00:00.09 |
*** join/#brlcad BigAToo
(n=BigAToo@pool-96-230-124-61.sbndin.btas.verizon.net) |
00:00.09 |
Ralith |
a model of one of those would be pretty neat
too. |
01:25.36 |
*** join/#brlcad JoanW
(n=5b698faa@bz.bzflag.bz) |
01:47.04 |
*** join/#brlcad samrose
(n=samrose@oh-69-69-33-40.sta.embarqhsd.net) |
02:11.19 |
*** join/#brlcad dreeves
(n=dreeves@67.130.253.14) |
02:13.08 |
dreeves |
I sent the patch to you brlcad I will work on
pushing it to sf but I was having problems so decided to be lazy
for now |
02:13.18 |
dreeves |
and send it directly to you |
02:23.27 |
dreeves |
ok got it submitted to sourceforge |
02:34.35 |
brlcad |
nice, just saw it |
02:35.09 |
brlcad |
didn't get anything direct though (at least
not yet |
02:35.29 |
brlcad |
but no matter, I see the tracker |
02:44.02 |
*** join/#brlcad dreeves_
(n=IceChat7@67.130.253.14) |
03:02.40 |
starseeker |
<snort> they're already offering
discounts to try and get people to come to SIGGRAPH 2009 |
03:03.11 |
starseeker |
<sarcasm>wonder if
perhaps they're worried about the
economy...</sarcasm> |
03:13.54 |
dreeves |
Sorry about that I guess I should have spent a
little more effort figuring it out |
03:13.55 |
bjorkintosh |
are SIGGRAPH papers any good? |
03:15.22 |
dreeves |
there is still room for improvement on the
performance. Actually lost 15% somewhere on the end maybe
something for later... |
03:16.49 |
dreeves |
probably move on to the NURBS stuff I
suppose...although I think that is quite abit more
involved. |
03:22.53 |
starseeker |
bjorkintosh: yesh |
03:22.55 |
starseeker |
er yes |
03:23.25 |
starseeker |
all conferences will have ups and downs in
paper quality, but SIGGRAPH has lots of very good papers |
03:34.08 |
bjorkintosh |
really? |
03:48.16 |
*** join/#brlcad ashishrai
(i=d2d43dfb@gateway/web/ajax/mibbit.com/x-39632265f0993dad) |
06:31.03 |
*** join/#brlcad ewilhelm_
(n=ewilhelm@pool-71-111-78-159.ptldor.dsl-w.verizon.net) |
07:13.50 |
*** join/#brlcad madant
(n=madant@117.196.133.86) |
07:21.29 |
madant |
slept for 15 hours straight
:O |
08:19.29 |
*** join/#brlcad _sushi_
(n=_sushi_@84-72-93-63.dclient.hispeed.ch) |
09:24.10 |
*** join/#brlcad madant
(n=madant@117.196.142.174) |
09:51.15 |
*** join/#brlcad mafm
(n=mafm@223.Red-83-49-86.dynamicIP.rima-tde.net) |
10:00.17 |
madant |
howdy mafm |
10:55.36 |
d-lo |
Morning all. |
10:55.56 |
d-lo |
Bad traffic + thick fog = scary drive
:/ |
10:59.53 |
madant |
does not have a driving
license :P |
11:00.11 |
*** join/#brlcad mafm_
(n=mafm@223.Red-83-49-86.dynamicIP.rima-tde.net) |
11:08.02 |
d-lo |
Hrm, lemme think: Bad traffic + thick fog +
no license = even scarier!!! |
11:22.03 |
*** join/#brlcad _sushi__
(n=_sushi_@84-72-93-63.dclient.hispeed.ch) |
11:28.52 |
madant |
d-lo: :P true |
11:30.01 |
*** join/#brlcad ashishrai
(i=d2d43dfb@gateway/web/ajax/mibbit.com/x-cdc87fe2db0fc4ac) |
11:35.46 |
mafm_ |
hi |
11:40.14 |
madant |
hey :) |
11:41.30 |
madant |
freaky way to do pull-ups :p http://www.crossfit.com/mt-archive2/Queens-Palace-Pullup.html |
11:42.44 |
d-lo |
Just goes to show how well proper incentive
can motivate! |
11:45.26 |
d-lo |
This one looks strangely familiar......
http://www.crossfit.com/mt-archive2/crossfit-AnzioAnnie.html
;) |
11:47.04 |
brlcad |
it'd be a lot more impressive if he was doing
them off the barrel tip :) |
11:47.44 |
d-lo |
True, but with photoshot, anything is
possible! :) |
11:47.46 |
brlcad |
thick fog + rowing = eerie cool row
:) |
11:48.01 |
d-lo |
Yeah, i was wonding about that on the drive in
:) |
11:48.18 |
brlcad |
fortunately not all the crews have started so
the traffic was pretty low |
11:48.26 |
madant |
:D |
11:48.42 |
d-lo |
nice. Do you compete or is it just for the
halibut? |
11:48.45 |
brlcad |
it can get very scary |
11:48.58 |
brlcad |
I compete |
11:49.13 |
brlcad |
do ocd and competitive to just paddle
about |
11:50.24 |
madant |
has a backwater district
nearby :) |
11:50.31 |
d-lo |
On a somewhat related note, i found out that
my Patriot can corner VERY well :) 90 degree curve came up on me a
bit suddenly and I had to take it a bit faster than I wanted :)
Stiff suspension FTW. |
11:50.49 |
d-lo |
"every thing I know about performance driving
I learned from Top Gear" :) |
11:51.14 |
brlcad |
is starving |
11:51.36 |
brlcad |
madant: those look like fun pull-ups
actually |
11:51.56 |
brlcad |
I can see doing that |
11:52.24 |
madant |
yeah i mean height really doesnt matter ;) but
a nice pic nevertheless :) |
11:53.01 |
madant |
I can do 20 .. ok 19 :P |
11:53.27 |
madant |
though not confident about 20 at that height
:) |
11:54.04 |
brlcad |
hehe, nice ..
http://www.torch.aetc.af.mil/shared/media/photodb/photos/061228-F-1936B-010.JPG |
11:54.55 |
d-lo |
lol\ |
11:55.36 |
d-lo |
brlcad: If you can get to see the most recent
Top Gear (the one where the road test the Ford Fiesta) then you
really REALY should. |
11:56.09 |
d-lo |
I laughed my throat raw. That Humvee pic
reminded me of it btw :) |
11:57.18 |
madant |
yikes.. :) |
12:03.00 |
*** join/#brlcad BigAToo
(n=BigAToo@pool-96-230-124-61.sbndin.btas.verizon.net) |
12:18.41 |
CIA-40 |
BRL-CAD: 03indianlarry * r34055
10/brlcad/trunk/src/other/step/src/ (4 files in 4 dirs): Added
standard C++ library via LIBSTDCXX variable to build where
needed. |
12:35.23 |
*** join/#brlcad elite01
(n=omg@unaffiliated/elite01) |
14:40.21 |
*** join/#brlcad ashishrai
(i=d2d43dfb@gateway/web/ajax/mibbit.com/x-df09e762b5abd51c) |
14:53.27 |
brlcad |
madant: woot, top 20 :) |
14:54.33 |
madant |
huh / |
14:56.15 |
brlcad |
https://www.ohloh.net/p/brlcad/contributors |
14:57.42 |
madant |
:) aah.. it is going to get better
;) |
14:58.25 |
madant |
am sure i would have been way lower if they
didn't only count number of commits :D |
14:59.55 |
brlcad |
I've been wanting to do some mega
visualization at some point |
15:02.06 |
madant |
mega visualization ;) ? |
15:04.28 |
*** join/#brlcad BigAToo
(n=BigAToo@pool-96-230-124-61.sbndin.btas.verizon.net) |
15:15.01 |
*** join/#brlcad
Dr_Phreakenstein (n=phreak@216.151.24.198) |
15:22.53 |
*** join/#brlcad piksi
(i=piksi@pi-xi.net) [NETSPLIT VICTIM] |
15:25.01 |
*** join/#brlcad CIA-40
(n=CIA@208.69.182.149.simpli.biz) [NETSPLIT
VICTIM] |
15:27.36 |
dreeves |
how does the stuff work we submit as a patch?
