bug..
[10:11:04] <paul_c> Try that.
[10:12:17] <alex_joni> * alex_joni is checkin out
[10:13:36] <Imperator_> paul_c: about NML, do you think there is a alternative ???
[10:14:07] <Imperator_> something easy to understand by developers
[10:14:17] <Imperator_> let me say EMC developers
[10:14:48] <paul_c> For inter process communications on the same computer, there are several alternatives...
[10:15:18] <Imperator_> at the moment in EMC2 NML is only used for the GUI
[10:15:27] <paul_c> Once you start using networked processes, the choices get very limited...
[10:15:42] <Imperator_> the motion tasks are speaking different languages, I think ???
[10:15:53] <paul_c> and when you throw in OS independant support...
[10:15:58] <alex_joni> paul_c: what you commited.. it stops the bug, but I get the string "msgcat::mc "All files" (*)" in the Open Dialog
[10:16:57] <paul_c> Ooops.
[10:18:27] <Imperator_> is NML realy safe ??? for example if you are running the GUI on a different computer and if that computer crash's, what is the machine doing then ??
[10:20:32] <paul_c> If the GUI crashes, or the computer gets disconnected, EMC will continue to function.
[10:21:01] <Imperator_> so it don't stop running the NC program
[10:21:38] <paul_c> All the GUI does is send instructions and read status.
[10:21:57] <paul_c> EMC does not care where the commands come from.
[10:22:29] <Imperator_> hm, but if the GUI crashs the machine has to estop in my opinion
[10:22:57] <alex_joni> Imperator_: not really
[10:23:16] <alex_joni> estop is hardware
[10:23:33] <alex_joni> and the task should run as long as nobody tells it it should stop
[10:23:49] <alex_joni> that part can be an IO-controller (who sees that estop is on)
[10:23:52] <Imperator_> today it can also be in software if it is implementaded very good
[10:24:03] <alex_joni> or GUI (if user pushes estop switch)
[10:24:12] <alex_joni> well.. EN775 would disagree ;)
[10:24:18] <alex_joni> at least it does for robots
[10:24:43] <alex_joni> it would make my life a lot easier if it would allow SW estop, but it doesn't
[10:24:52] <Imperator_> do you know that safty PLCs from Pilz or Siemens
[10:24:53] <paul_c> HSE also requires a mechanical e-stop
[10:24:53] <alex_joni> so I need 4 extra wires on any teach-pendant
[10:25:06] <Imperator_> HSE ?
[10:25:16] <alex_joni> EN rules specify that you need 2 channels for ESTOP, not one
[10:25:19] <alex_joni> both HW
[10:25:21] <paul_c> Health & Safety Executive.
[10:25:28] <alex_joni> and with separate wires
[10:25:36] <Imperator_> with that Pilz stuf you can have estops on the saftybus
[10:25:45] <alex_joni> that's bogus
[10:26:04] <Imperator_> nope
[10:26:13] <alex_joni> it is in my eyes
[10:26:14] <Imperator_> it's certifyed
[10:26:19] <alex_joni> don't care
[10:26:23] <Imperator_> :-)
[10:26:26] <alex_joni> I wouldn't bet my life on it
[10:26:31] <alex_joni> would you?
[10:26:55] <Imperator_> have you used a cable car the last years ?
[10:27:02] <alex_joni> cable?
[10:27:13] <Imperator_> to go on top of a mountain
[10:27:19] <alex_joni> nope ;)
[10:27:27] <Imperator_> somwhere in west europe
[10:27:29] <alex_joni> actually yes ;)
[10:27:36] <alex_joni> but not western europe
[10:27:39] <Imperator_> ok, then you havent trusted them
[10:28:04] <Imperator_> because Pilz gets big with safty PLCs for that cable cars
[10:28:51] <Imperator_> there are some new standards, for that safty stuff (EN/DIN)
[10:29:33] <Imperator_> we had a EU research project about that http://cimweb.cim.fh-aalen.de/cdi/
[10:32:38] <Imperator_> http://www.pilz.com/german/products/safety/bus/normen.htm
[10:33:23] <alex_joni> anyways
[10:33:27] <Imperator_> english version
[10:33:28] <Imperator_> http://www.pilz.com/english/products/safety/bus/normen.htm
[10:33:35] <alex_joni> the thing started from NML
[10:33:47] <alex_joni> I think estop is pretty OK implemented
[10:34:13] <Imperator_> what I would like to say if there are no safty functions I can also use VNC or something like that to controll EMC from a other computer
[10:34:28] <Imperator_> maybe It is also easyer
[10:36:07] <Imperator_> the question is: is NML realy a advantage for example to control EMc from outside, or are there other technics like VNC that do the same
[10:38:03] <paul_c> VNC does not scale well, and it is also a bandwidth hog
[10:38:15] <alex_joni> ssh -x is a better solution
[10:38:18] <alex_joni> but..
[10:38:21] <Imperator_> ok
[10:38:57] <Imperator_> but on a local network VNC is ok, I use it to administrade our servers
[10:41:13] <paul_c> alex_joni: The line triggering the error is the one with tk_getOpenFile
[10:41:46] <alex_joni> right
[10:42:00] <alex_joni> because types is not set up right
[10:42:27] <Imperator_> EMC2 is now developer friendly. With normal C knolege a developer can get inside in one ore two weeks, but the message layer is terrible
[10:42:45] <alex_joni> paul_c: I tried it like this
[10:42:46] <alex_joni> set allfilestring "All files"
[10:42:46] <alex_joni> set textfilestring "Text files"
[10:42:46] <alex_joni> set ncfilestring "NC files"
[10:42:46] <alex_joni> set types {
[10:42:46] <alex_joni> {$allfilestring *}
[10:42:47] <alex_joni> {$textfilestring {.txt}}
[10:42:49] <alex_joni> {$ncfilestring {.nc .ngc}}
[10:42:51] <alex_joni> }
[10:43:00] <alex_joni> but it won't expand $allfilestring
[10:48:58] <paul_c> there's probably some trick with eval or subst that can be used
[10:49:10] <alex_joni> I'm checking lappend now
[10:49:19] <alex_joni> but.. I'm on unknown ground
[10:50:01] <paul_c> Imperator_: It's not libnml or NML messaging that is the problem with EMC
[10:52:03] <Imperator_> * Imperator_ listens
[10:52:11] <anonimasu> good morn
[10:52:12] <anonimasu> ing
[10:52:18] <Imperator_> hi
[10:53:51] <Imperator_> paul_c: you mean the implementation is the problem
[10:54:27] <paul_c> Partly, yes.
[10:55:06] <Imperator_> and mostly ?
[10:55:43] <alex_joni> hey an0n
[11:01:52] <anonimasu> what are you guys talking about?
[11:02:01] <alex_joni> nml
[11:02:10] <anonimasu> ah ok
[11:12:23] <paul_c> Whilst you will get verbal support from it's detractors, you won't get any real help from the same people when it comes to coding a viable alternative.
[11:13:41] <alex_joni> paul_c: even if I don't like NML, I think it's too deep in emc to be cut-out
[11:16:16] <CIA-4> 03alex_joni * 10emc2/tcl/tkemc.tcl: fixed that stupid tk bug, Open file should work now regardless of LOCALE
[11:16:28] <Imperator_> don't think so alex
[11:16:55] <alex_joni> Imperator_: it can be replaced..
[11:17:01] <alex_joni> but it would be a HUGE amount of work
[11:17:04] <Imperator_> axis for example don't uses it, so its mostly interpreter -> motion
[11:17:12] <alex_joni> and I don't see the programmers stumbling in
[11:17:15] <alex_joni> axis uses it
[11:17:22] <Imperator_> shure ?
[11:17:29] <alex_joni> yup
[11:19:03] <Imperator_> if there is a real aleternative i can try to do it
[11:19:39] <Imperator_> but I need a good concept
[11:19:40] <alex_joni> Imperator_: you'll have about 50k lines to change/replace
[11:20:02] <anonimasu> :)
[11:20:05] <alex_joni> roughly
[11:20:07] <paul_c> There are three main choices....
[11:20:28] <alex_joni> paul_c: you had some statistics about source code if my memory serves me right
[11:20:44] <paul_c> Povray uses a n/w ready IPC lib, but it is some 12Meg in size
[11:21:50] <alex_joni> my question is: what doesn't NML provide?
[11:21:51] <paul_c> then there is simpl, but that requires a major amount of work and would need to be recompiled for most configurations
[11:22:03] <alex_joni> what is it's limitations for developers?
[11:22:13] <paul_c> then there is corba, but that also has a major payload.
[11:22:29] <alex_joni> * alex_joni did some corba once...
[11:22:36] <alex_joni> wouldn't wanna go there
[11:22:45] <alex_joni> different compilers, idl-interpreters
[11:22:50] <alex_joni> all kind of head-ache
[11:23:28] <paul_c> If you go looking for lightweight platform independant IPC libraries that can be reconfigured at runtime, your choices are very limited.
[11:23:29] <alex_joni> we might "pervert" NML a little
[11:23:30] <Imperator_> we need a clean and slim implementation
[11:23:43] <alex_joni> define some generic messages
[11:23:52] <alex_joni> and the data within defines what they actually are/do
[11:25:31] <paul_c> You've spent too long listening to the whingers
[11:25:49] <alex_joni> heh.. talking about robin?
[11:25:55] <alex_joni> *evil grin*
[11:26:06] <paul_c> I didn't mention any names.
[11:26:12] <alex_joni> lol
[11:26:14] <alex_joni> right
[11:26:17] <alex_joni> anyways...
[11:26:29] <alex_joni> I think that could / SHOULD get done for IO
[11:27:39] <paul_c> You need to get to the codeFest to discuss that dort of issue.
[11:28:27] <alex_joni> too bad that's not possible
[11:28:34] <alex_joni> but I'll try to be online during the FEST
[11:28:44] <alex_joni> maybe we can "talk"
[11:36:46] <anonimasu_> * anonimasu_ stretches
[11:36:50] <anonimasu_> anonimasu_ is now known as anonimasu
[11:45:24] <Imperator_> paul_c: what do you mean with RPC libs that can be reconfigured at runtime ???
[11:45:36] <Imperator_> what has to be reconfigured ??
[11:46:48] <Imperator_> I mean if there is a change in the message level nessesary, then the sources have also changed and that causes alwas a new compiler run !
[11:47:13] <Imperator_> or maybe i don't understand something basic
[11:47:37] <paul_c> Timeouts can be defined in the configs,
[11:47:55] <paul_c> as can the protocols (udp, tcp, etc)
[11:48:24] <Imperator_> i think a IPC that runs localy is enough
[11:48:38] <paul_c> you can also define which processes are master or slave for a particular buffer
[11:48:57] <Imperator_> that is more or less fixed in EMC
[11:50:21] <alex_joni> the main thing is with messages that get sent around
[11:50:42] <alex_joni> and the possibility to let the user add/modify/use certain messages without recompiling
[11:51:01] <paul_c> If you want to restrict EMC to being run on a single computer, NML can be dropped entirely.
[11:51:33] <Imperator_> replaced by a memory based IPC
[11:54:00] <Imperator_> I think that it is a nice feature to run the GUI on a diferent computer, but I think we pay a very high price for it. Specialy because there are some other posibilitys to control the GUI from outside, like VNC or telnet
[11:54:25] <robin_sz> ssh -X works well
[11:55:01] <Imperator_> and if we have no special safty functions by NML then ther is no advantage by using NML instead of ssh -X, VNC, ....
[11:55:09] <Imperator_> Hi robin
[11:55:17] <robin_sz> Hi :)
[11:55:57] <Imperator_> maybe most of the users are using vnc or telnet instead of NML
[11:56:07] <alex_joni> I use ssh mainly
[11:56:13] <robin_sz> me too
[11:56:13] <Imperator_> I remember a discussion on that german forumm
[11:56:25] <Imperator_> i think they don't get it running
[11:56:26] <A-L-P-H-A> I use vnc + 128bit encryption
[11:56:53] <Imperator_> hm if nobody uses that feature of emc ????
[11:57:00] <A-L-P-H-A> probably should use ssh... but what am I worried about.
[11:57:10] <robin_sz> nml as a remote extender, is fine ... multi-platform interoperablity etc etc
[11:57:25] <Imperator_> it only displaced new developers, maybe
[11:57:51] <A-L-P-H-A> bbl.
[11:58:04] <robin_sz> what sucks about nml is the mess in the code from the implementation of it ... its preventing the development of the core really
[11:58:31] <robin_sz> paul_c has doen a nice job of tidying it up into some sort of order
[11:58:48] <Imperator_> like my gantry support. gantry support is nothing special but i need to send some messages to the gui
[11:58:50] <robin_sz> but that willnever correct the fundamental desing flaws in the core
[11:59:16] <robin_sz> exactly ... and to extend enc to handle those messages is difficult
[12:00:59] <robin_sz> anyway, its too big a job for one person to do .. to fix it.
