23:13 -!- ewjordan [n=ewjordan@69.182.19.78] has joined #sugar
23:14 -!- zerogate [n=zerogate@177-129.104-92.cust.bluewin.ch] has quit [Read error: 110 (Connection timed out)]
23:18 -!- mokurai [n=chatzill@75-25-130-215.lightspeed.sjcpca.sbcglobal.net] has joined #sugar
23:26 -!- alexlevenson [n=alex@wireless-19-183.media.mit.edu] has joined #sugar
23:26 < bjordan> bemasc: well played today, sir
23:27 < bjordan> jminor, ewjordan, alexlevenson: hi!
23:27 < alexlevenson> Just a heads up, there will be a physics, physics engine, elements etc. meeting taking place here shortly
23:27 < ewjordan> whatup brian
23:27 < alexlevenson> anyone interested in physics engines and the like are welcome
23:27 < bjordan> jminor: saw your post to the assembla message board, really cool stuff!
23:28 < bjordan> multitouch looks like TONS of fun with box2d!
23:29 -!- ewjordan_ [n=ewjordan@69.182.19.78] has joined #sugar
23:29 -!- ewjordan_ [n=ewjordan@69.182.19.78] has quit [Remote closed the connection]
23:29 < bjordan> (everything is more fun with box2d...)
23:29 -!- ewjordan [n=ewjordan@69.182.19.78] has quit [Read error: 104 (Connection reset by peer)]
23:29 -!- ewjordan_ [n=ewjordan@69.182.19.78] has joined #sugar
23:29 < ewjordan_> meh, crummy connection
23:30 < bemasc> ewjordan_: yo.
23:30 -!- jg [n=jg@pool-141-153-205-76.mad.east.verizon.net] has quit [Read error: 110 (Connection timed out)]
23:30 < ewjordan_> bemasc: hi
23:30 -!- ewjordan_ is now known as ewjordan
23:31 < homunq> there will be an unfreeze window for adding features between 8.1.1 and 8.2, right?
23:31  * homunq loses track sometimes.
23:31 < bemasc> homunq: you missed it
23:32 < bemasc> you can still submit special feature requests for 8.2.  You have to convince m_stone that they are useful, essentially
23:34 -!- robsonmwoc [i=c915e243@gateway/web/ajax/mibbit.com/x-0a2dda99584a6bb4] has joined #sugar
23:34 -!- _sj| [n=sj_@wireless-47.media.mit.edu] has joined #sugar
23:36 < nteon> _sj|: here for fizzicks?
23:36 < alexlevenson> Everyone's here for physics, whether you know it yet or not
23:38 < alexlevenson> sj, are you here?
23:39 < mokurai> Physics rules.
23:39 < _sj|> ohhh yea
23:39 < _sj|> would I be anywhere but the fizziks talk?
23:39 < bjordan> ewjordan: bemasc is ben schwartz
23:39 < alexlevenson> naturally
23:39 < mokurai> "If it weren't for Physics and Law Enforcement, I'd be unstoppable!"
23:39  * _sj| waves at physicists bernie, blaketh, cjl, cassidy, daf, danjared, dgilmore, eben, and erikg
23:39 < cjl> hey sj|
23:39 < alexlevenson> Ok, so I'm sure you're all wondering why I've gathered you here today...
23:40 < ewjordan> ah, hi ben (and SJ and all), this is eric
23:40 < _sj|> also ewjordan, hijordan, cmpjordan, m_stone, jminor, *c
23:40 -!- MartyMcFly [n=martin@dslb-088-064-179-070.pools.arcor-ip.net] has quit [Remote closed the connection]
23:40 < nteon> so, personally I would like to see activities to teach physics concepts, or a physics activity that can run files that teach concepts
23:41 < nteon> in addition to any 'playing around'
23:41 < alexlevenson> SJ, this meeting was your idea, so anything you'd like to talk about before we launch into Brian and My part of this?
23:41 < mokurai> Has everybody seen Alan Kay's demo on teaching Galilean gravity?
23:41 < alexlevenson> no I haven't
23:41 < alexlevenson> and as far as physics concepts go that would be a good idea
23:41 < _sj|> there are Awesome things happening in physics simulation, and we should find people excited about this
23:41 < _sj|> help them design things to work with it (in avdance of even having an engine :)
23:41 < bjordan> nteon: so I've written a bit about this on the Wiki / talked with ewjordan about this
23:42 < alexlevenson> we can even graphically show all the forces in play etc
23:42 < _sj|> pick out other cool projects like crayonfizziks and its clones
23:42 < bjordan> nteon: having a common file format would do a lot towards those goals
23:42 < _sj|> and get their contributors excited about physics-on-XO
23:42 < alexlevenson> So I think that brings us to one important topic
23:42 < alexlevenson> we need a standard file format
23:42 < alexlevenson> for a 'physics scene'
23:42 < bjordan> nteon: between ewjordan's http://jbox2d.org engine and the XO's Physics activity
23:42 < _sj|> and have some regular dicsussions with other awesome noncoders like marvin minsky who love playing with these sorts of tools
23:42 < alexlevenson> yes
23:42 < alexlevenson> but
23:42 < alexlevenson> if we have a format
23:42 < bjordan> _sj| +1
23:42 < _sj|> so that FizXO becomes an unstoppable force in the universe of ignorance
23:42 < alexlevenson> then we can have an editor
23:43 < mokurai> http://wiki.laptop.org/go/Presentations/May_2008_Country_Workshop
23:43 < mokurai> http://download.laptop.org/content/conf/20080520-country-wkshp/Video/2008-05-20/13-Beyond-Printing%20(medium).ogg
23:43 < _sj|> alexlevenson, totally
23:43 < alexlevenson> so that between all these cool physics ideas
23:43 < alexlevenson> u can load a scene and do different things w/ it
23:43 < alexlevenson> for example
23:43 < _sj|> alex, are there existing formats we could adopt as an interim standard?