Does it get included in the next release or is somewhat unknown
because someone needs to make time to look it over? |
15:32.11 |
*** join/#brlcad ``Erik__
(i=erik@c-76-111-12-116.hsd1.md.comcast.net) |
15:36.41 |
madant |
dreeves: depends on the type of patch i
guess.. mostly it gets added to the trunk pretty fast once someone
has a look at it .. some need a lot of regression tests and might
take longer.. if the person submitting the patch has been given
commit access then he is given the right to close the patch tracker
as well.. brlcad would obviously be able to clarify all this better
:) |
15:46.49 |
dreeves |
no big I was just curious if there was a
formal process or if it was just somewhat loose. I think what I
submitted is fairly low priority because I would guess not many
people are using that feature. Also it is just a performance
improvement. |
15:51.57 |
*** join/#brlcad ``Erik___
(i=erik@c-76-111-12-116.hsd1.md.comcast.net) |
16:00.35 |
brlcad |
madant, yeah visualization kinda like the code
storm, but actually somewhat more practical :) |
16:01.30 |
brlcad |
dreeves: the latter, one of the existing devs
takes the time to review and test/evaluate the patch |
16:04.33 |
brlcad |
yeah, basically what madant said -- it's not
at all a formalized process very much intentionally |
16:06.47 |
brlcad |
highly formalization in open source projects
is often a sign of a dying project, most operate under meritocratic
and high-iteration agile practices (or at least gravitate towards
them) |
16:07.29 |
dreeves |
Yeah I'm not a big fan of very formalized
process at all |
16:07.49 |
dreeves |
good way to bring a project to it's
knees |
16:07.52 |
brlcad |
that said, yeah, your patch isn't what I'd
call "high-priority" but it's actually very important because it's
from a new contributor :) |
16:08.06 |
dreeves |
:) |
16:08.37 |
brlcad |
and it's not just a performance improvement,
it's a _substantial_ performance improvement |
16:08.40 |
brlcad |
:) |
16:09.25 |
dreeves |
I think there is still room for more
improvement but I want to get on to the high priority stuff I
really just did that get warmed up |
16:10.31 |
dreeves |
I didn't want to start with something where I
was trying to get familiar with what was what at the same time as
tackling a tough problem i.e. NURBS |
16:10.52 |
brlcad |
nods |
16:11.53 |
madant |
brlcad: hmm.. yeah it would be fun ..
information visualization is always nice |
16:26.43 |
*** join/#brlcad ``Erik___
(i=erik@c-76-111-12-116.hsd1.md.comcast.net) |
16:30.31 |
brlcad |
dammit, sf.net is down right as I
submit |
16:30.50 |
brlcad |
~ping sourceforge.net |
16:30.51 |
ibot |
pong sourceforge.net |
16:34.30 |
brlcad |
there we go |
16:35.29 |
d-lo |
<PROTECTED> |
16:35.30 |
ibot |
pong sourceforge.net |
16:35.44 |
d-lo |
~google |
16:35.45 |
ibot |
hmm... google is at http://www.yahoo.com |
16:35.55 |
brlcad |
~d-lo is a BRL-CAD developer |
16:35.56 |
ibot |
brlcad: okay |
16:36.05 |
brlcad |
~d-lo |
16:36.06 |
ibot |
hmm... d-lo is a BRL-CAD developer |
16:38.19 |
dreeves |
didn't mean to include that other stuff I
forgot about messing around that |
16:41.52 |
dreeves |
actually only 3 files should have been in
there |
16:43.15 |
*** part/#brlcad ewilhelm
(n=ewilhelm@pool-71-111-78-159.ptldor.dsl-w.verizon.net) |
16:47.00 |
brlcad |
dreeves: I know, it was obvious |
16:47.00 |
brlcad |
i wouldn't have even mentioned it save for the
other issues |
16:47.00 |
brlcad |
all minor, but patch acceptance generally puts
it in the submitter's hands to fix it so the issues are resolved
earlier instead of later |
16:47.01 |
brlcad |
(as the goal is to give commit) |
16:49.06 |
dreeves |
absolutely yeah don't let me slide on anything
I want comply no problem |
16:49.35 |
starseeker |
A bat hanging on to the space shuttle as it
launches... awesome :-) |
16:49.43 |
dreeves |
I tried to follow the style but obviously
hasn't been the style I have been follow as of late so habits and
such |
16:50.26 |
dreeves |
So do I just re attach a new patch or is it a
new submit altogether |
16:51.08 |
dreeves |
btw it is only 3 files |
16:53.13 |
d-lo |
Hrm, if a small peice of foam can cause a
catastrophic loss of the orbiter.... i wonder what fate that bat
has condemned them to? :/ |
16:54.41 |
dreeves |
I'm sure the "small" piece of foam was
considerably larger than the bat |
16:55.45 |
dreeves |
was the brace style the only thing you noticed
when reviewing the code? |
16:55.47 |
starseeker |
yeah, no one is worried that the bat will
cause launcher damage |
16:56.35 |
starseeker |
wonders if there is a bat
version of the Darwin award... |
16:56.51 |
dreeves |
However I pretty sure the bat didn't make
it |
16:57.08 |
d-lo |
Dunno, sounds like the basis for Samuel
Jacksons next movie: Bats on a Shuttle... |
16:59.13 |
brlcad |
dreeves: not a problem .. there are probably
four distinct styles throughout the brl-cad sources -- just try to
police new code and hit up files for consistency as they are worked
on |
16:59.39 |
brlcad |
most of the files have issues, it'd takes many
weeks of tedium to hit up everything (it can only partially be
automated) |
17:00.13 |
dreeves |
ok no problem |
17:00.14 |
brlcad |
dreeves: braces was the only stylistic thing I
noticed that mattered |
17:01.08 |
brlcad |
the other stylistic changes are not worth
policing unless someone specifically wants to go on a
rampage |
17:01.41 |
dreeves |
I was trying to follow the correct style but
in my day job I use the other style so I will need to get the
settings in my editor so I don't forget |
17:01.42 |
brlcad |
or is just egregiously annoying like spaces
before commas |
17:01.50 |
brlcad |
nods |
17:01.56 |
brlcad |
I use the other style on a few other
projects |
17:02.04 |
dreeves |
hey the 30 setting that should should have
been 64 |
17:02.05 |
brlcad |
some people have religion on one vs the
other |
17:02.20 |
brlcad |
to me, it's just more important that the code
is consistent |
17:02.46 |
dreeves |
Yeah I'm not religious on that kind of stuff
would rather spend my energy on more useful things |
17:02.54 |
brlcad |
minimally per-file consistent, ideally
globally consistent (sans src/other) |
17:02.56 |
dreeves |
yes I agree |
17:03.54 |
dreeves |
that setting was in the bezier_2d_isect.c
before but if I'm to pre allocate the buffers then the setting
needs to come from extude.c |
17:05.48 |
dreeves |
I tried to think of a way to get rid of it all
together but at the time I was tired (no sleep in 30+ hours) and I
couldn't think of anything. So I gave in a moved the setting to
extude.c and passed into bezier_2d_isect.c |
17:07.02 |
dreeves |
I was just messing around and wanted to see
what it would like like if I changed the setting to 30 vs 64.