[12:01:14] <robin_sz> it needs a policy decision to fix it as a group
[12:01:30] <robin_sz> and that we have never managed to get
[12:02:03] <Imperator_> I can help, but Im not able to make the concept
[12:02:48] <Imperator_> but i can do the work
[12:03:14] <robin_sz> there are plenty of people to help, but it needs direction. it needs an agreed decision on what the aim and the final position will be, thats been the hardest part.
[12:03:40] <Imperator_> thue
[12:03:46] <Imperator_> thrue
[12:04:18] <Imperator_> last try: true
[12:04:30] <robin_sz> personally, I don't think that decision will ever be made ... but I could be wrong. I see its not on the horizon for this years code-fest ... so I guess at least another year.
[12:07:04] <Imperator_> maybe that year is the last chance for a while
[12:07:18] <robin_sz> yeah
[12:07:27] <robin_sz> I had hoped emc2 would address the issue
[12:07:30] <Imperator_> because if EMC2 gets ever stable, nobody touches the message laver ever
[12:07:42] <robin_sz> yep
[12:08:04] <robin_sz> shame really, because that is the biggest setback for emc.
[12:08:12] <robin_sz> the hard-coded messages
[12:08:33] <robin_sz> and the completely incorrect use of C++
[12:09:52] <Imperator_> paul_c: do you think it is possible to spend a day at EMC-Fest for that ???
[12:11:09] <paul_c> On which part ?
[12:11:25] <Imperator_> on finding a alternative to NML
[12:11:44] <Imperator_> a local running IPC
[12:12:20] <Imperator_> implementated in C
[12:12:52] <robin_sz> or sensible C++
[12:13:13] <paul_c> Regardless of opinions expressed here, I doubt if many of the key people within the EMC community will want to dump NML
[12:14:27] <robin_sz> I think keeping the external appearance of NML would be OK, if the internale implementation was fundamentally different
[12:15:34] <robin_sz> and I thought the whole idea of emc2 was to break free from some of those constraints ..
[12:17:10] <paul_c> do you have a viable alternative ?
[12:17:11] <Imperator_> how are the key people ?? You Paul and John have made th core of EMC2. John don't like NML
[12:18:15] <robin_sz> viable alternatives have been discussed at length ... several possible implenetations for a messaging framework have been discussed at lenght ..
[12:19:43] <robin_sz> until the basic requirements of an alternative have been laid down, and a decision made to work towards it as a group ...its all just hot air
[12:20:13] <paul_c> Those lengthy discussions must have taken place either in private /msg or when I've not beem here - Certainly have not seen any discussions about it on the mailing lists.
[12:20:15] <Imperator_> how has changed any messages ever ??? I think only you Paul !! don't know how is also able to change/add messages
[12:20:49] <alex_joni> I think Imperator means who
[12:21:01] <alex_joni> I did add some messages.. just for the fun of it
[12:21:05] <Imperator_> oh, thanks ALex
[12:21:18] <robin_sz> theres been hours and hours of IRC chat on the subject ..
[12:21:28] <paul_c> Matt Shaver has added a couple of messages.
[12:22:26] <Imperator_> two people !
[12:22:44] <Imperator_> we need something that every developer can handle
[12:22:45] <robin_sz> so now those special requirements are hard coded in everyones emc?
[12:22:45] <paul_c> Fred Proctor has added many..
[12:23:05] <alex_joni> robin: let's separate the talks a bit
[12:23:19] <alex_joni> overall NML is pretty ok, and I don't think we would need to drop it
[12:23:35] <alex_joni> afterall we still got some people who can adjust it
[12:23:49] <alex_joni> if we switch to smthg else, we all would be at step 0
[12:24:13] <alex_joni> but I totally agree NML needs some changing (implementation) to get more flexible
[12:25:12] <alex_joni> Imperator_: a PROGRAMMER (and I don't count myself as one) would/should be able to figure NML out in a day or two
[12:25:21] <alex_joni> maybe a c++ one, not a plain-c one
[12:25:25] <alex_joni> but still
[12:25:36] <Imperator_> the question is what do we realy need !!
[12:25:44] <alex_joni> right
[12:25:56] <alex_joni> ok.. what is the answer?
[12:26:43] <Imperator_> do we need realy a system that can run on more than one computer ?? If there are no special safty functions, then we definatly don't need it. Then VNC or ssh can do the same job
[12:27:08] <Imperator_> i think that is the main question !!! Local or distributed !!!
[12:28:02] <alex_joni> I would want to be able to run 2 GUIs on the same machine
[12:28:07] <alex_joni> one remote
[12:28:13] <alex_joni> maybe only for status
[12:28:21] <Imperator_> and if it is running on different computers it must be easy to use
[12:28:21] <alex_joni> so distributed is not a bad choice
[12:28:34] <robin_sz> id like to see NO hard-coding of message tyopes for routing
[12:28:34] <alex_joni> and .. it's already implemented.. so why scrap it?
[12:28:46] <alex_joni> robin: I know you wouldn't
[12:28:59] <alex_joni> how about a middle way?
[12:29:13] <robin_sz> why cant we have both?
[12:29:14] <alex_joni> leave some hardcoded (ones that are improbable to change)
[12:29:22] <alex_joni> and add some generic messages
[12:29:31] <robin_sz> sounds a mess
[12:29:34] <alex_joni> robin: both what?
[12:29:38] <robin_sz> why have two seperate systems?
[12:29:52] <alex_joni> they are not separate systems
[12:29:53] <robin_sz> ignore the both comment
[12:30:02] <robin_sz> yes they are ..
[12:30:27] <Imperator_> with VNC you can do that ALex
[12:30:28] <alex_joni> why?
[12:30:35] <alex_joni> not really Martin
[12:30:58] <alex_joni> with VNC I access the same desktop
[12:30:58] <anonimasu> because of the work involved to change nml to do non hardcoded ones only..
[12:31:03] <robin_sz> because the "hard coded" ones are using a different routing system to the non-hard coded ones
[12:31:16] <alex_joni> why would they?
[12:31:34] <robin_sz> because they are fundamentally different
[12:31:48] <alex_joni> I disagree
[12:31:50] <anonimasu> I disagree with that..
[12:32:01] <robin_sz> one option would be to add a "generic" router ...
[12:32:03] <alex_joni> they don't even exist... how can they be fundamentally different?
[12:32:08] <alex_joni> :P
[12:32:23] <robin_sz> then port all the messages slowly to the generis ssytem
[12:32:26] <alex_joni> robin: if you have a generic_io_NML_command
[12:32:33] <alex_joni> I would accept that
[12:32:39] <alex_joni> but I wouldn't do it myself
[12:33:08] <robin_sz> let me put it another way:
[12:33:26] <robin_sz> I shoule be able to add a message that goes end-to-end throught he system,
[12:33:45] <alex_joni> from where to where?
[12:33:52] <robin_sz> without adding ANY more switch statememtn
[12:34:11] <alex_joni> and how do you plan to modify the components that actually use the message?
[12:34:22] <anonimasu> hm, loading up all possible ones from a config file into a table.. and a routing function.. telling what goes where..
[12:34:46] <robin_sz> from say, coming in from a GUI, to some endpoitn IO, or internal logic
[12:35:06] <robin_sz> they key is routing based on message content, NOT message type or class
[12:35:40] <alex_joni> ok.. let's take IO
[12:35:45] <alex_joni> you add a button to the GUI
[12:36:02] <alex_joni> and say that should be for warm-weather-enable
[12:36:08] <robin_sz> right
[12:36:17] <anonimasu> I think what robin wants is a way to pass messages without recompiling the core of emc...
[12:36:21] <alex_joni> you add a message type for it (or whatever makes your heart beat faster)
[12:36:26] <robin_sz> * robin_sz nods to anonimasu
[12:36:36] <robin_sz> stop
[12:36:41] <anonimasu> from one module(part) to another without re-compiling everything..
[12:36:44] <alex_joni> and you need support for it in the io-controller
[12:36:47] <robin_sz> message type?
[12:36:56] <alex_joni> whatever... ini-param,
[12:37:03] <robin_sz> right ok .. keep going
[12:37:10] <robin_sz> right
[12:37:13] <anonimasu> only having the part handle the message and the other parts just route it through...
[12:37:19] <alex_joni> the io-controller needs to have support for that
[12:37:23] <robin_sz> right
[12:37:25] <anonimasu> since it dosent matter for their operation
[12:37:29] <alex_joni> it gets the message (because it's IO-message)
[12:37:38] <alex_joni> but the data says warm-weather-enable
[12:37:51] <alex_joni> which is unknown right now.. and needs to get implemented
[12:38:03] <alex_joni> but then I ask.. who does the logic?
[12:38:27] <anonimasu> the person that implements the handler for the warm weather enable part..
[12:38:28] <robin_sz> well, it gets the message there becasue the WarmWeatherController I just added to the IO has told the system it lieks to hear about an warm-weather-enable messages it migh have
[12:38:35] <alex_joni> say warm-weather needs to be active 5 seconds, then it needs to pulse 3 times
[12:38:54] <robin_sz> right .. thats in the WeatherController logic
[12:39:25] <robin_sz> it has told the system a list of messages its interested in
[12:39:29] <alex_joni> so you need to modify the IO-controller (extend it with weathercontroller)
[12:39:36] <anonimasu> yeah
[12:39:38] <robin_sz> yep
[12:39:43] <robin_sz> but the key point ..
[12:39:53] <anonimasu> but to add youe message you dont have to re-compile everything inbetween..
[12:39:53] <alex_joni> and the GUI (extend it with warm-weather-enable)
[12:39:56] <anonimasu> yeah
[12:40:08] <robin_sz> is i DO NOT want to add a cases to a hundred switch statements ...
[12:40:08] <alex_joni> you still need to recompile parts of it
[12:40:16] <robin_sz> not the core
[12:40:18] <anonimasu> alex_joni: ofcourse..
[12:40:26] <alex_joni> depends on what you see as the core
[12:40:27] <anonimasu> you change the gui, and you add your handler telling the system what messages you are interested in..
[12:40:33] <anonimasu> and you get the messages there..
[12:40:34] <alex_joni> if you take core=libnml
[12:41:21] <anonimasu> * anonimasu is ranting
[12:41:23] <anonimasu> brb tea :)
[12:41:33] <robin_sz> we've done all this so many times ...
[12:41:50] <robin_sz> we've been saying (pretty much the same thing) for 2 years now
[12:42:15] <Imperator_> so we have to find a solutian to stop it
[12:42:34] <Imperator_> and we have to implementade the solutian
[12:42:53] <Imperator_> i can't beleve that it is that complicated
[12:43:04] <robin_sz> the implementation isnt complex
[12:43:10] <robin_sz> but it affects ALL of emc
[12:43:21] <robin_sz> so it needs to be a group effort
[12:43:29] <robin_sz> too big for one person
[12:43:52] <robin_sz> I started at one point .. but it was too much for me to do.
[12:44:09] <Imperator_> ok who wants to help ???
[12:44:17] <Imperator_> I want to
[12:44:39] <Imperator_> robin ???
[12:44:42] <Imperator_> Alex ???
[12:45:09] <robin_sz> If there was a positive decision with real direction, I would help
[12:45:18] <les> hello...i'm here for a minute
[12:45:28] <robin_sz> my efforts are on G2002 at the moment though
[12:46:11] <alex_joni> hey les
[12:46:27] <anonimasu> I can help.
[12:46:42] <Imperator_> ok i think we need also Alex
[12:47:13] <alex_joni> I can help.. but I agree that we need a direction first
[12:47:18] <anonimasu> * anonimasu nods
[12:47:22] <anonimasu> that's the first step
[12:47:24] <Imperator_> of cause
[12:47:25] <robin_sz> I think you need paul_c and jmkasunich too
[12:47:33] <Imperator_> right
[12:47:42] <Imperator_> ok time for a new Wiki Page
[12:47:53] <robin_sz> oh god not another ;)
[12:48:03] <alex_joni> lol
[12:48:15] <Imperator_> ok Robin your part
[12:48:20] <Imperator_> :-)
[12:48:50] <Imperator_> btw. I think i have created two dead pages yesturday
[12:49:02] <Imperator_> can you kill them robin ?
[12:49:10] <Imperator_> it's your Wiki, right ???