23:43 < mokurai> Too many formats, actually.
23:43 < bjordan> ewjordan: same question
23:43 < alexlevenson> hmmm
23:43 < alexlevenson> well
23:44 < bjordan> http://en.wikipedia.org/wiki/COLLADA
23:44 < mokurai> Matlab, Mathematica, Squeak,...
23:44 < bjordan> er NVM
23:44 < alexlevenson> for one thing since we're currently using python we could just dump all the physics info into a pickle file
23:44 < ewjordan> I think so far every game designer has pretty much made up their own format that incorporates game commands/scripting as well
23:44 < bjordan> http://en.wikipedia.org/wiki/PAL_%28software%29 - supports Box2D
23:44 < alexlevenson> well that's the thing
23:44 < alexlevenson> Box2D does this cool thing
23:44 < alexlevenson> where every body has lots of built in properites
23:44 < alexlevenson> and
23:44 < mokurai> Physics engines for games use radically different models.
23:45 < nteon> bjordan: what kind of information is suppose to be in this file format
23:45 < alexlevenson> well
23:45 < nteon> is it just info on the bodies in the simulation
23:45 < alexlevenson> yes
23:45 < alexlevenson> and
23:45 < alexlevenson> in each body
23:45 < alexlevenson> is userdata
23:45 < alexlevenson> to be used optionally
23:45 < bjordan> right, but box2d based physics playgrounds have remarkably similar models
23:45 < alexlevenson> yes
23:45 < alexlevenson> and
23:45 < bjordan> so it's definitely possible
23:45 < mokurai> In principle, we could have a format specifying the laws of nature to use.
23:45 < alexlevenson> yes...
23:45 < nteon> so there wouldn't be any text, or any way for teachers to package instructions for what to do with simulations
23:45 -!- clajo04_ [n=clajo04_@cpe-72-229-214-63.nyc.res.rr.com] has joined #sugar
23:45 < alexlevenson> but so far everything here is box2d
23:46 < nteon> no way for text labels, or maybe a couple sequential simulations
23:46 < mokurai> Each simulation would need a lesson plan to go with it.
23:46 < bjordan> ewjordan: what do you think in terms of a common format between pybox2d and jbox2d playgrounds
23:46 < nteon> ?
23:46 < bjordan> s/playgrounds/editors
23:46 < alexlevenson> right so
23:46 < alexlevenson> both consist of box2d bodies
23:46 < alexlevenson> so if we can find a way to easily store + retrieve that info independent of language
23:46 < ewjordan> I think it would be relatively easy to set up a simple xml format to get at least most of Box2d's functionality working across the various versions
23:47 < alexlevenson> yes
23:47 < alexlevenson> or
23:47 < ewjordan> (including Flash, C#, and C++)
23:47 < alexlevenson> if we can get a handle to the actually C++ info
23:47 < alexlevenson> does JBox2D run on the C++ box2D?
23:47 < ewjordan> no, JBox2d and the AS3 Box2d are both actually ports
23:48 < alexlevenson> in the same fashion that pyBox2D is a swig binding
23:48 < alexlevenson> oh...
23:48 < ewjordan> (not wrapped)
23:48 < alexlevenson> well there goes that idea haha
23:48 < mokurai> We would need a way to call a number of engines from many programming languages.
23:48 < alexlevenson> ?
23:48 < mokurai> And not just a programming interface, but a way to snap modules together at the user level.
23:48 < bjordan> mokurai: how's it?
23:48 < mokurai> bjordan: How's which?
23:48 -!- newlands [n=newlands@wkstn.wycliffe.ca] has quit [Remote closed the connection]
23:49 < ewjordan> but it would be easy enough to do a loading utility in the C++ version, which would eventually flow downstream to the others, too, assuming there was interest
23:49 < nteon> mokurai: what do you mean snap modules together?
23:49 -!- dcabo [n=dcabo@92-236-36-35.cable.ubr11.dals.blueyonder.co.uk] has quit []
23:49 < alexlevenson> ya same questoin
23:49 < alexlevenson> ewjordan: true
23:49 < bemasc> how does pybox2d currently save scenes?
23:49 < alexlevenson> it doesn't!
23:50 < mokurai> You may have seen the tile-based programming in Etoys. I am thinking about how to snap objects into a world, regardless of their sources.
23:50 < _sj|> Can someone make a [[physics]] page and link from it to a page on [[physics engines]]?
23:50 < _sj|> I'd like to have a list of availale engines and how close they are to running on-XO
23:50 < alexlevenson> it exists!
23:50 < alexlevenson> http://wiki.laptop.org/go/Physics
23:50 < nteon> mokurai:  objects like what?  objects like bodies in the simulation?