Didn't make much of an impact on either. I'm sure it would with a
more complex shape |
17:10.52 |
*** join/#brlcad BigAToo
(n=BigAToo@pool-96-230-124-61.sbndin.btas.verizon.net) |
17:15.02 |
brlcad |
yeah, it sounds like the maximum number of
intersects through the sketch |
17:15.11 |
brlcad |
just at a quick glance |
17:15.22 |
brlcad |
if that's true, even 64 is kinda sketchy (no
pun intended) |
17:15.41 |
brlcad |
some simple grided sketch would blow out the
limit |
17:19.33 |
dreeves |
no i don't think that is what that number
is |
17:20.20 |
dreeves |
I think that is the number of times to
subdivide the curve in order to resolve down to a single segment of
the curve to find the particular intersection |
17:26.09 |
*** join/#brlcad madant
(n=madant@117.196.144.14) |
17:26.10 |
brlcad |
ah, hm |
17:26.33 |
brlcad |
then that sounds particularly
asinine |
17:26.52 |
brlcad |
especially if it has a ton of control
points |
17:27.08 |
brlcad |
unless that's between each pair of control
points |
17:28.19 |
brlcad |
probably should be based adaptive based on the
amount of curvature in a given piecewise region |
17:30.12 |
dreeves |
yeah checks the curvature |
17:31.01 |
dreeves |
it tries to subdivide to get down to little or
no curvature but at some point it stops which the maxlevel
setting |
17:33.18 |
dreeves |
let me restate |
17:34.53 |
brlcad |
aha.. so then after a given minimized
curvature, it then performs MAXDEPTH samples of that
piece |
17:34.57 |
brlcad |
that's not so bad then |
17:35.11 |
brlcad |
probably even overkill depending on that
curvature setting |
17:35.36 |
dreeves |
exactly that is why I was playing around with
the number a little |
17:36.09 |
dreeves |
no I misspoke |
17:36.19 |
dreeves |
I was saying exactly to the overkill
part |
17:36.40 |
dreeves |
but the algorithm doesn't exactly work that
way |
17:38.06 |
dreeves |
it divides the control polygon down to attempt
to consider a single segment of the polygon but it also looks at
curvature to make sure it is only considering a case where the ray
will only cross that segment once |
17:38.40 |
dreeves |
if the curvature is not "flat" enough it will
continue to subdivide |
17:40.26 |
dreeves |
however it gives on continuing to subdivide if
it reaches the max subdivides |
17:41.30 |
dreeves |
the is actually no limit on how far it will
subdivide |
17:41.56 |
dreeves |
it only consider the depth after it has gotten
to a single segment of the control polygon |
17:42.42 |
dreeves |
I can imagine some situations that could cause
it problems but I think they are pretty extreme |
17:44.43 |
dreeves |
If I right I think control polygons 2 to 32nd
power is where you will see the algorithm start to have problems.
Well maybe a few powers lower but in that area |
17:45.43 |
dreeves |
no actually 64th power |
17:53.33 |
dreeves |
Hope what I wrote makes since here reading
backup I realize I'm dropping words... |
17:54.16 |
dreeves |
Of course I could be completely off track but
that is what I think is going on |
18:01.14 |
*** join/#brlcad ashish
(i=d2d43dfb@gateway/web/ajax/mibbit.com/x-f77d495992c1e9eb) |
18:22.29 |
*** join/#brlcad madant_
(n=madant@117.196.133.222) |
18:24.04 |
brlcad |
dreeves: you could toss in a zero-tolerance
check on the polygon points -- if they're all within computation
tolerance, halt |
18:43.47 |
*** join/#brlcad BigATo1
(n=BigAToo@pool-96-230-124-87.sbndin.btas.verizon.net) |
18:56.57 |
*** join/#brlcad
Dr_Phreakenstein (n=phreak@216.151.24.198) |
19:09.54 |
CIA-40 |
BRL-CAD: 03bob1961 * r34056
10/brlcad/trunk/src/librt/shoot.c: Modified rt_res_pieces_clean()
to skip uninitialized rt_piecestate structures. |
19:47.23 |
*** join/#brlcad cad58
(n=c752f349@bz.bzflag.bz) |
19:58.55 |
*** join/#brlcad andrecastelo
(n=chatzill@201008160086.user.veloxzone.com.br) |
20:04.37 |
*** join/#brlcad elite01
(n=omg@unaffiliated/elite01) |
20:07.33 |
*** join/#brlcad dreeves
(n=c752f349@bz.bzflag.bz) |
20:09.27 |
dreeves |
brlcad had to go to the office after my
comments so I didn't see what you wrote a response if
anything |
20:30.39 |
*** join/#brlcad kanzure
(i=bryan@66.112.232.233) |
20:31.06 |
kanzure |
considers working on a
project listed on http://brlcad.org/wiki/Google_Summer_of_Code/2009/Project_Ideas |
20:42.15 |
*** join/#brlcad piksi
(i=piksi@pi-xi.net) |
20:42.18 |
*** join/#brlcad madant
(n=madant@117.196.130.208) |
21:03.04 |
brlcad |
ah, dreeves left |
21:03.06 |
*** join/#brlcad _sushi_
(n=_sushi_@77-58-236-246.dclient.hispeed.ch) |
21:03.07 |
brlcad |
kanzure: glad to hear it! |
21:06.29 |
kanzure |
brlcad: Should I just submit a project
proposal, or how does this work? |
21:10.57 |
madant |
congrats :) |
21:19.51 |
madant |
hmm.. and they have changed the timeline as
brlcad predicted :D |
21:21.33 |
brlcad |
kanzure: discussion |
21:22.19 |
brlcad |
what do you want to work on? why? who are
you? (the 2cent gist version) why do you want to work on brl-cad,
what are your goals, etc |
21:22.34 |
brlcad |
madant: :) |
21:25.08 |
madant |
next 10 days are going to be fun on the
channel and mailing list :) |
21:29.18 |
madant |
brlcad: gsoc blog says "larger student peer
groups in each project" ;) |
21:41.43 |
kanzure |
hm, it's entirely possible that you guys don't
remember me :) |
21:41.46 |
kanzure |
http://heybryan.org/om.html |
21:42.06 |
kanzure |
http://heybryan.org/~bbishop/docs/gears/gears.html |
21:42.43 |
kanzure |
interchangeable/intercompatible parts and
automatic instruction generators, researcher here at the Automated
Design Lab at UT Austin :-) |
21:43.45 |
kanzure |
so in particular I'd probably end up working
on the parametric library or web thingy for design repository, in
particular my background has been in open source hardware design
packaging (much like "dot deb" and "dot rpm" files except, uh, for
hardware) |
21:45.57 |
*** join/#brlcad BigAToo
(n=BigAToo@pool-96-230-124-87.sbndin.btas.verizon.net) |
21:47.36 |
kanzure |
some stuff I've recently been doing (sadly,
it's not with BRLCAD):
http://heybryan.org/books/Manufacturing/pythonocc/2009-03-15_3.png |
21:50.23 |
*** join/#brlcad samrose
(n=samrose@oh-69-69-33-40.sta.embarqhsd.net) |
21:51.47 |
brlcad |
oh yeah, I guess I didn't announce it here yet
:) |
21:51.51 |
brlcad |
we made it into gsoc 2009 |
21:52.11 |
brlcad |
hurrahs |
21:52.35 |
kanzure |
http://heybryan.org/bioreactor/membraneless_filtration/2009-03-18_spiral.jpg
(warning: face) |
21:52.43 |
kanzure |