[12:53:06] <anonimasu> * anonimasu is back
[12:53:33] <alex_joni> wb
[12:54:01] <anonimasu> thanks
[13:01:37] <robin_sz> Imperator_: yeah, I can kill them
[13:01:42] <alex_joni> alex_joni is now known as alex_joni_away
[13:01:47] <robin_sz> its not *my* wiki ...
[13:01:50] <Imperator_> ok
[13:01:55] <robin_sz> its emc's wiki :)
[13:02:05] <Imperator_> :-)
[13:02:11] <Imperator_> on your server, right
[13:02:15] <robin_sz> right
[13:02:18] <Imperator_> ok
[13:02:37] <robin_sz> if anyone wants the admin password ... just ask
[13:04:27] <robin_sz> I dont see your "dead" pages ...
[13:07:33] <robin_sz> sigh .. the new yahoo groups layout is bloody awful
[13:07:48] <robin_sz> bunch of fsckwits
[13:08:36] <Imperator_> I have added yesturday the FirstEuCodeFest page
[13:08:47] <robin_sz> thats not needed?
[13:09:22] <Imperator_> I don't know, but i think i have added three pages to get it
[13:09:30] <robin_sz> no,
[13:09:35] <robin_sz> I think you didnt
[13:09:41] <Imperator_> ah ok, very good
[13:10:10] <robin_sz> I think you added links that were in 3 differetn formats until you got the one you wanted
[13:10:21] <robin_sz> but creating a link doesnt create a page ...
[13:10:36] <Imperator_> so ther is only one "FirstEuCodeFest" and not a second one with a name like "EUCodeFest" or EUEMCFest" ?
[13:10:46] <robin_sz> correct
[13:10:49] <Imperator_> ah ok
[13:11:02] <robin_sz> when you make a link .. it comes LikeThis[?]
[13:11:12] <Imperator_> jep
[13:11:20] <robin_sz> when you click [?] .. edit thpage and SAVE ... thats when the new page is created
[13:11:28] <Imperator_> then i have to klick on it an to fill something in
[13:11:32] <Imperator_> ok
[13:11:34] <robin_sz> jep
[13:11:42] <robin_sz> doh .. I did it now ..
[13:11:46] <robin_sz> yep :)
[13:12:27] <robin_sz> ja and jep , ya and yep .. .makes sense I guess :)
[13:20:09] <les> ho hum taking a few minutes to buy tires for the truck
[13:23:35] <alex_joni_away> alex_joni_away is now known as alex_joni
[13:26:25] <alex_joni> * alex_joni starts to add an IO generic message
[13:38:03] <robin_sz> mmm
[13:38:05] <robin_sz> SOAP
[13:38:08] <robin_sz> ;)))
[13:38:09] <alex_joni> soap?
[13:38:12] <robin_sz> only joking
[13:38:13] <alex_joni> that rings a bell
[13:38:45] <robin_sz> http://www.w3.org/TR/soap/
[13:39:10] <robin_sz> http://www.w3.org/TR/2003/REC-soap12-part1-20030624/
[13:39:15] <robin_sz> the mesaging layer ;)
[13:39:17] <alex_joni> * alex_joni goes home ;)
[13:39:26] <robin_sz> yeah, Im joking ..
[13:39:28] <alex_joni> or not...
[13:39:42] <anonimasu> haha
[13:39:43] <anonimasu> :)
[13:40:19] <robin_sz> SOAP Version 1.2 (SOAP) is a lightweight protocol intended for exchanging structured information in a decentralized, distributed environment.
[13:40:59] <robin_sz> It uses XML technologies to define an extensible messaging framework providing a message construct that can be exchanged over a variety of underlying protocols. The framework has been designed to be independent of any particular programming model and other implementation specific semantics.
[13:41:13] <anonimasu> shame jymm isnt here ;D
[13:41:19] <robin_sz> tee hee
[13:41:46] <anonimasu> * anonimasu is all for xml :p
[13:42:03] <alex_joni> maybe we can do SOAP over NML
[13:42:10] <alex_joni> LOLOLOLOLOLOLOL
[13:42:20] <anonimasu> alex_joni: a slippery solution
[13:42:29] <alex_joni> heh
[13:44:47] <alex_joni> I was kidding !!!!!!
[13:46:05] <anonimasu> I was dead serious ;)
[13:47:08] <alex_joni> s/serious//
[13:47:36] <anonimasu> :p
[13:47:43] <alex_joni> -_^
[13:58:23] <robin_sz> http://www.redpoint.org.uk/cgi-bin/emcinfo.pl?MessagingLayer
[13:58:49] <alex_joni> seen that ;)
[13:59:02] <alex_joni> * alex_joni finished adding an IO generic message
[13:59:14] <alex_joni> roughly 20 mins
[14:04:13] <anonimasu> nice
[14:05:54] <alex_joni> http://www.redpoint.org.uk/cgi-bin/emcinfo.pl?CustomNMLMessage
[14:06:33] <anonimasu> hm, that dosent seem too hard
[14:06:41] <alex_joni> an0n: right
[14:06:45] <anonimasu> * anonimasu is digging in segmentqueue
[14:07:31] <les> what are you finding?
[14:07:38] <alex_joni> segments
[14:07:41] <alex_joni> LOL
[14:07:42] <les> heh
[14:07:47] <alex_joni> a full queue of them
[14:07:51] <anonimasu> yeah
[14:07:55] <les> bet they are queued up too
[14:07:59] <anonimasu> * anonimasu breaks a segment
[14:08:14] <alex_joni> see.. now you have even more
[14:08:19] <anonimasu> I just wish I understood the math,
[14:08:41] <les> I wish I understood the code
[14:08:53] <alex_joni> I wish I understood the world
[14:09:04] <les> actually I can read the math part ok
[14:09:18] <anonimasu> :)
[14:09:35] <les> reminds me...waiting for the shop to warm up...
[14:09:47] <les> I think I will look at the diffs
[14:09:58] <anonimasu> diffs?
[14:10:05] <alex_joni> for the temp. of course
[14:11:11] <anonimasu> ok
[14:12:52] <anonimasu> hm maybe I should take a drive to the kiosk.
[14:13:00] <anonimasu> caffeine helps code.
[14:14:24] <les> oh I see...cradek worked on the negative length panic
[14:14:32] <alex_joni> yup
[14:14:38] <alex_joni> cradek fixed that
[14:14:52] <les> and what else here?
[14:19:22] <les> that seems mostly it
[14:20:22] <les> so I expect to see no more halting, but paerhaps still smooth motion with non optimal velocity
[14:21:31] <robin_sz> ooh look! ... nice new 160gb drive :)
[14:21:50] <les> if that is corrected I think we will have a pretty functional TP
[14:22:19] <les> 160s going for about $80 here
[14:22:40] <les> incredible
[14:23:44] <les> I have ended up with 3 80s in this box
[14:23:51] <les> 2 are hardware raid
[14:24:01] <alex_joni> 80s are the cheapest
[14:24:03] <les> other is just IDE for backup
[14:26:39] <Imperator_> you are looking at the TP of emc1, right ?
[14:27:01] <alex_joni> there's no TP in emc2
[14:27:11] <alex_joni> I mean no segmentqueue
[14:27:36] <les> might as well fix it first before it is included
[14:27:51] <anonimasu> * anonimasu nods
[14:28:10] <les> only problem I have is the inability to test it beyond 3 axes
[14:28:24] <anonimasu> beyond?
[14:28:32] <les> but as I understand it will only work with 3 cart axes anyway
[14:28:52] <anonimasu> rotary + tilt?
[14:29:00] <les> unknown
[14:29:03] <anonimasu> ok
[14:29:20] <anonimasu> I dont get the math so I better drop digging into it..
[14:29:35] <les> It looks complicated
[14:30:08] <Imperator_> in EMC2 there is more or less segmentqueue implemented, right ?
[14:30:12] <les> but it is just a series of initial, final, and in zone constraints on a series of cubic equations
[14:30:36] <les> martin, I understand not
[14:30:51] <anonimasu> you calculate the maximum velocity following a given path.. with a series of rules..
[14:31:19] <les> rules for continuity yes
[14:31:25] <anonimasu> yep
[14:31:51] <les> of cource with a cubic jerk will be discontiuous
[14:32:14] <les> But with the old TP accel was discontinuous
[14:32:25] <les> and jerk was unbounded
[14:32:34] <anonimasu> hm, isnt there a algorithm, that solves both?
[14:33:08] <les> continuous accel and jerk requires a fifth order plynomial
[14:33:39] <les> doable but the matrices for the conditions start to get pretty big
[14:33:44] <anonimasu> * anonimasu nods
[14:33:59] <anonimasu> hm, cant you implement that in a FPGA?
[14:34:11] <les> sure why not
[14:34:25] <les> would have to be big
[14:35:07] <SWPadnos> use 64-bit FP math
[14:35:15] <SWPadnos> (F as in fixed)
[14:35:28] <les> book is not handy but imagine solving an 8x8 diagonal matrix
[14:35:41] <les> that I think is required for just cubic
[14:35:44] <SWPadnos> Though a 5th order polynomial might need floats
[14:35:54] <alex_joni> hey SWP
[14:35:57] <SWPadnos> Hi there
[14:36:09] <alex_joni> missed some discussions about NML today
[14:36:11] <alex_joni> :)
[14:36:19] <SWPadnos> I'll look through my logs :)
[14:38:23] <SWPadnos> hmmm - too bad I was asleep at the time :)
[14:39:15] <alex_joni> heh
[14:40:14] <anonimasu> * anonimasu sighs
[14:40:18] <anonimasu> I need to order a book.
[14:41:36] <anonimasu> brb kiosk
[14:41:37] <SWPadnos> "20,000 Leagues Under The Sea"?
[14:41:43] <alex_joni> heh
[14:41:53] <alex_joni> Fine Thai Cooking for Dummies
[14:42:02] <SWPadnos> "Machinery's Handbook"?
[14:42:17] <alex_joni> Guide to EMC
[14:42:18] <alex_joni> :D
[14:42:23] <SWPadnos> "RCS Handbook"?
[14:42:28] <alex_joni> NML for Dummies
[14:42:29] <alex_joni> :D
[14:42:37] <robin_sz> babes monthly
[14:42:38] <SWPadnos> OOOhhhh - I'll take two
[14:42:44] <SWPadnos> of each
[14:43:02] <robin_sz> special "scratch-n-sniff" issue
[14:43:02] <SWPadnos> No -wait - I just got an ad for one (in email) :)
[14:43:10] <alex_joni> right
[14:44:10] <robin_sz> I cant understand all these "super viaggra" emails ...
[14:44:39] <SWPadnos> Bigger, Stronger, Slower that it was before
[14:44:49] <robin_sz> the selling point seems to be .. "rock hard for 48hrs continuous" .. I mean, thats going to be very inconvenient
[14:45:07] <robin_sz> you'll be taking a hacksaw to it after 24hrs ;)
[14:45:09] <alex_joni> lol
[14:45:16] <SWPadnos> Someone will be
[14:45:24] <alex_joni> robin: adding a message seems trivial ;)
[14:45:46] <robin_sz> yeah?
[14:45:56] <alex_joni> still doesn't seem right.. but it's not that hard
[14:46:13] <robin_sz> I think I found 12 locations to add a switch case when I did it
[14:46:21] <alex_joni> don't think so
[14:47:20] <SWPadnos> there is a 480 line case statement in that file - I'd say something isn't quite optimal
[14:47:30] <alex_joni> SWP: I agree
[14:47:39] <alex_joni> that needs reconsideration ;)
[14:47:45] <robin_sz> thats only ONE of the places
[14:48:02] <alex_joni> right.. the core
[14:48:14] <robin_sz> I still maintain a bloddy great big siren should have gone off in someones head wheh the wrote that
[14:48:14] <alex_joni> next you need the 2 parts that need to communicate
[14:48:34] <alex_joni> which you need to modify anyways
[14:48:38] <robin_sz> true
[14:48:43] <alex_joni> anyways... I'm going home now
[14:48:49] <robin_sz> good plan
[14:48:51] <alex_joni> I got an appointment ;)
[14:48:58] <alex_joni> massage :)
[14:49:03] <robin_sz> nice
[14:49:13] <alex_joni> yeah
[14:49:45] <alex_joni> bye guys
[14:49:50] <robin_sz> "im sorry, your regulur masseuse couldnt come, we sent Kurt The Babarian instead ;)
[14:50:19] <alex_joni> * alex_joni swaps apointments with robin_sz
[14:50:22] <robin_sz> please scream if you need assistance
[14:50:45] <alex_joni> even appointments
[14:50:55] <robin_sz> * robin_sz runs for cover
[14:51:06] <alex_joni> lol
[14:51:07] <alex_joni> bye
[14:51:56] <robin_sz> hmmm ...