23:51 < nteon> s/like/as in/
23:51 < alexlevenson> http://wiki.laptop.org/go/Python_Physics
23:51 < mokurai> nteon: Yes.
23:51 < alexlevenson> http://wiki.laptop.org/go/Physic_Engines
23:51 < mokurai> nteon: But also laws of physics. Fields. The shape of spacetime.
23:51 < _sj|> ohh the capitalization
23:51 < bemasc> mokurai: this is unrealistic.  Perhaps you have not written physics engines before.
23:52 < _sj|> awesome
23:52 < mokurai> bemasc: Indeed not. That is no reason to confine ourselves to what exists.
23:52 < nteon> mokurai: gotcha
23:52 < mokurai> bemasc: I am asking for a design for cross-language OOP.
23:52 < bjordan> mokurai: are you talking about exporting subsets of the physics elements and being able to put them into other scnees?
23:52 < ewjordan> mokurai: I'd agree with bemasc re: shape of spacetime stuff, that's REALLY hard - fields are fine, though
23:52 < bemasc> mokurai: but there is good reason to confine ourselves to what is likely to be accomplished
23:53 < alexlevenson> I agree with bemasc in that last statement
23:53 < alexlevenson> There are lots of very lofty goals we can have
23:53 < mokurai> bemasc: I don't know how hard the problem really is, and I don't think you do either.
23:53 < ewjordan> for educational purposes a good (static, at least) E+M simulator would be fantastic
23:53 < alexlevenson> though, there are some things we'd like to talk about for now
23:53 < bemasc> I'd like to talk about specific educational goals, and work from there
23:53 < bjordan> bemasc: I think that sounds reasonable.
23:54 < bjordan> ewjordan: absolutely
23:54 < mokurai> bemasc: But I'll put the question away until I can ask Alan Kay.
23:54 < alexlevenson> So for educational goals: I have one
23:54 < bjordan> ewjordan: _sj| alexlevenson and I have talked about adding e&m functionality to our activity/elements/(box2d?)
23:54 < alexlevenson> yes
23:54 < bjordan> e&m = electricity and magnetism
23:54 < bemasc> bjordan: how is box2d's support for long-range forces?
23:55 < alexlevenson> via the fact that box2D supports adding a force / torque / impulse
23:55 < ewjordan> bemasc: I'll answer that - there pretty much is none built in
23:55 < alexlevenson> what is meant by long range force?
23:55 < bemasc> my feeling is that e+m simulation is best kept separate from rigid-body, for the moment
23:55 < nteon> is there any suport for the weak and strong nuclear forces, as well?
23:55 < ewjordan> bemasc: you've got to do it yourself
23:55 < bjordan> ewjordan is very well acquainted with the Box2D internals
23:55 < alexlevenson> That's where elements comes in!
23:56 < bemasc> ewjordan: what do you have in mind for electrostatics?
23:56 < eben> bemasc++
23:56 < mokurai> bemasc: Educational goals should include Galilean and rudimentary Newtonian gravity, mechanics, optics, basic concepts of quantum physics (without the formulas and calculation), waves, rudimentary Special Relativity, and fundamental concepts of General Relativity without formulas and calculation.
23:56 < ewjordan> bemasc: if I was going to do fields, I'd overlay a grid and do a loosely coupled field simulation on that
23:56 < bemasc> mokurai: my feeling on this is that simulation is not really very useful for teaching physics
23:57 < alexlevenson> can i say something real quick?
23:57 < mokurai> ewjordan: Stephen Wolfram has some interesting models of fields in spacetime using cellular automata.
23:57 < bemasc> (For the record, I have a degree in Physics from MIT, and am pursuing a PhD in biophysics)
23:57 < alexlevenson> I have a different goal
23:57 < alexlevenson> (sort of)
23:57 < bjordan> bemasc: it's very effective in demonstrating mechanical concepts
23:57 < mokurai> alexlevenson: Jump in.
23:57 < alexlevenson> the incredible machine is not a 'let's learn physics game'
23:57 < bemasc> bjordan: but not as effective as a machine
23:57 < alexlevenson> it's a critical thinking game 
23:57 < bjordan> bemasc: DScrofani used it my year, it was pretty nice
23:57 < bjordan> bemasc: plus we could play with the engine at home (theoretically), which you can't do with physical demonstration machines
23:57 < alexlevenson> So it's more of a puzzle solving game, that happens to rely on a physics engine
23:57 < mokurai> bemasc: Physicists use simulations all the time. Why not students?
23:58 < bemasc> bjordan: interesting.  All I'm saying is, real pulleys are cheap, and simulated pulleys are a poor substitute.
23:58 < bjordan> bemasc: if they show information on the forces acting on them real-time?
23:58 < alexlevenson> My point is, teaching physics is good, but there are other aspects of a physics engine
23:58 < alexlevenson> i.e. building simple machines
23:58 < bjordan> bemasc: I think there's an opportunity for a lot of data to be visualized during these simulations
23:58 < alexlevenson> less about physis
23:58 < alexlevenson> more about thinking
23:59 < bjordan> bemasc: which is something more than is possible during physical classroom demonstrations
23:59 < mokurai> alexlevenson: We also need to work on how to think about the history and philosophy of science.
23:59 < bemasc> alexlevenson: I agree, but perhaps we should distinguish "game engines" from "physics engines".  When used in that context, it's a game engine, and its veracity as a simulation of reality is not critical.