brlcad: remember my work on spirals? |
21:52.45 |
kanzure |
:) |
21:52.51 |
kanzure |
now this just needs to be tested in real
life. |
21:53.02 |
brlcad |
<PROTECTED> |
21:53.13 |
kanzure |
DNS-wise? |
21:53.30 |
brlcad |
hah, yes |
21:53.35 |
brlcad |
resolves to 127.0.0.1 |
21:53.47 |
_sushi_ |
resolves to 127.0.0.1
too |
21:54.02 |
kanzure |
http://66.112.232.182/bioreactor/membraneless_filtration/2009-03-18_spiral.jpg
(warning: face) |
21:54.05 |
kanzure |
try that? |
21:54.24 |
brlcad |
kanzure: and to answer your question, I did
remember the name, just not the connection |
21:54.29 |
kanzure |
okay |
21:56.21 |
*** join/#brlcad louipc
(n=louipc@archlinux/trusteduser/louipc) |
21:56.31 |
brlcad |
madant: hm, that's actually an interesting
quote on the blog .. have to talk to lh about it |
21:57.52 |
brlcad |
bz's going to test almost the exact opposite
mentoring-wise as our experience indicated that we needed fewer
students in order to better engage them |
21:58.03 |
madant |
kanzure: what does pythonocc do ? |
21:58.04 |
brlcad |
probably just that there's a sweet spot that
orgs have to find |
21:58.08 |
kanzure |
do mentoring organizations have a minimum team
size? |
21:58.16 |
brlcad |
nope |
21:58.17 |
kanzure |
madant: pythonOCC == python bindings to
OpenCASCADE |
21:58.20 |
kanzure |
http://pythonocc.org/ |
21:58.24 |
madant |
ah.. :) |
21:58.32 |
kanzure |
so I was just putting together pythonOCC + the
HeeksCAD interface together |
21:58.35 |
kanzure |
since they both use wxWidgets |
21:59.43 |
madant |
k.. opencascade is pretty neat ;) |
22:00.03 |
kanzure |
but the API *sucks*. I mean, coming from the
perspective of trying to figure out how to write a simpler wrapper
for it, or something, it's just huge and hairy |
22:00.12 |
brlcad |
thinks they still fail the
freeness test |
22:00.26 |
brlcad |
open yes, just not free |
22:00.33 |
archivist |
kanzure, author of HeeksCAD often is in
#cam |
22:00.42 |
kanzure |
archivist: thank you for the heads
up |
22:01.06 |
madant |
freeness maybe :) but not many CAE tools in
open source scene otherwise right ? |
22:01.08 |
kanzure |
brlcad: apparently they are changing their
licensing actually |
22:01.35 |
brlcad |
they've said that for a couple years |
22:01.39 |
kanzure |
oh :( |
22:01.58 |
brlcad |
and they did tweak it once about two years
ago, as it was originally entirely ambiguous |
22:02.35 |
madant |
they had BRL-CAD in one of their CDs or
something if i remember correctly ? |
22:02.45 |
brlcad |
did they? |
22:02.48 |
brlcad |
that'd surprise me |
22:02.50 |
kanzure |
why? |
22:02.56 |
kanzure |
I mean, why would they put it on? |
22:02.58 |
brlcad |
or at least be news to me |
22:03.21 |
madant |
let me see :) |
22:03.27 |
kanzure |
we were hoping that pythonocc would make it
into gsoc, but that didn't happen this year |
22:03.37 |
kanzure |
but brlcad is a good alternative,
heh |
22:03.47 |
brlcad |
I'm all for competition and collaboration, but
that project has just smelled bad from the beginning |
22:05.12 |
brlcad |
seemed more of a failed commercial attempt
that was dumped onto a website with an "Open" label slapped onto it
in order to try to attract some business, then when that didn't
work, they kept tweaking the license to try to actually open it up
"just enough" to get some revenue through providing commercial API
support |
22:05.13 |
madant |
ah maybe it was the CAELinux dvd :) |
22:05.21 |
kanzure |
the giant DVD library? :) |
22:06.01 |
kanzure |
ah, what I think I meant to say was that
OpenCASCADE recently became "open" and "free" enough to warrant
debian packaging and inclusion into the main repositories |
22:06.05 |
brlcad |
i could be completely wrong in that
perception, but i was pretty excited to try and join our geometry
engines several years back only to be sorely disappointed on many
fronts |
22:06.09 |
kanzure |
(6.3.0 hit the repositories last
week) |
22:06.27 |
kanzure |
brlcad: can you describe what happened on that
front in particular? |
22:07.58 |
brlcad |
kanzure: well part of what I just mentioned,
that their entire intentions haven't felt like they're with any
actual regard or interest with F/OSS development of CAD |
22:08.02 |
madant |
hmmm opencascade in debian and not in non-free
? i remember somebody was trying to package salome and opencascade
a year ago.. |
22:08.34 |
kanzure |
things that sound interesting: 1.3, 1.7, 1.10,
2.8 (from the wiki page) |
22:08.38 |
brlcad |
doing only the minimum they can get away with
to conjure up business, not being open to the community at
large |
22:08.47 |
kanzure |
brlcad: no, I mean the joining of the geometry
engines :) |
22:08.54 |
kanzure |
sorry for the ambiguity |
22:08.59 |
brlcad |
I mean that too |
22:09.13 |
kanzure |
oh |
22:09.15 |
brlcad |
there's no technical discussion to be had if
there are legal issues and project management issues |
22:09.28 |
kanzure |
were their headers weird or
something? |
22:09.29 |
kanzure |
ah, I see. |
22:09.40 |
madant |
brlcad: news coming up at main page ;)
? |
22:10.17 |
brlcad |
their license originally was completely
proprietary iirc, the only thing "open" about them was that you
could get the source code (but you couldn't redistribute) and they
had the word "Open" in their name |
22:10.29 |
brlcad |
madant: mailing list first, but yeah |
22:10.33 |
brlcad |
probably later today |
22:11.11 |
kanzure |
where is the brlcad-devel mailing list
located? is it on the sf.net server? |
22:11.17 |
brlcad |
yes, sf |
22:11.58 |
kanzure |
http://lists.sourceforge.net/lists/listinfo/brlcad-devel |
22:11.59 |
kanzure |
aha |
22:13.58 |
louipc |
heh opennurbs seems kind of closed when it
comes to the community too |
22:14.07 |
Ralith |
yay, gsoc! |
22:15.04 |
madant |
hey Ralith, howdy |
22:15.14 |
Ralith |
hullo |
22:15.31 |
kanzure |
louipc: isn't that the blender
module? |
22:16.01 |
Ralith |
so I know we've got plenty of engs in
here |
22:16.08 |
Ralith |
anyone have any thoughts on the hp50g
calculator? |
22:16.42 |
kanzure |
huh, you guys were solving constraint
satisfaction problems with graphs |
22:16.44 |
Ralith |
I discovered my ti-89's not talking with my
computer anymore. |
22:16.47 |
madant |
has never used a graphing
calculator :D |
22:16.57 |
kanzure |
has done some work on graph
grammars for graph-based design automation. |
22:17.08 |
kanzure |
http://brlcad.org/wiki/Libpg_:_A_parametrics/constraint_library |
22:17.12 |
louipc |
kanzure: no that's the rhino 3d nurbs
library |