[14:52:00] <robin_sz> in C++
[14:52:10] <robin_sz> when someone refers to a variable as :
[14:52:16] <robin_sz> ::someVar
[14:52:27] <robin_sz> does that just mean "in the current class"
[14:53:23] <anonimasu> yeah
[14:53:36] <anonimasu> if I remember it right..
[14:53:50] <anonimasu> usually you do "classname::variable"
[14:53:55] <anonimasu> inside of the class..
[14:54:13] <anonimasu> that gets replaced with the name of your object.. during runtime
[14:54:13] <robin_sz> so I can just dump the :: infornt of the var then?
[14:54:21] <anonimasu> I dont know if that works
[14:54:24] <anonimasu> class::var
[14:54:29] <anonimasu> works :)
[14:54:30] <robin_sz> so ::someVar is the same as someVar
[14:55:00] <anonimasu> yes
[14:55:04] <robin_sz> 'k
[14:55:12] <robin_sz> because it was screwing up in indent ;)
[14:55:31] <anonimasu> ;)
[14:55:53] <robin_sz> baz = (foo == bar)? :: someVar : ::someOtherVar;
[14:56:02] <jepler> uh, you can write that?
[14:56:05] <robin_sz> baz = (foo == bar)? ::someVar : ::someOtherVar;
[14:56:11] <anonimasu> uh?!
[14:56:21] <anonimasu> let me try that..
[14:56:21] <anonimasu> :p
[14:56:23] <robin_sz> yes you *can* ... but you shouldnt :)
[14:56:28] <robin_sz> or something like it
[14:56:31] <anonimasu> you should write classname:var
[14:56:33] <robin_sz> the terhanry thing
[14:56:43] <robin_sz> yep
[14:56:45] <anonimasu> err ::
[14:56:47] <robin_sz> or just var
[14:56:49] <anonimasu> yeah
[14:57:13] <jepler> ever think we'd all be better off if EMC was written in C, not C++?
[14:57:18] <jepler> </c++ hater>
[14:57:23] <robin_sz> no
[14:57:25] <anonimasu> jepler: not really
[14:57:27] <robin_sz> thats C
[14:57:33] <jepler> :: isn't C.
[14:57:41] <robin_sz> well, apart from that bit
[14:57:51] <robin_sz> baz = (foo == bar)? someVar : someOtherVar;
[14:57:56] <jepler> robin_sz: in your ?: expression, are "foo" and "bar" classes or variables?
[14:58:01] <jepler> the one with the ::s
[14:58:03] <robin_sz> variables
[14:58:25] <robin_sz> ::foo just means "foo in the current class"
[14:58:30] <robin_sz> same as foo
[14:58:54] <robin_sz> SomeOtherClass::foo is foo in another class
[14:59:02] <anonimasu> hm.. you cant do ::
[14:59:15] <anonimasu> hej.cpp:10: `::var' undeclared (first use here)
[14:59:45] <robin_sz> well, you have to decalre it in the lcass header obviously .. wait while I find the (awful) code
[14:59:58] <anonimasu> either var or class::var
[15:00:16] <anonimasu> :)
[15:00:23] <jepler> http://craie.unpy.net/~jepler/colons.cc
[15:00:32] <robin_sz> (gtf = new FXTextField (mat, 8,
[15:00:32] <robin_sz> (new EMDataTarget (EPARM mach.pos[2]))->
[15:00:32] <robin_sz> units (&mach.progUnits,
[15:00:32] <robin_sz> units.linZscale)->
[15:00:32] <robin_sz> swit (::lin_printf_strings, 0,
[15:00:33] <robin_sz> 3)->conditional_on (axis_dt[2],
[15:00:35] <robin_sz> TRUE)->
[15:00:37] <anonimasu> what kind of c++ is that?
[15:00:37] <robin_sz> conditional_hide (), FXDataTarget::ID_VALUE,
[15:00:37] <anonimasu> :p
[15:00:41] <robin_sz> setFont (gaugeFont);
[15:00:42] <jepler> If the class has a variable x that shadows a global x, ::x refers to the global
[15:00:43] <robin_sz> oops. sorry.
[15:00:44] <anonimasu> printf?!
[15:00:45] <anonimasu> :p
[15:00:47] <robin_sz> shitty?
[15:00:55] <robin_sz> ar right
[15:01:11] <robin_sz> globals. CICK. worse than I though
[15:01:19] <jepler> my example isn't about stdio vs iostream, it's about the meaning of ::
[15:01:54] <robin_sz> right
[15:01:56] <anonimasu> yeah
[15:02:19] <robin_sz> actually, that wasnt the ?: example I was looking for
[15:02:38] <anonimasu> it dosent works the same.. when I do it..
[15:02:47] <robin_sz> swit (id < 3 ?::lin_gauge_strings_abs : ::ang_gauge_strings_abs,
[15:02:47] <robin_sz> 0, id < 3 ? 3 : 4)
[15:02:52] <robin_sz> ick .. lokk at that
[15:02:56] <robin_sz> almost unreadable
[15:03:23] <robin_sz> gives C++ a bad name :(
[15:03:41] <jepler> don't worry, you can't tarnish my opinion of C++ any further
[15:03:43] <SWPadnos> that gives the programmer a bad name
[15:04:22] <robin_sz> i think EMC's use of C++ is the worst example Ive seen ...
[15:04:32] <robin_sz> it would not be fair to judge C++ on that
[15:04:45] <SWPadnos> yeah - it looks like it's only being used for compile-time type checking
[15:04:48] <anonimasu> jepler: your example works only if you have x as a global..
[15:04:59] <robin_sz> right
[15:05:03] <anonimasu> and implictly wants to use your redefination of the variable..
[15:05:06] <jepler> anonimasu: yes, otherwise ::x refers to something that was not declared
[15:05:27] <robin_sz> so in ::var the class is '' ie .. the root, global base class
[15:05:32] <anonimasu> jepler: nope.. it's declared within the class..
[15:06:01] <anonimasu> still :: should always be in the class imo..
[15:06:05] <robin_sz> so you need a x declared as a global outisde the class
[15:06:18] <robin_sz> ick. globals.
[15:06:29] <anonimasu> yeah if you want to use :: to access the variable within your class..
[15:06:29] <robin_sz> there is NO excuse for globals
[15:06:35] <SWPadnos> if you have a function like SetMode(ModeType Mode), and there's a class global variable named Mode
[15:06:50] <SWPadnos> then you do ::Mode = Mode, and everything's fine
[15:06:57] <anonimasu> yeah
[15:07:03] <anonimasu> but well that looks so horrid..
[15:07:05] <anonimasu> it hurts my eyes :D
[15:08:15] <anonimasu> class::variable looks better..
[15:08:49] <jepler> anonimasu: I use "The C++ Programming Language" by Strostrup as my reference. Page 847, section C.10.1 has the line of source '::k++; // the global k'. I don't think that '::k' ever refers to a member of a class.
[15:09:06] <anonimasu> hm..
[15:09:18] <SWPadnos> it does if it's in a class member function
[15:09:22] <SWPadnos> (I think)
[15:09:34] <anonimasu> SWPadnos: you are right..
[15:09:51] <anonimasu> hm, actually it refers to the global..
[15:09:58] <robin_sz> jepler: that sounds right
[15:10:19] <anonimasu> I tried it out..
[15:10:20] <jepler> SWPadnos: If there is a parameter 'x' and a class member 'x', then 'x' refers to the parameter, and 'this->x' refers to the member. 'C::x' (if C is the type) seems to, too.
[15:10:46] <anonimasu> hm :: refers to the global and not to the class member..
[15:10:57] <robin_sz> * robin_sz checks The Book
[15:11:02] <anonimasu> very very strange.
[15:11:38] <robin_sz> oh yes, bottom of page 847, top of 848
[15:12:13] <anonimasu> hm, I thought :: always refered to things inside the current class..
[15:12:30] <robin_sz> seems not
[15:12:30] <anonimasu> in the same way class::thing always refers to the object..
[15:12:32] <anonimasu> yeah
[15:12:43] <anonimasu> it dosnet work that way..
[15:12:49] <SWPadnos> ok - I've been looking for my copy of the book for a week :)
[15:13:08] <anonimasu> dosent..
[15:13:14] <robin_sz> doesn't
[15:13:29] <anonimasu> I wrote a example and it references to the global..
[15:17:41] <anonimasu> #include <iostream.h>
[15:17:41] <anonimasu> int var = 1024; class hej {public:int var;void hej::test() { var=0; cout << "var: " << var << endl << "hej::var: " << hej::var << endl << "::var: " << ::var << endl; }};
[15:17:41] <anonimasu> int main(void) { hej te; te.test();return(0);
[15:17:41] <anonimasu> }
[15:17:46] <anonimasu> ;)
[15:18:12] <anonimasu> like that..
[15:18:29] <anonimasu> if you run indent on it you can read it again :)
[15:19:09] <jepler> what does "hej" mean?
[15:19:14] <anonimasu> hello in swedish
[15:19:21] <anonimasu> or more like "hi"
[15:20:28] <anonimasu> I need to brush up on my c++
[15:20:36] <anonimasu> it's been lots of time since I last coded any
[16:16:43] <robin_sz> http://www.mech.uwa.edu.au/jpt/shearmagic/tour.html
[16:16:45] <robin_sz> tee hee
[16:18:02] <anonimasu> :)
[16:19:06] <pemmet> anyone here know how to use steppermod?
[16:20:24] <paul_c> steppermod is depreciated.
[16:20:49] <pemmet> what should i use?
[16:21:10] <pemmet> (i got my sister's laptop, so i can sit next to my setup here in the shop_
[16:21:12] <pemmet> )
[16:21:31] <paul_c> if you're using RTAI, you only have one choice with steppers...
[16:21:53] <pemmet> RTAI? uh oh... what does that mean?
[16:22:33] <paul_c> using BDI-4.xx ?
[16:22:39] <pemmet> yea..
[16:22:46] <pemmet> .o2, i think
[16:23:12] <paul_c> 4.20 ?
[16:23:33] <pemmet> heh.. um... where's the quickest place to check within emc?
[16:23:52] <paul_c> uname -a
[16:25:17] <pemmet> oops
[16:25:21] <pemmet> got disconnected...
[16:25:22] <pemmet> but yea
[16:25:29] <pemmet> i typed in uname -a to the consol
[16:25:58] <pemmet> and i got localdomain 2.6.9 -adeos #1 sun Dec 19 15:52:35 GMT 2004 i686 GUN/Linux
[16:26:48] <paul_c> OK... Looks like you still have 4.08 installed.
[16:26:54] <pemmet> ok
[16:27:20] <paul_c> be advised, there is a critical bug in the EMC version on that release.
[16:27:21] <pemmet> all i've got is a 2D table. stepper motor driven...
[16:27:23] <pemmet> ?
[16:27:24] <pemmet> uh oh
[16:27:37] <pemmet> so i suppose i need an update/patch or soemthing
[16:28:18] <paul_c> that release also had a broken version of synaptic
[16:28:47] <pemmet> well all i need the software to do is tell my hardware how fast and in what direction to spin...
[16:29:07] <pemmet> but maybe those bugs might get in the way of that, huh
[16:30:18] <pemmet> i made myself a parallel cable that put all my hardware's pins in the right spot for the software on my computer..
[16:30:23] <paul_c> You should still be able to spin the motors
[16:30:25] <pemmet> but i want to make sure that it's sending the right signals...
[16:30:39] <pemmet> you mean manualy?
[16:31:00] <paul_c> yes - The EMC bug only bites when you home an axis.
[16:31:08] <pemmet> ahhh
[16:31:15] <pemmet> is that the thing where it just keeps going?
[16:31:29] <pemmet> i hit home a few times, and it just took off... (the software, not the hardware) and kept moving
[16:31:33] <pemmet> or thinking it was moving
[16:31:45] <pemmet> since it still doesn't run my hardware, nothing was in danger...
[16:31:47] <pemmet> but is that the bug?
[16:32:05] <paul_c> yes
[16:32:10] <pemmet> ok
[16:32:12] <pemmet> well...
[16:32:21] <pemmet> i can opt not to ever use that button...
[16:32:23] <pemmet> for now
[16:32:28] <pemmet> until i can get the update
[16:33:01] <paul_c> sure...
[16:33:38] <pemmet> i just dont have the means to get it on the machine just this moment... i need to burn it a CD... but i can do that a lill later, probably
[16:34:07] <pemmet> but anyway.. my main problem is with tweaking the output of EMC to send the correct signals to my hardware...
[16:34:19] <pemmet> is that somewhere in the ini file?