23:59 < bjordan> mokurai: thank you for that list, adding it to our wiki discussion page
23:59 -!- ffm is now known as ffm|walk
23:59 < ewjordan> I think for stuff that's dead simple to implement in software, it's fine to do so.  My nervousness starts acting up when we start thinking about simulations of quantum effects and relativity and all that, because that's a huge (PhD level) undertaking, esp. to integrate with the rest of an engine
23:59 < bjordan> http://wiki.laptop.org/go/Physics
23:59 < mokurai> How did Galileo think about gravity with the primitive tools available? No clocks, for example.
--- Day changed Fri Jul 11 2008
00:00 < alexlevenson> well, i'd like it realistic, but just because it has physics it's not about learning physics by default
00:00 < alexlevenson> he used chimes
00:00 < mokurai> How did Kepler think about orbits, as opposed to Newton?
00:01 < alexlevenson> before I leave
00:01 < alexlevenson> please look at our features / requests page
00:01 < mokurai> swjordan: I have examples of QM and GR simulations I can point to later. They aren't that hard.
00:01 < bjordan> ewjordan: that is a good point
00:01 < ewjordan> alexlevenson: I think I'm agreeing with you here mostly, if I understand correctly that you're saying that a realistic goal is to put together a fun and flexible physics playground rather than to shoot for the type of thing that a physics teacher would use to demonstrate formulae and stuff like that
00:01 < alexlevenson> for some topics we'd like added to elements
00:01 < bjordan> mokurai: are they open sourced?
00:02 < alexlevenson> ewjordan: yes
00:02 < alexlevenson> ewjordan: i have two ideas for projects
00:02 < alexlevenson> 1) playground game
00:02 < mokurai> bjordan: Many are Open Source.
00:02 < alexlevenson> 2) incredible machine game
00:02 < alexlevenson> number 1 for learning physics
00:02 < alexlevenson> number 2) for puzzle solving
00:02 < mokurai> We also need to think about what to do as real physics, not depending on models and simulations.
00:03 < bjordan> mokurai: interesting, do they demonstrate these concepts in a way that a child could understand?
00:03 < ewjordan> mokurai: don't get me wrong - I understand you can simulate this stuff (my senior thesis was simulating EM waves going into a black hole, and I've done a lot of SR simulation, too), the problem is to integrate it with a whole engine
00:03 < nteon> alexlevenson, I then think you need to create an example activity that can be used to TEACH physics, beyond playing with it
00:03 < mokurai> Galilean relativity is easy to demonstrate with almost no apparatus.
00:03 < nteon> mokurai, what do you mean?
00:03 < mokurai> We should point out to children that they can observe the parabolic path of a stream of water, which the Greeks could have done but missed completely.
00:03 < alexlevenson> My point is that is a very different goal
00:04 < alexlevenson> not a bad one just not the one I have at the moment
00:04 < bemasc> mokurai: so point it out... but you don't need a simulation
00:04 < ewjordan> alexlevenson: IIRC, someone at the Boston Game Jam a couple years back started an incredible machine game, any idea if that ever progressed?
00:04 < bjordan> mokurai: is the laptop a best way to accomplish this?
00:04 < alexlevenson> they gave up for some reason
00:04 < mokurai> bjordan: They were not written for children. But we can use the same techniques to write something for children.
00:05 < bjordan> mokurai: we could
00:05 < ewjordan> mokurai and bemasc: as a FWIW, Box2d might be getting fluids support (similar to what Phun has) soon, I've been working a bit on that
00:05 < mokurai> A note on method: Alan Kay demonstrates that we can teach children to create simple simulations, and then have them record phenomena that follow that model, with Record or Measure.
00:05 < bjordan> mokurai: I am more interested in the near term using this Box2D engine to its full potential
00:05 < alexlevenson> OK I have to leave, if you get a chance please discuss: breakable objects (polygon decomp?), particles (sand / liquid), motors, springs, and simple magnets
00:05 < alexlevenson> Thanks!
00:05 < alexlevenson> Bye!
00:06 < bemasc> mokurai: I hope you will send an e-mail to the list detailing these simulations.  I'm especially surprised that you think there's a GM simulator that has any real educational value, given that the only ones I'm aware of require supercomputers to produce results of any interest.
00:06 < bjordan> mokurai: allowing kids to make these demonstrations and play with the physics of mechanics in a controllable and analyzable environment
00:06 < ewjordan> bemasc: GM?
00:06 < bemasc> err.... GR
00:06 < ewjordan> ah
00:06 < bemasc> curses
00:06 < mokurai> bemasc: What if you do the simulation of conic sections on the computer, and of motion that is constant on one axis and uniformly accelerated on another (parabolic)? And then get them to observe the parabola in real life?
00:07 < mokurai> bemasce: Is that not more powerful than dealing with these ideas separately?
00:07 < bemasc> mokurai: I draw a distinction between graphing and simulation.  Certainly, I think 2D and 3D plotting software can be very useful
00:07 < mokurai> bjordan: There is room here for many approaches. Please do.
00:07 < ewjordan> "real" GR is out, IMHO - even the simulation experts have taken over 20 years to get a stable simulation of black hole collisions on supercomputers computing for months on end
00:08 < jminor> hi
00:08 < bjordan> jminor: hey!
00:08 < jminor> sorry my meeting ran late.