22:17.14 |
louipc |
http://www.opennurbs.org/ |
22:17.31 |
madant |
kanzure: graphs .. well my initial idea was
that bgl would be useful.. but now i think maybe not.. |
22:17.44 |
kanzure |
bgl? |
22:18.03 |
madant |
boost graph library |
22:18.10 |
kanzure |
so, our lab uses software called
"GraphSynth" |
22:18.17 |
kanzure |
it's a graph grammar solver for engineering
problems, more or less |
22:18.28 |
madant |
boost's hypergraph support the last time i
checked was not that great.. |
22:18.29 |
kanzure |
one student is doing gear-based automated
design and optimization, another is doing linkages, and another is
doing product disassembly |
22:18.41 |
louipc |
kanzure: brl-cad requires opennurbs |
22:18.45 |
kanzure |
so it might be interesting to integrate the
graph functionality into libpg |
22:19.15 |
madant |
kanzure: what exactly do u mean by graph
functionality ? |
22:19.19 |
archivist |
hmm gears |
22:19.31 |
kanzure |
archivist: did you see the link? |
22:19.49 |
kanzure |
http://heybryan.org/~bbishop/docs/gears/gears.html |
22:20.10 |
kanzure |
madant: I mean the graphsynth classes for the
representation of graphs, generating thousands of solutions for a
given graph, and so on |
22:20.37 |
kanzure |
(just the other day we were generating a
thousand 'design solutions' from a given 'specification graph'
(these aren't the actual terribly convoluted names they have, but
that's what they are)) |
22:20.51 |
madant |
:) |
22:21.06 |
madant |
just checking out the graphsynth
page |
22:21.45 |
kanzure |
think of it as a search-and-replace system
that goes through the entire tree of possible substitutions given a
set of 'replace rules' (a "rule set") |
22:22.02 |
louipc |
those are some funky gears |
22:22.07 |
kanzure |
it's a hack to glxgears :) |
22:22.29 |
*** join/#brlcad
Dr_Phreakenstein (n=phreak@216.151.24.198) |
22:22.43 |
kanzure |
I was going to write a gear
visualizer/CAD-generator with pythonocc |
22:22.49 |
Ralith |
there's a cool xscreensaver module that
generates very (infinitely?) long serieses of properly interlocking
gears |
22:22.54 |
kanzure |
but I haven't got around to that yet |
22:23.00 |
madant |
hmm. and the design is effectively a
'configuration' expressed in XML ? |
22:23.02 |
kanzure |
Ralith: oh? I've never seen that, is that in
the standard X11 package? |
22:23.05 |
kanzure |
madant: yes |
22:23.08 |
Ralith |
no idea |
22:23.13 |
Ralith |
it might not even be on my system |
22:23.21 |
archivist |
glxgears |
22:23.23 |
kanzure |
Ralith: ah, maybe it's in the xscreensaver
package (sorry, didn't read) |
22:23.24 |
Ralith |
I always set my screensavers to random, so I
can't help beyond that |
22:23.29 |
kanzure |
glxgears doesn't do interlocking of infinitely
many gears |
22:23.35 |
Ralith |
might even haev only been in the freebsd
distrib |
22:23.42 |
Ralith |
yeah this is much more advanced than
glxgears |
22:23.46 |
Ralith |
nicer looking models, too |
22:23.49 |
kanzure |
anyway, it sucks because it's not actual
CAD |
22:23.55 |
kanzure |
it's just a cylinder and some OpenGL
calls |
22:24.03 |
kanzure |
kinda useless for lab research :) |
22:24.12 |
Ralith |
I wouldn't think parametric gears would be
very hard |
22:24.17 |
kanzure |
not at all |
22:24.17 |
madant |
kanzure: does graphsynth support hypergraphs
? |
22:24.31 |
kanzure |
madant: please excuse me, I'm not as well
versed in hypergraphs as I should be |
22:24.36 |
madant |
nods |
22:24.38 |
kanzure |
is this where there are vertices with multiple
edges? |
22:24.45 |
Ralith |
hm |
22:24.55 |
Ralith |
brlcad: does BRL-CAD have any sort of geargen
tool? |
22:25.05 |
bjorkintosh |
geargen? gear generator? |
22:25.07 |
Ralith |
yeah |
22:25.08 |
kanzure |
yes |
22:25.13 |
kanzure |
madant: is that what a hypergraph
is? |
22:25.14 |
Ralith |
okay, cool |
22:25.23 |
kanzure |
Ralith: no, I wasn't answering you |
22:25.26 |
Ralith |
oh |
22:25.45 |
Ralith |
I ask because we've got weird things like tire
generators and fence generators but I don't recall ever seeing a
gear generator |
22:25.45 |
madant |
kanzure: the hypergraph idea is basically that
you can have edges between edges.. |
22:25.58 |
bjorkintosh |
fence generator? |
22:26.00 |
bjorkintosh |
whoa! |
22:26.03 |
Ralith |
chain link fence |
22:26.07 |
bjorkintosh |
is in white picket ... |
22:26.08 |
bjorkintosh |
oh okay. |
22:26.13 |
kanzure |
madant: blah? wha? |
22:26.13 |
bjorkintosh |
*as in. |
22:26.46 |
kanzure |
madant: http://en.wikipedia.org/wiki/Hypergraph |
22:26.48 |
kanzure |
"In mathematics, a hypergraph is a
generalization of a graph, where edges can connect any number of
vertices. " |
22:26.52 |
kanzure |
in which case, yes |
22:29.19 |
louipc |
hmm! those generators used to be prefixed with
mk_ |
22:29.32 |
louipc |
now I don't know what the generators are
blah |
22:29.48 |
kanzure |
just grep over the source for 'gear'
:-) |
22:29.53 |
kanzure |
grep -H -I -R "gear" * |
22:30.05 |
madant |
kanzure: if you consider a constraint
network.. it eventually has lets say a set of nodes( variables) and
constraints( edges between two or moredependent variables) but
constraints might depend on other constraints as well.. so you have
hyperedges connecting two or more edges which can effectively be
broken down into edges between many vertices.. |
22:30.31 |
kanzure |
I see |
22:30.56 |
Ralith |
kanzure: hehe |
22:30.59 |
louipc |
kanzure: that's not very efficient, and
doesn't help those who only installed the binary |
22:31.06 |
madant |
consider a constraint which depends on two
other constraints for instance .. |
22:31.11 |
kanzure |
madant: right |
22:31.27 |
madant |
so graphsynth supports hypergraphs ? |
22:32.47 |
kanzure |
the answer is complicated. yes in the sense
that the constraint network can be represented, no in the sense
that the way that you solve it is not by treating it as a
hypergraph |
22:32.50 |
kanzure |
I don't know if that makes sense |
22:33.19 |
kanzure |
see, when a constraint network would be added
into graphsynth, the "rules" would be called forth to solve a
particular subset of the graph |
22:33.32 |
kanzure |
then it would be possible, say, to find the
next part of the constraint network needs to be updated |
22:33.36 |
kanzure |
dependent on the last modification that was
made |
22:33.53 |
kanzure |
which is essentially solving the same thing,
no? |
22:34.20 |
madant |
looking at the
grammar.. |
22:34.35 |
kanzure |
well that's the thing, the grammars are
hand-crafted |