[16:34:22] <pemmet> or is it somewhere else
[16:34:57] <paul_c> /usr/local/emc/generic.ini
[16:35:14] <pemmet> does it matter if i'm using a custom file.ini?
[16:35:21] <pemmet> it's my emc.run thing points to it...
[16:35:25] <pemmet> or should i go with generic
[16:35:54] <paul_c> use a custom ini, and you have the original as a backup.
[16:35:54] <pemmet> tell you what.. i can bring up both
[16:35:58] <pemmet> yea
[16:35:58] <pemmet> ok
[16:35:59] <pemmet> good
[16:36:01] <pemmet> that's what I got
[16:36:47] <paul_c> in [EMCMOT], you have EMCMOT=freqmod ?
[16:37:07] <pemmet> that is the current setting, yes
[16:37:26] <paul_c> and IO_BASE_ADDRESS=0x378 ?
[16:37:32] <pemmet> yep..
[16:38:52] <paul_c> cat /proc/ioports | grep parport
[16:39:12] <pemmet> should i type that into the konsol?
[16:39:17] <paul_c> yup
[16:39:59] <pemmet> it told me 'permission denied'
[16:40:02] <pemmet> i AM logged in root
[16:40:09] <pemmet> so what might that be about?
[16:40:45] <paul_c> NNOOooooo.....
[16:40:52] <pemmet> uh oh
[16:40:57] <pemmet> i dont like the sound of that....
[16:40:58] <paul_c> Should never need to log in as root...
[16:41:02] <pemmet> no?
[16:41:14] <pemmet> well then
[16:41:26] <paul_c> but you're there now...
[16:41:34] <pemmet> yea...
[16:42:01] <pemmet> i read the quote at the end of your last email to me... "to err is human... to really F&%@ things up you need the root password"
[16:42:20] <paul_c> rm -fR /*
[16:42:30] <paul_c> Kills the whole system dead.
[16:42:37] <pemmet> well
[16:42:38] <paul_c> (so don't do it...)
[16:42:40] <pemmet> haha
[16:42:42] <pemmet> ok
[16:42:45] <pemmet> you convinced me
[16:42:59] <pemmet> and i wont fall for it when one of the young frehman CS students tells me to do it..
[16:43:27] <pemmet> something akin to deltree *.* or whatever
[16:43:37] <robin_sz> cat /dev/random > /dev/hda works pretty well too
[16:43:43] <paul_c> dd if=/dev/random of=/dev/hda
[16:43:57] <pemmet> heh..
[16:43:58] <paul_c> Another good way to kill a system as root
[16:44:13] <pemmet> ok.. so what should i REALLY type in? :P
[16:44:41] <robin_sz> the real trick is to keep it running long enough to do real damage to the disks, you dont want it falling over before its throuroughly stuffed
[16:44:41] <paul_c> exit
[16:44:58] <pemmet> ok
[16:45:03] <pemmet> typed that..
[16:45:07] <pemmet> closed my consol
[16:45:13] <pemmet> *konsol
[16:45:44] <paul_c> Open another console, and we'll go through your system.
[16:45:50] <pemmet> ok
[16:46:05] <pemmet> k.. got one open
[16:46:20] <paul_c> tip: If you need to know what a command does, type:
[16:46:24] <paul_c> man <command>
[16:46:30] <pemmet> cool
[16:47:15] <paul_c> First thing to do is to check for the parallel port....
[16:47:24] <paul_c> ls /proc/io*
[16:47:46] <pemmet> it gave me /proc/iomem /proc/ioports
[16:48:05] <paul_c> OK... Now try
[16:48:13] <paul_c> cat /proc/ioports
[16:48:30] <pemmet> mmmmkay
[16:48:35] <pemmet> gave me all kinds of stuff
[16:48:49] <paul_c> cat /proc/ioports | grep parport
[16:48:51] <pemmet> what specificly should i look for in this list?
[16:48:55] <pemmet> ok
[16:49:27] <pemmet> ok.. typed that last one in... and it just took it... no returned infor
[16:49:29] <pemmet> info
[16:49:45] <paul_c> paul@Babylon-117:~/EMC/emc2-bdi-4$ cat /proc/ioports | grep parport
[16:49:45] <paul_c> 0378-037a : parport0
[16:49:46] <paul_c> 037b-037f : parport0
[16:49:46] <paul_c> 0778-077a : parport0
[16:49:47] <paul_c> b000-b002 : parport1
[16:49:47] <paul_c> b800-b802 : parport2
[16:50:10] <pemmet> i got some parport1''s
[16:50:13] <pemmet> just 2 of 'em
[16:50:27] <pemmet> 10e0-10e2 : parport1
[16:50:27] <paul_c> but no parport0 ?
[16:50:42] <pemmet> oh wait.. yes i do
[16:50:45] <pemmet> didn't see 'em up there
[16:50:52] <pemmet> they are...:
[16:51:09] <pemmet> 0378-037a : parport0
[16:51:21] <pemmet> 037b-037f : parport0
[16:51:25] <paul_c> OK... That's the one we want
[16:51:32] <pemmet> excelent
[16:51:33] <paul_c> (the first one..)
[16:51:43] <pemmet> ok
[16:51:58] <paul_c> Now for a really silly question.....
[16:52:04] <pemmet> :)
[16:52:08] <pemmet> no worries... i'm new to all this
[16:52:21] <paul_c> You do have the cable plugged into the correct parallel port ?
[16:52:34] <pemmet> well... there is only one parallel port...
[16:52:51] <pemmet> actualy there are 2, but the second one was JUST installed by me a week ago when i thought i still needed 2
[16:53:11] <pemmet> it's plugged into my printer's parallel port...
[16:53:19] <pemmet> or the one that might have origionaly been for a printer
[16:53:42] <pemmet> i have the layout of my pins written down..
[16:54:01] <pemmet> and i tried to line them up for speed/direction with what the handbook wanted..
[16:54:26] <pemmet> suing the funcion step/dir column of the parallel help section in the handbook
[16:54:31] <pemmet> *using
[16:56:05] <pemmet> so to answer your question.. yes, i believe it to be in the right port :)
[16:56:12] <paul_c> Well... If the cable is wired up right, and everything is powered up, the motors shoud spin
[16:56:20] <pemmet> riiight
[16:56:24] <pemmet> that's what i would have thought
[16:56:48] <paul_c> Can you get hold of an 'scope ?
[16:56:58] <pemmet> a scope...
[16:57:02] <pemmet> oscilliscope?
[16:57:08] <paul_c> yes.
[16:57:08] <pemmet> or however you spell that?
[16:57:14] <pemmet> i bet i could
[16:57:25] <paul_c> "oscilloscope"
[16:57:28] <pemmet> i dont have one handy, but give me half an hour and i could whip one up, probably
[16:57:36] <pemmet> maybe less time
[16:57:46] <pemmet> the ECE building is a 1 min walk out the door
[16:57:52] <paul_c> OK - Gives me time to grab some tea.
[16:57:57] <pemmet> heh
[16:57:57] <pemmet> ok
[16:58:01] <pemmet> i'll go hunt one down
[16:58:04] <pemmet> what should i get with it?
[16:58:14] <pemmet> i mean what equipment should i bring with the scope?
[16:58:16] <paul_c> probes, power cord...
[16:58:20] <pemmet> heh
[16:58:21] <pemmet> ok
[16:58:25] <pemmet> brb
[16:58:31] <paul_c> back in 30 mins.
[17:20:50] <pemmet> wow... believe it or not, there was one 10' from my bench
[17:22:29] <anonimasu> paul_c: dd is nice ^__^
[17:23:24] <pemmet> and i liked your tea idea so much, i got some too
[17:23:37] <pemmet> man... this is an old scope
[17:24:28] <pemmet> shoot
[17:24:30] <pemmet> i need probes
[17:24:33] <pemmet> and the ones i have dont fit
[17:24:38] <pemmet> give me 5 more min
[17:24:41] <pemmet> :D
[17:25:39] <pemmet> unless a multimeter would work
[17:25:40] <pemmet> ?
[17:25:50] <pemmet> would one of those get us the info we need?
[17:28:11] <pemmet> sweet
[17:28:23] <pemmet> got an o-scope and probes AND a multimeter...
[17:33:43] <les> uh oh...another tsunami...
[17:33:46] <les> same place
[17:36:24] <les> hpefully not as big.
[17:37:20] <pemmet> nuts
[17:37:29] <pemmet> i got to get to a couple of classes that I've got...
[17:38:02] <pemmet> i'll be back at 2ish...
[17:38:13] <pemmet> it's 12:48 where i am now..
[18:55:39] <SWPadnos> Out of curiosity - does anyone know if RTAI works on 64-bit platforms?
[18:57:35] <Jymmm> Mornin!
[18:58:22] <SWPadnos> afternon!
[18:58:25] <SWPadnos> noon
[18:58:28] <SWPadnos> non-morning
[18:58:42] <paul_c> SWPadnos: Donate an AMD 64, and we'll find out ;)
[18:59:33] <SWPadnos> heh - I'll be ordering an Opteron system next week, so maybe you'll have to borrow it :)
[18:59:42] <Jymmm> wouldn't it be funny if Intel released a 128bit cpu - that blow AMD out of the water!
[18:59:51] <SWPadnos> It may be a laptop, so I might be able to bring it to Fest
[19:00:38] <paul_c> Jymmm: 10 years ago, I was mooting the possiblity of 128 bit processors...
[19:00:48] <SWPadnos> mooting?
[19:01:07] <Jymmm> paul_c we REALLY should have 256bit CPU's in this day and age.
[19:01:24] <paul_c> at the time, we were messing around with bit slice boards with 128 bit data bus.
[19:01:33] <SWPadnos> we really should have software that can run acceptably on 32-bit CPUs
[19:01:50] <Jymmm> SWPadnos blame lazy coders
[19:02:01] <SWPadnos> yeah - us bastards
[19:02:03] <SWPadnos> wait ;)
[19:02:06] <paul_c> * paul_c blames B. Gates.
[19:02:29] <Jymmm> Shit bring back OS2
[19:02:32] <SWPadnos> * SWPadnos too
[19:02:55] <SWPadnos> Remember the National Semiconductor NS32032? (in an Acorn PC, I believe)
[19:02:55] <paul_c> Jymmm: What's wrong with running Unix ?
[19:03:32] <A-L-P-H-A> os/2 has been dead for years. :( I used to run it, before win95.
[19:03:35] <Jymmm> paul_c HW support
[19:03:40] <paul_c> SWPadnos: Acorn started off with a 6502, then moved to a RISC chip.
[19:03:42] <Jymmm> A-L-P-H-A it's still alive
[19:03:57] <SWPadnos> Windows doesn't support that much hardware - all the vendors write their own drivers
[19:04:26] <A-L-P-H-A> jymmm, it's dead... you know it.
[19:04:29] <SWPadnos> Hmmm - I remember the 6502 machines, It may be a different company
[19:04:37] <Jymmm> A-L-P-H-A Goony goo goo
[19:04:47] <SWPadnos> I'll have to go through my Byte back issues :)
[19:06:22] <robin_sz> meep?
[19:06:30] <SWPadnos> moop
[19:06:57] <robin_sz> so .. what gives?
[19:07:18] <SWPadnos> the fault line near Indonesia
[19:07:24] <robin_sz> oh yes
[19:07:33] <Jymmm> robin_sz SWPadnos is giving everyone AMD 64 laptops fully loaded
[19:07:48] <SWPadnos> but you just missed the sign-up period
[19:08:54] <SWPadnos> Maybe it was Amstrad
[19:09:39] <robin_sz> my, those amstrads ..
[19:09:44] <robin_sz> thye were not nice
[19:09:58] <SWPadnos> paul - that wasn't a nice response to Jon ;)
[19:10:04] <robin_sz> made $$$ for alan sugar though
[19:10:26] <SWPadnos> do you remember the PC that was based on the National 32032 CPU?
[19:10:40] <paul_c> SWPadnos: He asked a question, and I answered it.
[19:10:52] <SWPadnos> yes - like MS tech support
[19:11:00] <SWPadnos> (You're in a helicopter)
[19:25:05] <anonimasu> hm, os2 was nice
[19:26:27] <robin_sz> sort of
[19:27:01] <anonimasu> nicer then windows will be..
[19:27:04] <anonimasu> will every
[19:27:05] <anonimasu> ever..
[19:27:17] <anonimasu> * anonimasu is on a typo spree
[19:27:28] <Jymmm> Our OS2 servers never went down
[19:27:33] <robin_sz> XP seem fine
[19:27:44] <anonimasu> I am running it and it's slow..