00:08 < alexlevenson> jminor: hi! I wanted to talk to you but I have to leave, maybe some other time
00:08 < alexlevenson> bye!
00:08 -!- alexlevenson [n=alex@wireless-19-183.media.mit.edu] has left #sugar []
00:08 < mokurai> ewjordan: I am not talking about the dynamics of neutron fluids. I am talking about demonstrating geodesics in spacetime for point particles.
00:08 < bjordan> jminor has worked with Elements and made an awesome multitouch demo for MakerFaire with a box2d testbed
00:09 < ewjordan> mokurai: that might be reasonably doable
00:09 < mokurai> ewjordan: Anyway, the supercomputer simulations have been done, and some of the Flash videos are publicly available.
00:09 < jminor> here is the thing I made: (with youtube video): http://lux.vu/blog/2008/07/01/physix/
00:10 < mokurai> ewjordan: They turn up on Astronomy Picture of the Day every few months.
00:10 -!- dfarning__ [n=dfarning@75-121-148-130.dyn.centurytel.net] has quit ["Ex-Chat"]
00:10 < bemasc> I would like to suggest that we start by talking to teachers, ideally teachers in deployments, and ask them what they would like to have
00:10 < ewjordan> mokurai: that's definitely true; the methods, however, are far too advanced for us to actually implement, though
00:10 < cjl> Sorry, coming in and slow in catching up, I'd like to point you to the bottom of this page  http://wiki.laptop.org/go/Science  UNESCO developed microscience course on electricity with students and teachers guides.
00:10 < cjl> The idea I love is every activity being paired with at least some textual content and in the context of science, nothings much more constructionist than lab practicals.
00:11 < bemasc> a 2D N-body gravity simulator with a decent interface that runs in real time might not be hard to whip up; it would be good to ask teachers if they would use it
00:11 < mokurai> bemasc: The problem with that is that they have no idea what is possible.
00:11 < bemasc> mokurai: and you have no idea what is useful
00:11 < jminor> so we have to bridge the gap by iterating
00:11 < mokurai> bemasc: But when we can show them what exists, then we can ask them what else they would like.
00:12 < bemasc> mokurai: I agree. Let's talk about what exists.
00:12 < jminor> There are several of playground / toybox applications that people have done very interesting things with.
00:12 < nteon> okay, this might be a little late, but:  1) what is the agenda, and 2) what do we want to take from this meeting
00:12 < mokurai> ewjordan: Well, the XO is about as slow as a Cray-1, so I have considerable hope for the future when we can move out faster hardware.
00:13 < nteon> lets sort that out quickly please
00:13 < ewjordan> hehe, true enough
00:13 < bemasc> nteon: agreed
00:13 < mokurai> cjl: Yes, I was saying earlier that each simulation needs to have a lesson plan with it.
00:13 < jminor> for example: a radial engine simulation/illustration in phun: http://youtube.com/watch?v=kcluNJ_HFCM
00:13 < nteon> or if i missed it tell me waht we're doing ;)
00:13 < bjordan> nteon: yes, thanks ( _sj| seems to have disappeared )
00:13 < bjordan> I'd really like to come out of this with...
00:13 < mokurai> bemasc: I'll collect some examples of what exists to prepare for the next round in that discussion.
00:14 < bemasc> mokurai: the physics page would be a great place to accumulate links to existing educational physics software
00:14 < mokurai> nteon: What can we simulate in existing software that will run on an XO? How do we build topic lessons around that, and a curriculum around the lessons? What next?
00:15 < cjl> um, physics and physics_collections seem merged on wiki, I thought they used ot be separate
00:15 < nteon> anyone else:
00:15 < jminor> I'd really like to come out of this with... brainstorming of ideas collected on the wiki
00:15 < bjordan> 1. Who is still interested in helping to develop Elements, which has applications in physics teaching, simple game creation for the XO
00:15 < jminor> from there we could consolidate ideas into more focused development
00:16 < bemasc> I know that in high school I played with a van der Waals noble-gas simulator that ran at 20 fps on a Pentium II 266
00:16 < bemasc> that was pretty cool
00:16 < mokurai> nteon: What other software can we bring to the table by relicensing or creating it? How can we get all the different kinds of software to work together? 
00:16 < bjordan> 2. If you have ideas/suggestions Physics, a Box2D based playground, I'd like to hear them
00:16 < bemasc> bjordan: I haven't heard about elements?
00:16 < bjordan> 3. I want to hear more about jminor's use of multitouch with respect to box2d
00:17 < bemasc> bjordan: we don't have multitouch...
00:17 < mokurai> "Most of biology is chemistry, and 99% of chemistry is physics."
00:17 < cjb> yes, it sounds like there's interesting gen2 work, at least
00:17 < jminor> bjordan: Elements is certainly interesting as an enabling technology, but needs a specific app (or set of apps) to help focus its development.
00:17 < bemasc> what is elements?
00:17 < mokurai> I run a screensaver that shows models of 3D biological molecules.
00:17 < nteon> bemasc: +1  i think we have enough to talk about without multitouch
00:17 < mokurai> There are interactive periodic tables.
00:17 < cjb> dsd_: we might work on my keyboard again
00:17 < jminor> Elements is a python wrapper around Box2D
00:17 < bemasc> ah.