22:34.51 |
kanzure |
like I said, one researcher has made a grammar
for gear generation, another for linkages (or is working on it),
another for product disassembly |
22:35.22 |
kanzure |
(and there's a few others- one for how to tie
a tie, electronics (filter circuits), and stuff I'm
forgetting) |
22:35.42 |
madant |
kanzure: i meant the basic designGraph
class |
22:36.29 |
madant |
what are the node classes ? |
22:36.48 |
kanzure |
vertices. |
22:36.58 |
kanzure |
what are you asking? the name of the node
class is 'node' IIRC |
22:37.14 |
madant |
i mean they directly correlate with the
variables right ? |
22:37.48 |
kanzure |
uhrm, sort of. each node has a list of labels,
and it's possible for a node to only have one label, so in this
case that label might be a variable related to the constraint
expression, yeah |
22:38.11 |
louipc |
starseeker: should src/libged/tire.c be
removed since it's already in src/shapes/tire.c? |
22:38.59 |
madant |
ah.. so designGraph is basically an
implementation of a directed graph right ? but hypergraphs are not
effectively reducible to directed graphs |
22:39.27 |
kanzure |
directed means that edges point in a certain
direction |
22:39.33 |
kanzure |
but in this case, edges are
bidirectional |
22:39.51 |
kanzure |
I mean, each node has a list of the edges that
are connected to it |
22:40.32 |
brlcad |
oof, turn my head for just a few
minutes.. |
22:40.45 |
madant |
:D |
22:43.17 |
brlcad |
louipc: mcneal & assoc. were very upfront
from the very beginning that there was no support, no community,
and a simple PD license, and open about the fact that their intent
was more converters and that they have no intention of helping
folks use it for analytic or modeling purposes .. big difference in
intent and flexibility |
22:43.58 |
madant |
kanzure: consider this hypergraph ( vertices :
A,B,C,D,E,F Edges/Constraints: A perp. B, Cperp. D, E parallel F if
A perp. B and C perp. D else E perp. F) how would this be
represented in graphsynth |
22:44.08 |
madant |
6 variables and 3 constraints |
22:45.05 |
kanzure |
I would have to play around with possible
representation methods :) |
22:45.14 |
kanzure |
one idea that comes to mind is using nodes for
perpendicularity/parallelism |
22:45.16 |
brlcad |
kanzure: 404 on http://66.112.232.182/~bbishop/docs/gears/gears.html |
22:45.31 |
kanzure |
so if two nodes are connected by "parallel" or
"perpendicular", then that's that. |
22:46.05 |
brlcad |
Ralith: don't think so, but pretty simple to
create with the pattern tool .. something I did in a couple minutes
a couple years ago here: http://brlcad.org/tmp/gear/ |
22:46.30 |
brlcad |
bjorkintosh: there's a picket fence generator
too |
22:46.43 |
kanzure |
ah, you were the one who did
/tmp/gear |
22:46.46 |
Ralith |
hm |
22:46.48 |
kanzure |
remembers seeing that a few
weeks ago |
22:47.17 |
Ralith |
not the most gearlike, but I imagine that's
not a limitation of the system |
22:47.23 |
brlcad |
louipc: ls src/shapes |
22:47.35 |
madant |
kanzure: just describe the new graph for me
? |
22:47.42 |
bjorkintosh |
haha |
22:47.46 |
bjorkintosh |
who put these things in there,
brlcad? |
22:48.19 |
brlcad |
louipc: look at the implementation of
src/shapes/tire.c .. ged_tire() |
22:48.26 |
madant |
kanzure: u mean two nodes and multiple edges
between them ? |
22:49.06 |
madant |
kanzure: in that case? lets say the nodes are
P(perpendicularity) and U ( parallel) .. how do u model the third
constraint ? |
22:49.21 |
kanzure |
madant: what does the if/then statement
signify? "if A per. B" does that mean "if A perp. B is possible"
? |
22:49.48 |
brlcad |
kanzure: /tmp/gear was a quick example for
someone based on known curvature parameters, forget who
though |
22:50.04 |
madant |
kanzure: if A perp. B is satisfied |
22:50.15 |
kanzure |
then that should be in the grammar
rules |
22:50.28 |
brlcad |
bjorkintosh: various developers -- creating a
procedural modeling tool that generates something is actually a
great way to become familiar with the code base |
22:50.34 |
kanzure |
it would be a new rule: "if you see this graph
structure, then make this modification" |
22:50.48 |
bjorkintosh |
you don't say. |
22:50.53 |
starseeker |
has considered making a gear
generator, but coils were more manual labor for a usable result and
less tractable to creation via tools like clone |
22:50.56 |
kanzure |
in particular it would be, using the model I
suggested off the top of my head, "if you see A-perp-B (vertices:
A, perp, B), then do blah blah blah" |
22:51.11 |
kanzure |
erm, sorry, vertices: A, perpSOME_ID_NUMBER,
B |
22:51.21 |
kanzure |
the perpendicularity between the two is
supposedly unique :-) |
22:51.26 |
brlcad |
so we often have new developers try to write
an app that makes something (of their choice), they end up learning
some of libbu, libbn, librt, libwdb and we end up with a simple
tool that generates some shape procedurally usually |
22:51.30 |
starseeker |
gears will probably come later, but the
variety is amazing in gears and much thought would be needed to
create a sufficiently general framework |
22:51.34 |
madant |
kanzure: k :) |
22:51.42 |
kanzure |
starseeker: oh? |
22:51.51 |
kanzure |
starseeker: that's kind of the work that I'm
supposed to be doing right now for the lab |
22:52.05 |
kanzure |
in particular, I'm also supposedly coming up
with a way to display helical gears, worm gears, bevel gears, etc.
etc. |
22:52.11 |
starseeker |
cooool :-) |
22:52.11 |
Ralith |
starseeker: isn't most of the variety just in
the details of the tooth shape? |
22:52.27 |
kanzure |
and I get to ignore tooth shape (more or
less)- which I know you guys are going to hate me for |
22:52.31 |
starseeker |
Ralith: tooth shape, tooth spacing, overall
gear shape, gear interior... |
22:52.40 |
brlcad |
e.g., I wrote the chain link 'fence' tool
probably a decade ago -- it comes in handy for generating scenes
for something that would otherwise be very painstaking to model by
hand (it generates a physically accurate chain link fence with all
the bends and individual wires) |
22:52.42 |
kanzure |
madant: I think it can be useful for
parametrics. I'll definitely talk it over with some labmates
soon. |
22:53.13 |
madant |
kanzure: so effectively the constraint graph
in itself does not represent the problem entirely ? |
22:53.23 |
Ralith |
starseeker: doesn't strike me as terribly hard
to abstract. |
22:53.40 |
brlcad |
you probably have starseeker drooling now ..