[19:27:45] <anonimasu> :)
[19:27:47] <Jymmm> XP is POSD
[19:27:49] <robin_sz> my windows hatred has relaxed after XP
[19:27:50] <Jymmm> XP is POS
[19:28:03] <anonimasu> well xorg/xf86 is POS also..
[19:28:26] <anonimasu> it might be better with hardware accel..
[19:28:27] <Jymmm> anonimasu ancient is more like it
[19:28:35] <robin_sz> linux desktop stuff is much slower than doze
[19:28:47] <anonimasu> I wish it were more like osx..
[19:28:57] <robin_sz> but I only ever use xterms so what do I care
[19:29:09] <anonimasu> oh, I would like to run it as desktop..
[19:29:11] <anonimasu> :)
[19:29:24] <Jymmm> anonimasu aqua ?
[19:29:37] <anonimasu> Jymmm: yeah but accelerated graphics..
[19:29:46] <robin_sz> * robin_sz spies another cnc that seems to be using the emc interp
[19:30:17] <anonimasu> Jymmm: I want my computer to make me happy when I sit at it.. windows dosent, osx does.. but I dont want a bloody mac.
[19:30:24] <Jymmm> anonimasu I wish aqua was streamlined; I have no need for drpshadows and the like
[19:30:29] <robin_sz> I thnk my windows hatred peaked with win95
[19:30:45] <robin_sz> me + 2k I ignored ...
[19:30:46] <anonimasu> Jymmm: i agree.. but it's way faster then anything on linux yet
[19:30:47] <anonimasu> :)
[19:31:04] <robin_sz> XP .. well, ive not really found fault with it from a suer point of view
[19:31:16] <robin_sz> user
[19:31:16] <Jymmm> I might be temped to move back to NT4SP6
[19:31:24] <robin_sz> ugh
[19:31:24] <SWPadnos> anonimasu: have you run Linux on the same hardware as Aqua?
[19:31:24] <anonimasu> I like 2k..
[19:31:31] <anonimasu> no, I havent..
[19:31:36] <anonimasu> I dont own a mac..
[19:31:44] <SWPadnos> Well then! :)
[19:31:45] <robin_sz> good choice :)
[19:32:00] <SWPadnos> I'd never be able to get used to only one mouse button
[19:32:03] <Jymmm> I have a dual G4 behind me, I should hook it up
[19:32:06] <robin_sz> we're about to upgrade a dozen or so Mac's
[19:32:20] <anonimasu> :)
[19:32:37] <robin_sz> various G3/Gx and iMac boxes
[19:33:07] <robin_sz> got bored of the imacs a LONG time ago
[19:33:21] <robin_sz> now, that was a POS.
[19:33:22] <anonimasu> SWPadnos: but still a p4 3.4 should be able to push X along at a good speed..
[19:33:41] <SWPadnos> reasonably good, but it does depend on how crappy the video card manufacturer is
[19:33:58] <anonimasu> a Radeon 9800
[19:34:04] <SWPadnos> I don't think there's a single modern video card with open source drivers
[19:34:29] <anonimasu> I havent tried with the ATI's own drivers since I need to unpack the .deb
[19:34:32] <SWPadnos> both ATi and Nvidia need closed binary modules for full functionality (if you can even get it)
[19:34:34] <robin_sz> Nvidia ..
[19:34:40] <anonimasu> you can
[19:34:41] <anonimasu> :)
[19:34:43] <robin_sz> so?
[19:35:00] <robin_sz> do the closed binary drivers work?
[19:35:05] <SWPadnos> I have heard that Nvidia's drivers are pretty good.
[19:35:10] <anonimasu> I've never had trouble with them..
[19:35:11] <robin_sz> * robin_sz nods
[19:35:17] <anonimasu> nvidias worked great..
[19:35:19] <robin_sz> well then ..
[19:35:21] <SWPadnos> they may, but what about with a new chipset?
[19:35:25] <anonimasu> but loads of people complained about them...
[19:35:34] <robin_sz> who cares if the driers are open or not .. I dont
[19:35:37] <anonimasu> the same with the ati..
[19:35:38] <Jymmm> http://kerneltrap.org/node/4057?PHPSESSID=6f47b94634d0ec2f7c195e8da2620758
[19:35:45] <anonimasu> loads of people complaining..
[19:35:54] <anonimasu> but people are running them happily also..
[19:36:04] <SWPadnos> I thikn the Nvidia method is to have the core functions in a library, and the system interface is open source, so a lot of the problems are worked around that way
[19:36:05] <Jymmm> anonimasu like they have a choice
[19:36:17] <robin_sz> right
[19:36:25] <SWPadnos> I remember that conversation (note that I took part in it)
[19:36:26] <anonimasu> Jymmm: either way, all drivers have trouble..
[19:36:50] <robin_sz> * robin_sz wonders which version of xf86 he has
[19:36:58] <anonimasu> * anonimasu runs xorg
[19:37:00] <Jymmm> 0.01
[19:37:22] <anonimasu> although I cant remember what differs them
[19:37:23] <anonimasu> :)
[19:37:44] <robin_sz> 4.1.03
[19:38:08] <robin_sz> is that current?
[19:39:05] <anonimasu> 4.4.0 is current..
[19:39:05] <anonimasu> :)
[19:39:16] <robin_sz> close then
[19:40:13] <robin_sz> no, 4.5.0 is current
[19:40:20] <Jymmm> wth... qcad advertises as open source, but when you go to their downloads, the demo runs 10 minutes.
[19:40:39] <robin_sz> and?
[19:40:48] <robin_sz> whats wrong with that?
[19:40:55] <Jymmm> robin_sz read it again
[19:41:00] <robin_sz> I read it
[19:41:11] <robin_sz> you pay them $200 you get a full version
[19:41:17] <robin_sz> I think its $200
[19:41:25] <Jymmm> how can something be open source and then limited at the same time
[19:41:35] <Jymmm> it's liek emc running for 10 minutes
[19:41:38] <robin_sz> you clearl have no clue what open source means
[19:41:39] <anonimasu> not "OPEN SOURCE"
[19:41:41] <anonimasu> as in GNU
[19:41:46] <anonimasu> just open source..
[19:41:56] <robin_sz> even GNU you could do that
[19:42:02] <anonimasu> hm could you?
[19:42:04] <anonimasu> ok
[19:42:13] <robin_sz> easy
[19:42:26] <robin_sz> you put a 10 minute timer in it
[19:42:34] <robin_sz> how hard is that?
[19:42:48] <anonimasu> oh not that, the pay to get the full source code..
[19:42:50] <anonimasu> and the application.
[19:43:01] <robin_sz> correct
[19:43:08] <robin_sz> and they can release it under GPL
[19:44:55] <Jymmm> oh so when you buy a license they give the source?
[19:45:06] <robin_sz> yes
[19:45:09] <robin_sz> its their code
[19:45:15] <Jymmm> which is gpl'ed?
[19:45:22] <SWPadnos> GPL doesn't require you to distribute the source to everybody - only to those who legally obtain your product
[19:45:24] <robin_sz> its is when they release it .. yes
[19:45:31] <robin_sz> correct
[19:45:41] <SWPadnos> and, you can charge for your code as well
[19:45:44] <Jymmm> so anyone can distribute the source?
[19:45:52] <robin_sz> yes ..
[19:46:03] <SWPadnos> but you can't prevent one of your customers from distributing your code to the rest of the world
[19:46:09] <robin_sz> correct
[19:46:15] <SWPadnos> however, GPL and Open Source aren't the same thing
[19:47:21] <alex_joni> greetings
[19:47:32] <SWPadnos> greetings and felicitations
[19:48:02] <SWPadnos> on second thought - scratch the felicitations - just greetings :)
[19:48:56] <alex_joni> heh.. feeling funny?
[19:49:05] <SWPadnos> I'm often funny
[19:49:56] <A-L-P-H-A> :( no mail today... silly Easter Monday, and government stuff. I need my mail, especially my new collet holder.
[19:50:32] <SWPadnos> damned theocracies
[19:54:41] <alex_joni> damn dvdwriter
[19:55:07] <alex_joni> writing at 4x... should have been finished ages ago
[19:55:17] <alex_joni> 66% in 30 mins
[19:55:20] <alex_joni> :((
[19:57:27] <alex_joni> anyways
[19:59:23] <A-L-P-H-A> 4x should only take lik 13 minutes... shouldn't it?
[19:59:34] <alex_joni> 14
[19:59:37] <alex_joni> yes
[20:00:16] <alex_joni> I think the buffer is fragmented (the data-cache on the disk)
[20:00:25] <alex_joni> so it starts/stops
[20:04:46] <alex_joni> interesting mail from paul today ;)
[20:10:06] <p3mm3t> hello
[20:10:17] <p3mm3t> sorry i totaly botched the times i thought i was gonna be available...
[20:11:10] <SWPadnos> 4x DVD? Shouldn't that be closer to 1 hour?
[20:11:27] <alex_joni> SWP: 15 mins tops
[20:12:14] <SWPadnos> isn't it roughly 2G/hour (for playback or 1x)?
[20:12:34] <SWPadnos> or is it the opposite - 9G and 17 hours?
[20:12:53] <alex_joni> no ideea :)
[20:14:17] <Jymmm> A-L-P-H-A since you mentioned it, there is BlenderCNC btw
[20:14:33] <A-L-P-H-A> oh, really?
[20:14:38] <A-L-P-H-A> in pre-alpha stages no doubt
[20:14:45] <Jymmm> nfc
[20:14:53] <alex_joni> who knows about exim & spamassassin on sarge?
[20:20:01] <Jymmm> A-L-P-H-A http://projects.blender.org/projects/blendercad/
[20:22:30] <A-L-P-H-A> dat ain't blendercnc. :/
[20:22:32] <cradek> I know about SA, but not exim or debian
[20:22:51] <alex_joni> I think I'll integrate it with procmailrc
[20:23:07] <alex_joni> I have spamd installed
[20:23:18] <alex_joni> oh.. btw, hello cradek
[20:27:46] <alex_joni> does that sound sane cradek?
[20:27:58] <alex_joni> http://www.stearns.org/doc/spamassassin-setup.current.html#sitewide
[20:28:20] <cradek> hello
[20:28:47] <cradek> if you're running spamd and using spamc inside procmailrc, exim doesn't need to know
[20:29:11] <cradek> that's definitely the simplest setup.
[20:29:30] <alex_joni> ok.. that's what I thought (from what I've been reding the last 10 mins)
[20:29:50] <alex_joni> afaict debian installed spamd
[20:30:05] <cradek> yeah procmail just pipes the message through spamassassin, which rewrites it (adding headers and stuff).
[20:30:16] <alex_joni> so all I need is do the magic in procmailrc (to filter the mail through spamc)
[20:30:19] <cradek> err, more specifically, through spamc
[20:30:25] <cradek> :0fw
[20:30:25] <cradek> |spamc
[20:30:26] <alex_joni> right
[20:30:59] <alex_joni> ok, and after that I can move tagged mail somewhere else?
[20:31:14] <cradek> sure
[20:31:25] <alex_joni> also through procmailrc ?
[20:31:29] <cradek> yes
[20:31:41] <cradek> in the next rule, check spamassassin's headers to decide what to do.
[20:32:38] <alex_joni> :0:
[20:32:38] <alex_joni> * ^X-Spam-Status: Yes
[20:32:38] <alex_joni> spamassassin-spam
[20:32:41] <alex_joni> ?
[20:33:44] <cradek> yeah something like that
[20:33:56] <alex_joni> ok.. will test tomorrow
[20:34:01] <alex_joni> thx
[20:34:09] <alex_joni> how's the pcb-milling?
[20:34:20] <cradek> just fine
[20:34:25] <alex_joni> coo
[20:34:34] <alex_joni> did yopu talk to les?
[20:34:35] <cradek> had a small mishap, but it'll be easy to fix up
[20:34:37] <alex_joni> you
[20:34:38] <cradek> about what?
[20:34:46] <alex_joni> segmentqueue
[20:34:55] <cradek> not for a while
[20:34:58] <cradek> what about it?
[20:35:14] <alex_joni> he said (today) some parts are still missing from a fully 100% ok SQ
[20:35:29] <cradek> that's not new news to me
[20:35:40] <alex_joni> well.. he said that SQ sometimes behaves very smoothly, but doesn't do the right speed
[20:35:46] <cradek> new news would be someone working on it
[20:36:02] <alex_joni> heh
[20:36:06] <cradek> yeah I know it has several bugs remaining
[20:36:12] <cradek> I put some in the tracker.