00:17 < bjordan> 4. Can we make a common format and integrate it with Eric's JBox2D engine so that physics classrooms that already have computers can help develop fun physics scenes for the XO as well
00:18 < mokurai> There are numerous math, physics, and chemistry programs in Free Software.
00:18 < jminor> nteon: I agree, multitouch is cool and all, but there are more valuable goals between here and there to focus on.
00:19 < bjordan> nteon, bemasc: jminor had a recent interesting experience with this, and I'd like to hear about the technicals of box2d with multiple inputs, but we can save that for another time (on #olpc-physics)
00:19 < mokurai> We also need to turn the problem around. What concepts need to be taught in grades 1-6? Which ones do we know how to simulate, and what still needs work?
00:19 < bemasc> cjb: how can we contact teachers in deployments?  This is something that I keep running up against.
00:19 < ewjordan> bcjordan: just a quick question, are there now two different physics projects going on related to the XO, or are you continuing with Elements, or scrapping it and starting something else?
00:19 < jminor> bjordan: a common file format (or converters) would go a long way to help consolidate the great work being done with Phun, iPhysics, etc.
00:19 < bjordan> jminor: exactly
00:19 < bjordan> jminor: ewjordan and I have talked about how this can be accomplished
00:20 < mokurai> bemasc: Anybody who speaks Spanish can get on OLPC-Sur and other lists. We can also raise the question on some of the English lists.
00:20 < jminor> For example, an SVG -> physics converter would help people build lesson plans, tutorials, examples without us needing to make a very fancy editor.
00:20 < cjl> If wiki is to be used for organizing this, some improvement in use of category tags on wiki would be helpful http://wiki.laptop.org/go/Category:Physics http://wiki.laptop.org/go/Category:Science
00:20 < nteon> bjordan, jminor: then can someone draft up an example for this file format or start writing a spec?
00:20 < bemasc> I have to say that I think a common file format for this is not likely to work beyond a few programs that share the same box2d engine, and even then it seems that they may be quite different.
00:20 < nteon> I don't see what needs to be talked about on that, besides specifics, and I don't want ot talk about specifics of a file format right now ;)
00:20 < mokurai> cjl: Category:Lesson? Category:Simulation?
00:21 < bjordan> jminor: I'm not sure if Phun or iPhysics necessarily would play nice (not OSS) specifically, but let's do this for at least the underlying engines, Box2D, JBox2D, pyBox2d/Elements, as3box2d
00:21 < jminor> getting geometry from one program to another could be very useful.  the simulations would be different, so some mechanics would not work from one to the other (think taking a carefully balanced machine to the moon)
00:21 < bjordan> ewjordan: Elements is the backend for Physics
00:21 < nteon> bjordan: can you draft up a version of the file format that works on the XO for starters?
00:21 < cjl> mokurai: not picky about names, just more consistent use
00:22 < mokurai> cjl: I think we do need finer-grained categories in some places.
00:22 < bjordan> nteon: yes, I'll work on that and activity saving this weekend and next week
00:22 < jminor> I have a simple xml file format in Physix.  Someone else (ken I think?) has a similar one in Elements.
00:22 < jminor> I can document it and put it on the wiki as a proof of concept.
00:22 < cjl> bemasc: mokurai working on getting GoogleTrans gadget on OLPC-Sur archive page to make more accessible to lang-en  "reverse localization"
00:22 < bjordan> jminor, ewjordan: would you like to be involved with the development of this?
00:22 < nteon> jminor: awesome
00:23 < ewjordan> bjordan: sure, I could definitely help out
00:23 < bemasc> cjl: I speak enough spanish; I'll ask on that list.
00:23 < jminor> yes
00:23 < ewjordan> jminor: I think that might be useful (putting up the xml format as a proof of concept)
00:23 < bjordan> jminor, ewjordan: this is a format, it's important we get this right and that it doesn't block anyone out from implementing it
00:24 < bemasc> most standards arise by one person creating it, and others adopting it
00:24 < nteon> bemasc: agreed.  
00:24 < jminor> We should also consider the fact that Phun is very cool and may become open source.  If it does then we may or may not want to switch to a different simulation engine.  If this happens, then a file format would help to carry over some work from one to the other (with caveats of course)
00:24 < bemasc> I recommend creating a sane save-format for Elements/Physics, and then publicizing it.
00:24 < ewjordan> bjordan: I'm with you; I can at least make sure we get as much of Box2d's functionality in
00:25 -!- gregdek [n=gdk@nat/redhat/x-8d39ce39cc63ed89] has quit [Read error: 110 (Connection timed out)]
00:25 < nteon> okay, can we move onto the next agenda item, which is...
00:25 < nteon> bjordan: (thats your cue)
00:25 < ewjordan> jminor: I don't know about that, from Phun's FAQ: Is the source code going to be released?
00:25 < ewjordan> Nope.
00:25 < bjordan> jminor: that is something to consider, though last I heard it wasn't going to be open sourced
00:25 < bjordan> ewjordan: ah, right
00:26 < jminor> Oh, last time I looked it said "maybe".  We could still try to convince the author.
00:27 < jminor> next agenda item?  how about ways to draw polygons fast on the XO... I was frustrated with both gtk and SDL's performance
00:28 -!- gregdek [n=gdk@nat/redhat/x-ae895f7d321976ee] has joined #sugar
00:28 -!- santiago [n=santiago@maisel-gw.enst-bretagne.fr] has joined #sugar
00:28 < bjordan> jminor: a good point, it is definitely worth approaching the author for help
00:28 < ewjordan> is pygame's sprite handling any faster?