he's been talking about gears and coils for a while :) (and now
has the latter mostly working) |
22:53.44 |
kanzure |
madant: maybe. I'm still not sure- why is
there an if/then in your constraint specification? |
22:53.59 |
starseeker |
sticks tounge out at
brlcad |
22:54.13 |
kanzure |
starseeker: you should definitely check out
the code in the directory I keep linking to |
22:54.19 |
madant |
kanzure: it is a possible logical constraint
right :) |
22:54.19 |
kanzure |
http://heybryan.org/~bbishop/docs/gears/ |
22:54.27 |
starseeker |
Ralith: It isn't hard to abstract - that's
one of the reasons it's a good candidate for a tool |
22:54.33 |
starseeker |
looks |
22:54.53 |
madant |
since constraints effectively have two states
: only constraints possible between them are in essence
logical |
22:54.54 |
kanzure |
madant: can you give me an example of a
situation where I inadvertedly used a logical constraint in a CAD
app without perhaps knowing it? |
22:55.10 |
starseeker |
kanzure: Can't see it from here - I'll have
to look later |
22:55.21 |
kanzure |
starseeker: blah, some of you are unable to
load the pages, while others in this channel can |
22:55.24 |
kanzure |
wtf? |
22:55.57 |
madant |
kanzure: :) am not really sure if any CAD
application supports it.. I mean i have only worked with CATIA and
it doesn't.. |
22:56.08 |
Ralith |
starseeker: well, what did you mean by "much
thought would be needed to create a sufficiently general
framework"? |
22:56.09 |
starseeker |
Ralith: a lot of the work involved with a
gear tool would be identifying what the standard methods are for
specifying gear type, size, etc. and wiring up the logic to
generate the proper geometry based on those numbers |
22:56.30 |
kanzure |
yay |
22:56.34 |
Ralith |
yeah, I imagine it would be mostly
research |
22:56.35 |
kanzure |
starseeker: we should definitely collaborate
on this |
22:56.35 |
starseeker |
for each gear type, there may be a whole
sub-set of specifications to look at |
22:56.39 |
madant |
kanzure: but not sure if many CAD applications
support it internally.. will have to think of a case :) |
22:56.41 |
Ralith |
it sounds ilke an interesting task,
though |
22:56.46 |
kanzure |
Ralith: I'm working with somebody who has done
most of that work |
22:56.49 |
kanzure |
that research work, I mean |
22:56.58 |
Ralith |
kanzure: cool; has he published his findings
anywhere? |
22:57.03 |
kanzure |
paper in progress |
22:57.04 |
kanzure |
:) |
22:57.10 |
starseeker |
kanzure: It'll be a while before I can devote
much time to it - I've got other more pressing tasks :-/ |
22:57.17 |
Ralith |
let me know when it's up |
22:57.19 |
Ralith |
I might just write this |
22:57.20 |
madant |
kanzure: how many people are working with
parametrics and constraints at UT ? |
22:57.37 |
kanzure |
madant: I don't know anybody who is explicitly
working on that problem. :( |
22:57.38 |
kanzure |
it's so sad |
22:57.49 |
brlcad |
kanzure: I believe unigraphics can get you
into some situations where you use constraints implicitly |
22:57.51 |
kanzure |
I guess there's me. |
22:58.16 |
brlcad |
it's got a red-light/green-light to let you
know if you're fully constrained or under/over too, interesting
concept |
22:58.33 |
starseeker |
Ralith: then for each gear type, you have to
think about precisely what geometry you want to use to represent
it. For example, would it be better to use sketches for gear
patterns that can be represented as extrusions, or should such
gears be combinations of arbs and tgcs? |
22:58.43 |
kanzure |
yes, I've seen something like that in
Solidworks re: red/green light for over/under constrained |
22:58.49 |
starseeker |
extrusions would do for a lot of types, but
not all |
22:58.59 |
madant |
brlcad: i like the similar feature in catia
sketcher.. where they show u overconstrained and underconstrained
and correctly constrained elements in different colours |
22:59.06 |
Ralith |
yeah, I realized that |
22:59.09 |
kanzure |
starseeker: in my case, the generated CAD file
would have to be dependent on the variables that the optimization
engine is spitting out |
22:59.12 |
Ralith |
we can't twist extrusions yet, can
we? |
22:59.14 |
kanzure |
so that kind of solves that problem
eh? |
22:59.22 |
kanzure |
but makes for a perhaps overly constrained
gear generator in the end of course :) |
22:59.36 |
brlcad |
has wanted similar lights for
constraining once it's all working, but even now for overlap
reporting |
22:59.55 |
starseeker |
kanzure: In part. You also want to be able
to manually specify gears from the command line (or, better, from
within MGED) using standard parameters |
23:00.00 |
kanzure |
right |
23:00.05 |
kanzure |
no argument there. |
23:00.16 |
Ralith |
some sort of extrude along a path describing
rotation as well as direction; is there a name for that? |
23:00.20 |
madant |
ah.. so it is effectively : if ( constraint1 =
true , constraint 2= true .... ) then constraints=true ;) hypegraph
:D |
23:00.20 |
kanzure |
madant: so was your logic constraint spec an
imaginary one? :) |
23:00.35 |
starseeker |
Ralith: sounds like sweep |
23:00.39 |
Ralith |
that's it |
23:00.44 |
Ralith |
we don't have that yet, right? |
23:00.47 |
starseeker |
right |
23:00.55 |
Ralith |
'cuz that would make things easier. |
23:01.00 |
starseeker |
indeed |
23:01.02 |
kanzure |
madant: yeah, the solutions would be those
graphs where "all constraints have been satisfied" (or eliminated,
or completely replaced, or something) |
23:01.32 |
madant |
kanzure: yep :) |
23:01.41 |
kanzure |
what would BRLCAD do with multiple solutions
to a constraints satisfaction problem anyway? |
23:01.46 |
Ralith |
in fact, I'm not sure how you'd represent
those twisted gears with simple operations on primitives--though
I'm no modeler. |
23:01.53 |
kanzure |
would it let the user pick from a list ? is
that how this goes? |
23:02.07 |
starseeker |
twisted gears? like on oil rig
drilling? |
23:02.13 |
Ralith |
I dunno about oil rigs |
23:02.13 |
starseeker |
yeah, those would be a problem |
23:02.15 |
kanzure |
helical gears? |
23:02.18 |
madant |
yeah the idea is to present the user with the
range of options.. |
23:02.26 |
Ralith |
I saw these in a lathe's gearbox |
23:02.28 |
madant |
asking him for either further constraints or
values |
23:02.48 |
starseeker |
runs through the primitives
in his mind... |
23:02.54 |
Ralith |
imagine taking a normal gear with fine teeth,
then twistingthe top and bottom half in opposite
directions |
23:03.03 |
Ralith |
such that the teeth are no longer parallel to
the axis |
23:03.12 |
*** join/#brlcad
Dr_Phreakenstein (n=phreak@216.151.24.198) |
23:03.16 |
starseeker |
Oh, that's less of a problem |
23:03.26 |
Ralith |
they're used to increase contact area between
the gears |
23:03.38 |
kanzure |
madant: so I guess all that I need to do now
is figure out how to make grammar rules for solving constraints
problems |
23:03.48 |
kanzure |
and then get the professor to put up a real
GPL license for graphsynth |
23:03.55 |
kanzure |
and then integrate the code and go from there
:p |
23:04.38 |
madant |
kanzure: i will read up on graphsynth today ..
