irclog2html for #brlcad on 20061105

02:26.52 CIA-6 BRL-CAD: 03johnranderson * 10brlcad/src/mged/red.c: Improved error checking
04:43.31 *** join/#brlcad Noobie (n=User@pool-71-240-238-151.fred.east.verizon.net)
04:44.42 Noobie can anyone get me past the ./configure portion of the install
04:45.17 Noobie i don't know if I can even get there i'm using Knoppix as my OS
04:53.13 Noobie can anyone help?
04:53.43 Noobie I will try again later
04:53.48 *** part/#brlcad Noobie (n=User@pool-71-240-238-151.fred.east.verizon.net)
05:27.30 ``Erik o.O
07:37.18 *** join/#brlcad clock_ (i=clock@84-72-92-39.dclient.hispeed.ch)
07:43.20 *** join/#brlcad clock_ (i=clock@84-72-92-39.dclient.hispeed.ch)
08:01.45 *** join/#brlcad ibot (i=ibot@pdpc/supporter/active/TimRiker/bot/apt)
08:01.45 *** topic/#brlcad is http://ftp.brlcad.org (mirror of http://brlcad.org, down for 'maintenance') || BRL-CAD is an open source solid modeling software suite || Developers needed! Read the HACKING file for details on getting involved || Release 7.8.4 is posted, next release will be 7.10.0
08:35.23 *** join/#brlcad ibot (i=ibot@rikers.org)
08:35.23 *** topic/#brlcad is http://ftp.brlcad.org (mirror of http://brlcad.org, down for 'maintenance') || BRL-CAD is an open source solid modeling software suite || Developers needed! Read the HACKING file for details on getting involved || Release 7.8.4 is posted, next release will be 7.10.0
13:40.38 Maloeran Erik, this ray-triangle benchmark of yours is completely broken
13:41.12 Maloeran I print out planes from the test, and I get a bunch of 0.0 equations, leading to infinities and NaNs ; obviously it's slow
14:17.50 Twingy he's ray-tracing imaginary planes, duh
14:27.06 Maloeran I would fix the code but that's just one of several things that don't seem quite right in there...
14:33.47 Twingy consider yourself a code archaeologist then, fixing remnants of the past
14:34.45 Twingy I'd say similar to the myth busters, but they have a habit of taking 500 year old technology and making it worse
15:41.25 Maloeran Hrmph. The public letter of Husseim to the united states contain way too many religious believes and statements, which is not technically a good move when dealing with followers of different... hypothesises
15:55.35 Twingy hypotheses
16:27.23 Maloeran Thanks
16:28.12 Twingy I have soooo many optimizations to do in gcam it's daunting
16:28.19 Twingy I have no motivation to do it either cause just about everything is instantaneous
16:28.39 Maloeran Eheh, nice. It has been some time since I had wrote this kind of software
16:29.00 Twingy all 2d intersection math so far
16:29.00 Twingy 2.5 dimensional
16:29.19 Twingy after getting this far with the project, you could transform just about any CAD package file format into a CAM ready file format if ordering is preserved
16:29.19 Maloeran Neat
16:29.29 Twingy but nobody seems to do that
16:29.37 Twingy order is not preserved and hence individuals have to come up with elaborate algorithms to figure out what the designer was thinking at the time of modeling because it was lost
16:29.41 Twingy I wouldn't be surprised if there are some cad packages that preserve this
16:29.42 Twingy preserved in a manner that was originally intended, not derived from a timestamp or something that just "happened" to be there
16:29.42 Twingy the disconnect between BRL-CAD and CAM is huge
16:29.42 Twingy you'd need g-code
16:29.42 Twingy even something as 'trivial' as a box requires breaking it into layers, computing offsets, adding depth to break the part away at the end
16:29.43 Twingy you might end up with 500 lines of g-code for a box
16:29.43 Twingy especially if it's steel
16:29.43 Twingy foam might be 50 - 80 lines
16:29.43 Maloeran Why the difference between steel and foam?