00:28 < bemasc> jminor: yeah, this is a big issue.
00:28 < jminor> does anyone have (successful) experience asking someone to open source their software?
00:28 < cjb> if you install sysprof, we can profile where the time's going
00:28 < bjordan> jminor: yes, it is slowing it down a lot
00:29 < bjordan> jminor: you http://wiki.laptop.org/go/Physics_engines/Speed_tests
00:29 < ewjordan> jminor: might be tough, in the case of Phun, as it's a thesis project and a lot of times PhD candidates want to hold back their best stuff for the sake of having an edge...I'm not sure, though
00:29 < nteon> bjordan: are those tests on the xo?
00:30 < bjordan> nteon: yes
00:30 < jminor> I haven't tried the latest Physics-0.2 on real XO hardware (only vmware).  What sort of fps do you get with a screen full of objects?
00:30 < bjordan> er.. jminor you might have a better idea
00:30 -!- cjl [n=chatzill@12.44.50.248] has quit ["ChatZilla 0.9.83 [Firefox 3.0/2008052906]"]
00:30 < bjordan> jminor: we haven't done any speed tests/improvements yet
00:31 < bjordan> jminor: we've just been using elements as is
00:31 < mokurai> bemasc: Talk to Antonio Ognio <antonio.plug@gmail.com> about the Peruvian list with 800-1000 members.
00:31 < jminor> We did speed tests between box2d and chipmunk (another simulator) but not too much with just raw drawing speed.
00:32 < jminor> Are there any other non-physics activities that do a lot of polygon drawing?
00:32 < bemasc> jminor: fast drawing is a big problem on XO, 2D and 3D.  My feeling is that the solution is a fast software OpenGL implementation, running at low resolution and piped into XVideo for scaling.
00:32 < bjordan> jminor: if you look at that graph, Box2D being simulated without drawing ran much more reasonably
00:32 < bemasc> someone sent some code to the mailing list a month or so ago to do this, but no one has put enough effort in to really get it working
00:32 < jminor> oh scaling would be very handy.  are there code examples out there?
00:33 < bemasc> jminor: sort of.  Every video player on the XO uses the hardware scaler.
00:33 < bjordan> jminor: it is simulating at a ridiculously high resolution
00:33 < nteon> bemasc: how would you feel about cairo with its opengl backend?
00:33 < cjb> (I don't think we ship mesa currently.)
00:33 < bemasc> nteon: it doesn't make sense.  The XO doesn't have any hardware acceleration for OpenGL.
00:33 < jminor> I bet other game-related activities could benefit from drawing speed tests too.
00:34 < bjordan> jminor: absolutely
00:34 < nteon> bemasc: then why do software opengl for 2d?
00:35 < bemasc> nteon: the 2D/3D distinction is not as hard as it appears
00:35 < mokurai> Has anybody talked about having one higher-speed computer per classroom so that teachers can run simulations in realtime? It would be better if students could do it, but a compromise might be worthwhile.
00:35 < jminor> anyone have contacts in the demo scene?  like scene.org?  we could get someone who knows their assembly language to write us a wicked fast polygon renderer.
00:35 < bjordan> nteon: with low resolution --> scaling it could speed up a lot
00:35 < ewjordan> for flat shaded 2d polys, someone could probably whip up a quick C rasterizer, right?
00:36 < bemasc> ewjordan: you can do the math: the XO has a 1 megapixel screen, and a memory bus that's only so wide...
00:36 < jminor> bemasc: do you have a link to more info on XVideo?  I'm not sure what that is.
00:37 < ewjordan> bemasc: how are things done now?
00:37 < bemasc> jminor: xvideo is how X displays video.  It's an interface provided by the GPU driver to scale video on the GPU.
00:37 < bemasc> xvideo can also do yuv<->RGB conversion, but that's not important here
00:38 < jminor> which activity would have code that does a 2x scale in it that I could dig into?
00:39 < bemasc> jminor: it's all hidden much deeper in the libraries.
00:39 < bemasc> gstreamer's xvimagesink, for example
00:40 < bjordan> there's no way of doing any scaling/getting any speed benefit using pygame?
00:40 < jminor> does the XO's X11 have a way to pick a different video mode - or does the display not support that?
00:41 < bemasc> bjordan: pygame uses SDL, and for some reason SDL doesn't use XVideo for scaling except when showing video
00:41 < bemasc> jminor: it's unimplemented.  It's coming.... eventually.
00:41 < bemasc> but xvideo scaling would get us the same result, without the nasty mode-switch
00:42 < jminor> well maybe we can add a vote for that feature in the roadmap
00:42 < bemasc> I thought about patching SDL, but it looks hard
00:42 < ewjordan> it seems to me that most of the performance drain at least with the physics apps is due to the number of shapes drawn, not the total pixel throughput, no?
00:42 < nteon> bemasc: I have a gut feeling that cairo with some compositing would be fast enough, but no numbers to back it up
00:43 < _sj|> cjl, physics (activity) and physics (collection) are different.  