is there any documentation other than http://www.me.utexas.edu/~adl/graphsynth/ |
23:05.43 |
kanzure |
not really. |
23:05.54 |
kanzure |
there's some related stuff here:
heybryan.org/~bbishop/docs/repo/ |
23:06.04 |
kanzure |
some example graphs of designs: http://heybryan.org/~bbishop/docs/repo/2009-02-20_FS/ |
23:06.14 |
starseeker |
I was thinking of connections like those
visible in page two of this patent: http://www.google.com/patents?id=9WEUAAAAEBAJ |
23:06.32 |
kanzure |
madant: this might be helpful, some
powerpoints? http://heybryan.org/~bbishop/docs/repo/presentations/ |
23:06.37 |
madant |
kanzure: :( i don't have access to windows
:P |
23:06.48 |
Ralith |
starseeker: those count as gears? |
23:06.59 |
starseeker |
It's remotely possible that some application
of tgc subtractions could produce that sort of effect... |
23:07.12 |
kanzure |
madant: there should be some odf presentations
there, and if not, don't worry- I don't have access to windows
either :) |
23:07.16 |
starseeker |
Ralith: they can. It's called a twisted
prism |
23:07.23 |
Ralith |
sounds... arcane. |
23:07.26 |
Ralith |
what're they used for? |
23:07.40 |
starseeker |
Lots of HP printer cartridges have used
it |
23:07.49 |
madant |
nah.. i meant the actual graphsynth
program |
23:07.52 |
Ralith |
I mean I can see in the drill application it
would keep the connections tight |
23:07.57 |
Ralith |
but in gears? |
23:08.08 |
starseeker |
it has the advantage of the turning mechanism
also pulling the component toward the socket |
23:08.10 |
kanzure |
madant: nope. |
23:08.26 |
starseeker |
Ralith: hang on, I'll see if I can find a
picture |
23:08.29 |
kanzure |
btw, I've been able to compile it on
monodevelop under linux |
23:09.12 |
brlcad |
kanzure: how long ago was dns
changed? |
23:09.18 |
brlcad |
(on heybryan.org) |
23:09.21 |
kanzure |
brlcad: it wasn't! :) |
23:09.27 |
kanzure |
I mean, I've never set it to
127.0.0.1 |
23:09.34 |
kanzure |
and it hasn't been changed recently |
23:09.46 |
brlcad |
then you have a bad name server somewhere in
there or an expired name or something.. |
23:09.57 |
kanzure |
hrm. |
23:10.01 |
madant |
hmm.. this is interesting : http://minion.sourceforge.net/ |
23:10.07 |
kanzure |
but then why doesn't the IP address
work? |
23:10.27 |
Ralith |
heybryan.org works for me |
23:10.58 |
madant |
works for me too @opendns |
23:11.12 |
kanzure |
haha, somebody's using opendns? :) |
23:11.16 |
Ralith |
:P |
23:11.25 |
Ralith |
is using level3's
servers |
23:11.32 |
starseeker |
Ralith: not ideal, but
http://www.google.com/patents?id=dGyEAAAAEBAJ&pg=PA37&dq=twisted+prism+gear+helical&source=gbs_selected_pages&cad=1_1 |
23:11.32 |
brlcad |
kanzure: ah, there's your problem |
23:11.37 |
brlcad |
all four of your name servers are
dead |
23:11.56 |
kanzure |
but then what about you being unable to use my
IP to access the same content? |
23:12.08 |
starseeker |
Ralith: element 37a |
23:12.08 |
kanzure |
madant: so, minion spits out a list of
solutions? |
23:12.13 |
brlcad |
looks like afraid.org lost their
name |
23:12.22 |
Ralith |
starseeker: yeah, I saw... isn't that another
interlock? |
23:12.32 |
kanzure |
http://minion.sourceforge.net/gsoc/index.html |
23:12.41 |
starseeker |
Ralith: It's a combination of interlock and
driver |
23:12.43 |
Ralith |
or is 39a a matching gear |
23:12.55 |
madant |
not sure.. just came across it :) scalability
seems to be one of their great aims ;) |
23:13.01 |
starseeker |
yes, 39a is the matching socket |
23:13.05 |
Ralith |
socket? |
23:13.13 |
Ralith |
still not a gear, then |
23:13.29 |
starseeker |
many of these gear types have helical
components, but some of them get rotation only from that twisted
prism |
23:13.36 |
Ralith |
well, either way, it's an interesting
shape |
23:13.42 |
kanzure |
madant: it's weird, they seem to be about
visualization of results? wtf |
23:13.45 |
Ralith |
and apparently one we can't model. |
23:14.18 |
starseeker |
I doubt it. I've been trying to envison some
sort of tgc subtraction which could do it, but I doubt it |
23:14.46 |
Ralith |
would a sweep cover those? |
23:14.58 |
starseeker |
If sufficiently general, yes |
23:15.14 |
Ralith |
I'm guessing sweeps are pretty nontrivial to
implement, then. |
23:15.23 |
Ralith |
seems very useful, though... |
23:15.42 |
starseeker |
yes and yes :-) |
23:16.03 |
madant |
kanzure: ? visualization |
23:16.08 |
kanzure |
in their gsoc page |
23:16.43 |
madant |
ah those are just their ideas page right
:) |
23:18.00 |
madant |
hmm they have a solver independent modeling
language too.. nice .. my idea for libpc is somewhat the same in
the long run.. i mean multiple solvers for the same constraint
network |
23:18.47 |
starseeker |
Well, enough tgc playing with an arb6 MIGHT
produce something interesting, but I don't think it can do
it |
23:18.54 |
kanzure |
shouldn't it just implicitly find all possible
solutions, madant ? |
23:19.22 |
madant |
kanzure: i liked their work too a bit http://www.gecode.org/ |
23:20.16 |
madant |
kanzure: oh the mutliple solvers ? i was just
thinking it might be possible that some solvers are better for some
graphs / problems |
23:20.34 |
madant |
kanzure: i was not implying running a set of
solvers for getting the complete set of solutions |
23:21.43 |
kanzure |
ah, better as in more efficient |
23:21.43 |
kanzure |
okay |
23:23.14 |
madant |
is diybio fun ;) ? |
23:23.18 |
kanzure |
very |
23:23.35 |
kanzure |
scroll down to the bottom of my homepage and I
have a link to some recent work I've been doing with 'sharpie
microfluidics' |
23:23.51 |
kanzure |
i.e. prototyping a "lab on a chip" with a
sharpie and a few spare CDs using surface tension and the capillary
force ;-) |
23:24.42 |
brlcad |
kanzure: huh, did minion make it? |
23:24.55 |
brlcad |
ah, no |
23:25.27 |
brlcad |
going to say, just having two ideas and being
that domain-specific .. nearly impossible |
23:25.53 |
madant |
kanzure: can't find it :P link ? |
23:26.03 |
brlcad |
madant: that does look like an interesting
project, though :) |
23:26.13 |
kanzure |
madant: http://groups.google.com/group/diybio/msg/1197606e3c3dc439 |
23:26.55 |
kanzure |
(there were 40 messages in that thread .. some
of the more recent messages have been about prototyping
microchannels via human hair + sharpie.. but anyway) |
23:27.03 |
madant |
brlcad: yeah.. i will check out their papers
tomorrow .. er.. today ( 5 am here) .. |
23:27.06 |
*** join/#brlcad ``Erik___
(i=erik@c-76-111-12-116.hsd1.md.comcast.net) |
23:27.16 |
brlcad |
pretty bold performance claims |
23:27.17 |
madant |
should get going to try the
Crossfit Murph |
23:27.33 |
brlcad |
not just an order.. orders* faster |
23:27.55 |
brlcad |
license is a non-starter though, maybe they'd
consider changing it |
23:27.59 |
madant |
kanzure: see you around.. will definitely
check out graphsynth today |
23:28.05 |
kanzure |
okay, neat :) |
23:28.13 |
kanzure |
mentions again that he's
considering brlcad for gsoc |
23:28.14 |
kanzure |
heh |
23:28.18 |
brlcad |
heh |
23:28.20 |
madant |
brlcad: maybe.. most probably students or
profs somewhere.. |
23:28.43 |
brlcad |
yeah, http://minion.sourceforge.net/people.html |
23:29.28 |
madant |
ok.. :) be back in an hour.. |
23:29.34 |
brlcad |
would be fantastic if they relicensed as lgpl
or bsd |
23:29.40 |
brlcad |
cya madant |
23:48.01 |
starseeker |
hmm, this is as close to a twisted prism shape
as I can come in a short trial: http://bzflag.bz/~starseeker/twisted.png |
23:48.07 |
starseeker |
not so good |
23:59.02 |
brlcad |
heck of a lot easier with brep :) |
23:59.10 |
brlcad |
but probably not as fast |