16:30.23 Twingy steel is slightly less hard than the end-mill (carbide usually) and therefore requires that you be very slow and conservative on your cuts otherwise you risk causing the stepper motors to skip a 'step' (1.8/20 deg) or cause and end-mill to snap
16:30.25 Twingy it's always a function of the end-mill, mill cnc steppers, and material being cut
16:30.25 Maloeran Okay. I wouldn't have thought going very slow would require 10 times more code
16:30.27 Twingy it's a multidimensional problem
16:30.27 Twingy and smaller the end mill doesn't necessarily mean you can cut faster
16:30.28 Twingy there is a curve, once you hit about 1/8" you have to start slowing down
16:30.28 Twingy otherwise you will snap it
16:30.28 Twingy Maloeran, your layers are .001" instead of .01"
16:30.29 Twingy each layers requires 4 or more lines of g-code, depending on tapering
16:30.29 Maloeran Right, okay
16:30.30 Twingy so as you can imagine you have machine profiles and endmill profiles
16:30.31 Twingy so that the software can figure out 'safe' operating limits
16:30.31 Twingy the g-code driving software is stupid
16:30.31 Twingy it doesn't care if you try to make it go 100km/h with a 2cm diameter endmill
16:30.32 Twingy it doesn't even know what the tool diameter is
16:30.32 Maloeran Well, it's low level and does what it's told I am assuming ; which is a good thing
16:30.33 Twingy yep
16:30.33 Twingy now emc has tried to bridge that gap
16:30.33 Twingy and make the driver smarter
16:30.34 Twingy a little too smart in my opinion
16:30.34 Maloeran You probably lose some flexibility that way
16:30.34 Twingy well, if you have a good cam package, all you need is a stupid driver
16:30.35 Twingy if you have a crappy cam package, or you do g-code by hand (gasp) you need a smart driver
16:30.35 Twingy gcam is like a g-code compiler
16:30.35 Twingy by pure definition, it's a compiler
16:30.36 Maloeran Eheh yes, a visual compiler in some ways too
16:30.36 Twingy right
16:31.14 Twingy with like 8 asm instructions :)
16:31.23 Twingy very analogous to coding an 8-bit pic chip
16:31.49 Twingy doing floating point math (your csg cube) can be done, but it requires many instructions
16:37.17 Twingy I suppose at some point one will be able to measure the efficiency of their code by the number of electrons required to solve the problem
16:41.37 Maloeran I don'T think it is possible to compute the minimal electron count required to implement an algorithm within a context of hardware processing
16:59.57 Twingy today's current hardware, no
17:00.48 Twingy a practical measurement for today would be something around nano watts
17:09.56 *** join/#brlcad Noobie (n=User@pool-71-240-238-151.fred.east.verizon.net)
17:10.23 Noobie is anyone available?
17:10.46 Noobie I need some help installing brlcad
17:11.06 Noobie i'm using Knoppix as my OS so I'm not even sure if it is possible
17:22.37 Twingy just idle until you find somone
17:24.29 Noobie will do
18:30.19 Noobie gotta go... out
18:30.25 *** part/#brlcad Noobie (n=User@pool-71-240-238-151.fred.east.verizon.net)
23:08.11 ``Erik me too, when I play games.. but I did that before with my c64, heh :D
23:09.06 Twingy I envision in like 20 years this disturbing looking cable with like 17 adapters on it
23:09.21 Twingy sticking 3 feet out of the back of the computer
23:09.27 Twingy angled 30 degrees
23:09.35 Twingy held in with duct tape
23:10.10 Twingy whose combined processing power is 1 week of a C64 just to send a character code
23:13.23 ``Erik hehehe
23:13.43 ``Erik I have new ps2 kbd's, but I like my old mitsumi precision din kbd a lot more *shrug*
23:14.12 ``Erik tho I do like the apple clear&white usb kbd's a lot
23:17.32 Twingy yes, they are good
23:17.43 Twingy I'm just amazed this natural touch has lasted under such abuse
23:38.57 *** join/#brlcad test34 (n=test34@unaffiliated/test34)

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with blootbot logs, split per channel, etc.