[20:49:55] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[20:50:14] <p3mm3t> p3mm3t is now known as pemmet
[20:52:41] <robin_sz> coo
[20:52:47] <robin_sz> the net fell apart
[20:52:53] <Jymmm> SSDD
[20:54:37] <robin_sz> ah well
[20:57:19] <dmwaters> {global notice} Hi all! one of our main rotation servers just kernel paniced. The box is on it's way back up, when it's back up, I'll give it the fix, and reboot it again. Thank you all for your patience, and thank you for using freenode!
[20:57:25] <alex_joni> hey robin
[21:01:33] <alex_joni> cradek: do I need to set upa a mbox to use procmail?
[21:02:39] <cradek> alex_joni: I think procmail can deliver to several mailbox types
[21:04:53] <dmwaters> {server notice} Alright boys and girls, I'm gonna reboot this box to fix the problem that caused the kernel to panic, sorry about this.
[21:07:53] <alex_joni> robin: kernel panic ;)
[21:13:45] <les> cradek you there?
[21:14:31] <les> heh
[21:15:45] <les> anyway if you see this
[21:16:09] <les> I was checking your diffs in segmenqueue in preparation for testing
[21:17:03] <cradek> nope
[21:17:56] <les> also hunting for the equation 4.4.31 in the code
[21:18:02] <les> ha
[21:18:17] <les> that arbitrary value x
[21:18:25] <les> currently assigned 5
[21:18:54] <les> I think is a "tightness" factor for the cubic
[21:19:05] <les> may want to play around with that
[21:19:09] <les> but must find it
[21:20:14] <les> It could also greatly affect velocity adaptation
[21:20:35] <les> which prob is still a problem
[21:20:45] <les> I will look some more
[21:23:56] <cradek> see s->p in line 590
[21:24:07] <cradek> also line 629
[21:26:48] <cradek> no, maybe that's not it
[21:27:05] <cradek> yes, I think it is
[21:27:16] <cradek> number of cruising steps
[21:28:18] <alex_joni> right.. night guys
[21:28:22] <les> ok i'll check that out
[21:28:32] <alex_joni> thx again cradek
[21:28:39] <cradek> alex_joni: welcome
[21:28:56] <A-L-P-H-A> later aj
[21:29:09] <alex_joni> bye les
[21:29:16] <les> bye alex
[21:29:18] <alex_joni> ciao alpha
[21:29:24] <A-L-P-H-A> adious
[21:29:28] <A-L-P-H-A> audious?
[21:29:38] <les> adeos?
[21:29:47] <les> no adios
[21:29:49] <A-L-P-H-A> ahh, c'est oui.
[21:29:49] <cradek> odd-youse?
[21:29:53] <les> heh
[21:30:06] <A-L-P-H-A> sianarah?
[21:30:14] <A-L-P-H-A> bye
[21:30:45] <A-L-P-H-A> joy-gen.
[21:30:56] <A-L-P-H-A> jai-jang.
[21:47:43] <Jymmm> to mill something 2"x2"x24" at an angle, are you guys making a jig for that?
[21:47:53] <Jymmm> lengthwise
[21:48:34] <SWPadnos> with a CNC, it shouldn't be necessary (unless you consider clamps a jig)
[21:49:18] <Jymmm> This sin't with CNC, but lets say it is... how could you mill at a 30 degree angle?
[21:50:21] <Jymmm> isn't
[21:50:34] <SWPadnos> You don't have to mill at the specific angle with a CNC - you use a ball end mill and do multiple passes in the appropriate direction
[21:51:04] <Jymmm> ok, and without CNC?
[21:51:05] <SWPadnos> like slices in an architectural landscape :)
[21:51:24] <SWPadnos> you use an angle vise (sine vise, I think they're called)
[21:51:36] <Jymmm> it's 24" long
[21:51:59] <SWPadnos> You can also use angle blocks, and clamp the part to the blocks
[21:52:15] <SWPadnos> milling something 24" long is trouble anyway, even on a Bridgeport
[21:52:21] <SWPadnos> you need to do multiple passes
[21:52:31] <paul_c> clamp it down flat on top of a sacrificial plate....
[21:52:33] <SWPadnos> (unless you know whet you're doing, which I don't)
[21:52:46] <paul_c> and use a 60 degree cutter
[21:53:07] <SWPadnos> well - that's the easy way ;)
[21:53:26] <paul_c> then there's the hard way...
[21:53:36] <SWPadnos> tilt the head :)
[21:53:41] <paul_c> clamp in vice, and use a file.
[21:53:42] <Jymmm> ew
[21:55:09] <Jymmm> paul_c when you say cutter, what are you talking about exactly? a face mill of some sort?
[21:55:33] <paul_c> Are you using a vertical mill ?
[21:55:52] <Jymmm> paul_c yep
[21:56:50] <paul_c> lemme find you a pretty picture
[21:56:59] <Jymmm> cool beans
[21:58:50] <Jymmm> like this? http://www.microcutusa.com/dblangle.html
[21:59:31] <paul_c> not quite
[22:02:04] <robin_sz> not still trying to make your own rail from solid eh Jymmm
[22:02:27] <Jymmm> robin_sz, yep
[22:03:11] <robin_sz> but hiwin is so cheap
[22:03:13] <paul_c> http://www.microcutusa.com/chamfersng.html - Bit small...
[22:04:04] <robin_sz> oh joy, that reminds me .. that silly sod who bought our press should be taking it this week :)
[22:04:17] <robin_sz> s/press/punch/
[22:04:23] <Jymmm> robin_sz didn't we talk about this? Your_cheap / 10 == My_Cheap
[22:04:44] <robin_sz> shrug
[22:04:46] <paul_c> http://www1.mscdirect.com/CGI/NNSRIT?PARTPG=NNLMK3&PMPXNO=4760653 - bigger version
[22:05:10] <robin_sz> I suspect the cutters cost more than the rail ...
[22:05:36] <robin_sz> and anyway, milling it is unlikely to be accurate enought to make useable rail I suspect
[22:05:39] <Jymmm> robin_sz: That's ok though; I'm not replacing the cutter THAT often
[22:05:51] <Jymmm> paul_c that thing just LOOKS scarry! lol TY
[22:06:06] <Jymmm> * Jymmm watches way too many movies sometimes =)
[22:06:30] <robin_sz> deffo more than the rail
[22:06:57] <robin_sz> hiwin rail, in 15mm is like .. 100 bucks for 1.5m
[22:07:00] <Jymmm> robin_sz it's parts not tooling that I'm being cheap about.
[22:07:11] <robin_sz> either way, they cost
[22:08:18] <Jymmm> ok, that cutter cost $300. For $300 I can't even buy enough railing for one cnc router. If I make my own, I should be able to get at least 5 routers worth of railing.
[22:09:50] <Jymmm> just curious, can cutters like that be resharmed, or is that usually a bad idea?
[22:09:52] <robin_sz> hahaha
[22:09:53] <gezr> howdy yall
[22:10:07] <robin_sz> that cutter has replaceable tungsten inserts
[22:10:12] <robin_sz> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=633&item=7504145917&rd=1
[22:10:23] <Jymmm> oh cool
[22:10:31] <Jymmm> robin_sz, ok so why you laughing at me?
[22:10:42] <robin_sz> because you will not make useable rail
[22:11:11] <gezr> gonna mill square slide rail i guess?
[22:11:47] <robin_sz> nah, someting with 60 degree chamfers
[22:12:29] <gezr> you mean a < sitting ontop of another < to make a rail and saddle that looks like <<-->> that?
[22:12:48] <robin_sz> no idea
[22:13:58] <robin_sz> rail is like 60 bucks/metre from hiwin, precision ground. cariages are 40 bucks ...
[22:15:11] <SWPadnos> this is the $100 version: http://www1.mscdirect.com/CGI/NNSRIT?PARTPG=NNLMK32&PMPXNO=1960800
[22:16:01] <Jymmm> SWPadnos not as large of cutters, is it?
[22:16:20] <SWPadnos> perhaps not - I didn't check that
[22:16:36] <Jymmm> SWPadnos it's cool; just difficult to tell from img.
[22:17:39] <robin_sz> and this is the $50 version
[22:17:42] <robin_sz> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=25284&item=7504196474&rd=1
[22:18:09] <Jymmm> somebody gimme op's so I can kick robin_sz please!
[22:18:30] <SWPadnos> one thing I've learned - Jymmm is trying to design a cnc router that can be built for <$500 - in quantities larger than one
[22:18:46] <SWPadnos> He's not trying to build one router for <$500
[22:19:04] <Jymmm> what SWPadnos said
[22:19:12] <robin_sz> well, been there .. tried it with my plasmas
[22:19:49] <SWPadnos> I think we should all get a cut of the profits :D
[22:20:23] <robin_sz> if he could do basic maths hed see its not worth trying to make rail ... the IGUS stuff for instance ...
[22:20:35] <robin_sz> buy it, cut it, bolt it on.
[22:20:40] <robin_sz> job done.
[22:20:50] <robin_sz> try making your own rail, feel free.
[22:20:51] <gezr> Jymmm : how long a rail do you plan on milling?
[22:21:04] <SWPadnos> that's certainly better if you count labor in the budget :)
[22:21:06] <robin_sz> but it will cost far more in time than the igus stuff costs
[22:21:06] <Jymmm> gezr max 48"
[22:21:23] <gezr> Jymmm : whats the size of the milling machine you have at your use, and how many vises does it have?
[22:21:28] <robin_sz> you HAVE to count labor .. that was my BIG mistake
[22:21:43] <Jymmm> gezr HF mini mill
[22:22:06] <robin_sz> Its *exactly* what I tried to do on my mk1 and 2 machines ... rail from cold-rolled bar
[22:22:10] <Jymmm> robin_sz curious, do you count cnc mill time as labor time?
[22:22:17] <robin_sz> certainly
[22:22:24] <SWPadnos> machine time has a cost and therefore a price attached to it
[22:22:27] <Jymmm> robin_sz excluding setup tiem
[22:22:37] <robin_sz> oh, ok, come and work for me
[22:22:51] <robin_sz> you will set my machiens up for free?
[22:22:54] <SWPadnos> yes - most shops charge from $150 to $250 for machine time (could be more these days)
[22:22:57] <gezr> setup, run, teardown are all quoted
[22:23:01] <robin_sz> exactly
[22:23:13] <Jymmm> I meant actual run time is what I was saying
[22:23:17] <SWPadnos> yes
[22:23:23] <Jymmm> of course you have to charge for setup time
[22:23:37] <SWPadnos> realize that there are costs involved:
[22:23:39] <SWPadnos> electricity
[22:23:43] <SWPadnos> lubricants / coolants
[22:23:46] <SWPadnos> cutters
[22:23:49] <gezr> a nice job shop will run somehting for anywhere from 50 an hour up, depends on hunger, that hourly rate includes setup run programing teardown,
[22:23:53] <SWPadnos> machine wear and tear
[22:24:01] <SWPadnos> machine purchase amortization
[22:24:05] <SWPadnos> floor space
[22:24:06] <SWPadnos> etc.
[22:24:06] <Jymmm> SWPadnos is that typically at the same (or close) rate as labor though?
[22:24:17] <gezr> shop gets paid wether that thing is cutting chips or not
[22:24:22] <SWPadnos> no - a machinist might be $50/hour, but a CNC is 5x as much
[22:24:35] <robin_sz> Jymmm: heres a simple question: would you rather build a $500 router that you have to spend days making all the parts, or a $800 router that you just take the bits of the shelf and assemble? sometimes it is worth spending more to make your life 10x easier
[22:25:12] <gezr> SWPadnos : its sorta different then that, but all expenses are calculated into a "shop rate"
[22:25:21] <SWPadnos> I agree with robin except in one case - if you can't sell it for more than $500, then it's of utmost importance to keep the cost down
[22:25:53] <robin_sz> well, still, buy off the shelf
[22:26:03] <SWPadnos> sure - whenever possible
[22:26:10] <robin_sz> drawer runners?
[22:26:14] <Jymmm> robin_sz: If is was a 20% margin more to build a better/faster/easier machine sure, but the numbers are really not there for that. If I can get this design to work, I can reproduce them quickly.
[22:26:33] <robin_sz> no, you'll die under the load of making the parts
[22:26:50] <robin_sz> your margin is only 20%
[22:26:52] <robin_sz> ??
[22:27:05] <Jymmm> robin_sz $500 + 20% (my cost)
[22:27:11] <robin_sz> yore fucked then
[22:27:20] <robin_sz> take my advice. quit now
[22:27:23] <Jymmm> I'm not SELLING them for that,
[22:27:38] <robin_sz> thank god
[22:27:52] <robin_sz> 3 to 5 x cost
[22:28:02] <robin_sz> or you'll be bankrupt
[22:28:24] <Jymmm> robin_sz I said I'm willing to spend $500 up to 20% more for easier/faster
[22:28:54] <gezr> Jymmm : I would be very hesitant to mill a 48" long rail on a full size bridgeport and expect a good clean surface, I did clean up a 48" southbend lathe bed on a mill once, but that was just dressing the edges, and that operation took an entire day, and I still screwed it up, I get to do the saddle again.
[22:29:13] <robin_sz> * robin_sz nods
[22:30:12] <robin_sz> Jymmm: stick to simple, off the shelf, repeatable parts. custom made stuff will not beat IGUS on price .. thats stuff is dirt cheap.
[22:30:42] <Jymmm> robin_sz divide by 10
[22:30:50] <gezr> im not trying to be mean/rude or anything of the sort, I was just cleaning up a .010 dip in the tables surface, and I only had to move the part once, and re-indicate, I should have hired a shop to use a big surface grinder to do the work for me,
[22:31:48] <robin_sz> Jymmm: you are not costing your time
[22:31:51] <Jymmm> gezr: Not at all =) I'm not actually milling a 48" piece of stock
[22:32:03] <gezr> Jymmm : I agree with you, if you have all the time to work on things then have at it, I dont have that sort of time, and my time is expensive
[22:32:05] <Jymmm> robin_sz That is true.
[22:32:31] <robin_sz> Jymmm: how much do you pay your employees?
[22:32:40] <robin_sz> per hour ..
[22:32:47] <Jymmm> robin_sz: But here's the thing... I need to get started somewhere. And if right now I have to give up sleep, so be it.
[22:33:00] <robin_sz> no no no ...
[22:33:09] <robin_sz> you.re making the same misatke I did
[22:33:16] <robin_sz> it cost me THOUSANDS
[22:33:18] <robin_sz> no ..
[22:33:22] <robin_sz> tens of thoudands
[22:33:27] <robin_sz> look,
[22:33:31] <robin_sz> there just you right?
[22:33:42] <robin_sz> your time is VALUABLE
[22:34:00] <robin_sz> FAR too valauble to be making stuff you can buy
[22:34:01] <gezr> If I fart I charge myself
[22:34:25] <robin_sz> you will end up a busy fool. seriously, this is what I did.
[22:34:27] <SWPadnos> I would too
[22:34:35] <robin_sz> "oh I can just drill and tap this"
[22:34:45] <robin_sz> blah ..
[22:35:09] <robin_sz> work on the design, make it simple quick and cheap
[22:35:14] <robin_sz> but LOW labour
[22:35:19] <robin_sz> labour is the killer
[22:35:56] <Jymmm> robin_sz: Ok, here's the bottom line..... I am going to attempt to compete with this guy with something better than what he's offering and anticipating getting more for what he's getting too (on avg he's been getting $1100) http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=57122&item=7503826197&rd=1
[22:36:43] <Jymmm> But I need to get my cost LOW. I calculated his costs less mechancial to be $300
[22:37:09] <Jymmm> driver, motors, router, PS
[22:38:07] <gezr> your double his work envelope, there is no X2 factor when talking machine size, its exponential cost for size growth
[22:38:11] <paul_c> The guy is taking the piss isn't he ??
[22:38:28] <robin_sz> hes probably making them at a loss
[22:38:43] <robin_sz> or close to a loss
[22:38:45] <paul_c> wot... Out of chipboard ?
[22:38:46] <Jymmm> paul_c yeah he is, he start the auction at $375 which make me think that's his cost.
[22:39:07] <gezr> he isnt including his time either
[22:39:13] <robin_sz> quite
[22:39:19] <gezr> oh my gowd,
[22:39:31] <gezr> it looks clean,
[22:39:33] <SWPadnos> well, all I can say is I would never sell that machine
[22:39:34] <Jymmm> no he's not, but he's sold every one of them too.
[22:39:42] <gezr> Jymmm : you are double his size, your cost should at least be double his
[22:39:54] <SWPadnos> closer to triple
[22:40:01] <robin_sz> yeah
[22:40:07] <gezr> accually its squared if not more
[22:40:17] <SWPadnos> yeah - I was just thinking that
[22:40:20] <Jymmm> gezr: Well, the 48" is for personal use; will try for 18" x 25" or somethign for sell
[22:40:25] <robin_sz> so ... $500 is not a reallistic budget
[22:40:49] <gezr> Jymmm : then make your machine the best it possiable be, so it can be used to crank out the smaller ones
[22:41:19] <Jymmm> gezr the plan is the personal machine dbl as prototype
[22:41:20] <robin_sz> I figure my MINIMUM costs for a 4x4 plasma to be around $2k ... and thats not including time
[22:42:19] <gezr> wow, if that thing gets wet, its gonna grow a lot
[22:42:26] <robin_sz> heh :)
[22:42:28] <Jymmm> gezr lol, yep
[22:42:34] <SWPadnos> note that he says nothing about accuracy
[22:42:42] <SWPadnos> he mentions resolution, but not accuracy
[22:42:54] <robin_sz> Jymmm: did AXYZ sell any machines this month?
[22:43:12] <Jymmm> he also says ball bearing slides too
[22:43:19] <Jymmm> robin_sz who?
[22:43:47] <SWPadnos> a drawer slide is also ball bearing
[22:43:51] <robin_sz> http://www.axyz.com/table_mill.html
[22:44:12] <robin_sz> SWPadnos: drawer slides woudl have been a suggestion for a smaller mill, but 48" .. probably not
[22:44:28] <gezr> Jymmm : who knows, you got a plan stick with it.
[22:44:36] <SWPadnos> I was only pointing out that the description in that auction is compatible with that machine being made with drawer slides
[22:44:43] <robin_sz> yep
[22:44:48] <robin_sz> I think it probably is
[22:44:57] <SWPadnos> and plastic screw blocks
[22:45:02] <robin_sz> yep
[22:45:12] <gezr> I have a drawer slide axis built, that cost me a lot to get just to that stage
[22:45:18] <robin_sz> heres the thing,
[22:45:35] <robin_sz> I thnk you cold build the same thing and do it slightly better for slightly more
[22:45:58] <robin_sz> say, beeter wood, and better slides for $500 isntead of $375
[22:46:21] <Jymmm> ok you made me pull out the dman calculator... $500 + 20% == $600
[22:46:31] <SWPadnos> I could have told you that
[22:46:54] <Jymmm> * Jymmm smacks SWPadnos then why didn't ya?! =)
[22:47:07] <robin_sz> because it was too obvious to even mention
[22:47:11] <gezr> and to think im fretting over the cost of a stg card in the future
[22:47:20] <Jymmm> * Jymmm is drain bamaaaged today (taxes)
[22:47:21] <SWPadnos> you didn't ask :)
[22:47:31] <SWPadnos> aaaaaaaahhhhhhhhhhhhhh taxes
[22:47:39] <robin_sz> look, you probably think its a great plan
[22:47:40] <gezr> I need to sign and mail mine off hmm
[22:47:45] <robin_sz> but ...
[22:47:48] <robin_sz> heres the ting
[22:48:12] <robin_sz> that machine is cheap, but crap
[22:48:19] <robin_sz> hes probably making close to zero profit
[22:48:28] <robin_sz> and its probably hard to build
[22:49:07] <robin_sz> a cheap, but well built 48x48 woudl sell well, be simple to build with shop bought components and make 5x the profit
[22:49:20] <robin_sz> liek, say, a shopbot
[22:49:25] <robin_sz> seen them?
[22:49:36] <Jymmm> kinda a 3-in-1 one machien?
[22:49:39] <robin_sz> yeah
[22:49:45] <Jymmm> ok, yeah
[22:50:03] <robin_sz> http://www.shopbottools.com/
[22:50:12] <Jymmm> robin_sz: Here's the deal...
[22:51:42] <Jymmm> I'm in a apartment and dont' have room to build something large at this time. I can't justify renting space in a small industrial area till I least can justify it (with proven sales). If I start making these lil ones I get the word out and as more ppl learn about them I can get to the point of saying I can now justify the space.
[22:51:55] <robin_sz> right
[22:52:00] <robin_sz> exactly how I started out
[22:52:24] <robin_sz> 2 years and about $80K later I have learnt a bit
[22:52:45] <Jymmm> brb... coffee
[22:52:54] <robin_sz> so build a small one, but better
[22:53:15] <robin_sz> $1100 is NOT what the market will max out at either
[22:58:16] <gezr> Ill be back later on guys, I have hously things to take care of
[22:58:45] <Jymmm> cya gezr
[23:01:05] <Jymmm> robin_sz I think one of *MY* drawbacks is the lask of parts locally. I want to make my first one good enough for me to use, but I need to be very cost effective. Primarily mixing what I'll sell with what I'll use, improving the disgn of what I sell along the way. It's like a software company using their own products. (so to speak)
[23:01:45] <robin_sz> OK,
[23:01:52] <robin_sz> lets put that into context
[23:02:02] <robin_sz> you are a small one man software house
[23:02:19] <Jymmm> If I buy everything mail order, It's much more difficult to say "hey, lets try this out and see what happens" I now have to seriously THINK before I buy or I'll run out of budget
[23:02:40] <robin_sz> you are telling me how the best way to proceed is to start by writing your own OS and compiler
[23:02:55] <robin_sz> Jymmm: one thing will save you ... 3d CAD
[23:03:13] <Jymmm> robin_sz I'm working on the CAD as we speak =)
[23:03:25] <Jymmm> i fired up the mac
[23:04:26] <Jymmm> as in finding some CAD program to use/learn I mean.
[23:04:53] <Jymmm> (the mac has dual G4 + 2GB ram)
[23:06:49] <Jymmm> robin_sz I was using the software analogy as an example. I use to work at Netscape (pre-aol). Some employees were using IE/Mosaic/etc at the time. The hammer dropped and everyone was told to use Netscape. Made sense to me; easier to find bugs and the like.
[23:07:18] <p3mm3t> p3mm3t is now known as pemmet
[23:10:12] <Jymmm> robin_sz Please believe me when I saw that I am hearing what you're saying, I truely am, I'm just not there yet is all. I can try the design I came up with for under $100, if it works GREAT! If not, there is always IGUS which is in LA.
[23:10:25] <Jymmm> s/saw/say/
[23:19:47] <robin_sz> Jymmm: fair enough, I am just truing to save you from yourself
[23:20:47] <robin_sz> that machine will take, minimum 10 hours build time, minimum
[23:21:00] <robin_sz> probably closer to 20 if you make some parts
[23:22:11] <robin_sz> a 48" .. well, its going to take even longer
[23:22:15] <Jymmm> robin_sz: I really do appreciate it, carry on!!! I think I just need to try (at least on paper) first then take it from there.
[23:22:27] <robin_sz> absolutely
[23:22:59] <Jymmm> I want to use MDF as the base material, that should save some cash
[23:23:08] <Jymmm> dry adn seal it
[23:23:17] <robin_sz> MDF is more expensive than chipboard
[23:23:21] <robin_sz> but a good choice
[23:23:33] <Jymmm> full sheet of 3/4" is $22
[23:23:33] <robin_sz> I used 40mm on mine
[23:23:44] <Jymmm> and that's 49" x 96"
[23:24:03] <Jymmm> not a typo... 49" not 48"
[23:24:07] <Jymmm> wierd
[23:24:14] <robin_sz> 1250mm
[23:24:35] <Jymmm> ah
[23:26:51] <acemi> in debian, I install kernel-source and I patch it with adeos patch. I install rtai-source (from vulcano branch) to /usr/src/modules/rtai/ . I can create kernel-image using "make-kpkg kernel-image" but I can't create rtai module deb package using "make-kpkg module-image". What must I do to create rtai deb package?
[23:47:42] <paul_c> acemi: You need the debian/rules
[23:48:04] <acemi> debian/rules?
[23:48:22] <acemi> i don't understand
[23:49:33] <paul_c> in the modules/rtai directory, do you have a debian/ sub-dir ?
[23:50:16] <acemi> no
[23:50:32] <acemi> how can I prepare this
[23:52:13] <paul_c> grab the rtai package from Debian's Sid repository - You should find a usable debian/rules in there.
[23:52:56] <acemi> can I use rtai-source form sarge repository?
[23:53:02] <acemi> from
[23:53:59] <acemi> I'll try. thanks paul_c
[23:54:13] <robin_sz> does Sarge have an RTAI kernel?
[23:54:25] <robin_sz> * robin_sz had never looked.
[23:55:12] <paul_c> unstable is 3.1.0-3
[23:55:17] <acemi> no RTAI kernel but there are kernel_patch_adoes and rtai-source package
[23:55:32] <paul_c> dunno if the rules have changed...
[23:55:42] <robin_sz> right
[23:55:45] <robin_sz> interesting