00:43 < bemasc> nteon: there's no reason it would be faster
00:43 < _sj|> alex, you should put your ideas for projects up on the physics page
00:43 < bemasc> anyway, those graphs are surprisingly encouraging
00:43 < bjordan> bemasc: absolutely
00:44 < bjordan> bemasc: I realized just last week how big a hit the drawing was...
00:44 < bjordan> *how big a hit in performance
00:44 < bemasc> I'll stop complaining now; if we can do 20 fps with 25 objects, that's great.  The graph shows box2d doing 25 fps with 50 objects, which is amazing.
00:44 < bjordan> _sj|: moving [[Physics]] the day I send out an email saying that's the URL for hte activity... dirty ;)
00:44 < nteon> bemasc: I don't know how they're doing it now with PyGame/SDL.  I think alex mentioned they might be redrawing every shape, each frame, instead of just translating and rotating an image surface.
00:44 < jminor> alternately we can figure out a reasonably pleasing wireframe mode
00:45 < nteon> (and then stamping it)
00:45 < ewjordan> bemasc: box2d should be seeing some pretty significant performance improvements over the next few months, too, as Erin has promised to do some cache-friendliness optimizations
00:45 < bjordan> jminor: a good idea
00:46 < jminor> time for another agenda item?
00:46 < bjordan> Ok, let's wrap this up. I think the consensus is we should do speed tests of different drawing engines for Elements on the XO
00:46 < _sj|> bjordan, whoops :-)
00:47 < _sj|> bjordan, it's now a disambiguation page, a sure sign of  success...
00:47 < bjordan> This weekend + next week ewjordan, jminor?, and I will be adding a Box2D-friendly physics style format to elements/physics and jbox2d
00:48 < bemasc> jminor: http://lists.laptop.org/pipermail/devel/2008-June/015302.html
00:48 < bjordan> myself and AlexLevenson will definitely be around #olpc-physics
00:48 < bjordan> and I invite you all to come there and idle like it's your job!
00:48 < jminor> bemasc: sweet!  I'll give it a go
00:49 < bemasc> I'll send an e-mail to olpc-sur requesting opinions and floating a few ideas
00:49 < bjordan> cjb: what was the name of that diagnostic tool you had mentioned? (to figure up what was eating time)
00:49 < cjb> sysprof
00:49 < ewjordan> oh, Alex had wanted us to discuss: breakable objects (polygon decomp?), particles (sand / liquid), motors, springs, and simple magnets - bjordan, if he needs any advice on that, give him my e-mail
00:49 < bjordan> thanks
00:49 < cjb> ask tomeu or riccardo about how to get it running, they just fid
00:49 < cjb> did
00:49 < bjordan> alright
00:49 < jminor> I'll post some info on my XML format to the wiki today or tomorrow.
00:50 < bjordan> ewjordan: ah right, yes, we want to add concave polygon decomp to elements
00:50 < ewjordan> bjordan: alright, I can def. advise on that
00:50 < bjordan> ewjordan: and particles may be rough given current drawing speeds, but definitely looking in to
00:50 < jminor> bjordan: there is working code for concave polygon decomp in Physix
00:51 < bjordan> jminor: yeah, I saw, it looked great :)
00:51 < jminor> Could springs be just an added joint type in box2d?
00:51 < bjordan> ewjordan: you have it in jbox2d as well, correct?
00:51 < ewjordan> bjordan: nope, not jbox2d yet (haven't ported my decomp code to Java)
00:51 < bjordan> ah
00:51 < bjordan> which OSS engines is it in now?
00:51 < bjordan> or is it just your post on the forums
00:52 < ewjordan> jminor: the distance joint actually works as a spring if you set the frequency and damping
00:52 < jminor> ewjordan: cool - we should try to get that added to the box2d documentation
00:52 < ewjordan> sorry, I mean frequencyHz and dampingRatio
00:53 < jminor> i should go soon.
00:53 < bjordan> oh, and _sj|, jminor, ewjordan, nteon, et al: we're considering running an OLPC Physics Jam (Physics Game Jam?) in late August in Boston
00:53 < ewjordan> jminor: absolutely, I'll ask Erin if he minds if I update that
00:53 < jminor> If anyone has questions about Physix, feel free to post comments on my blog, or contact me directly.
00:54 < ewjordan> jminor: just a quick q, what are you using to do the decomposition right now?
00:54 < bjordan> jminor: it was great to talk with you, we should all meet again soon
00:54 < jminor> Physics Game Jam sounds awesome.  Travel for me in August is pretty unlikely though.
00:55 < jminor> ewjordan: the polygon decomp I used was found here:  http://www.flipcode.com/archives/Efficient_Polygon_Triangulation.shtml
00:56 < bjordan> jminor: have you heard from Erik006 or Chris or kne lately?
00:56 < jminor> I also turned off a couple of b2Asserts() in box2d to allow for thin triangles.  I put some details on this in the README.txt in the source for Physix.
00:56 < jminor> bjordan: not a peep
00:57 < jminor> This was awesome - thanks for organizing :)
00:57 < ewjordan> jminor: ah, ok - yup, those asserts are a real hassle with decomposition code, that's the main problem I ran into
00:57 < bjordan> jminor: alright, do you have any thoughts on assembla vs. the OLPC git?
00:57 < bjordan> jminor: right now we have elements as a portion of Physics' git repository, and we have made a few changes
00:57 < jminor> ewjordan: I didn't see any adverse effects from turning them off though.
00:57 < bjordan> jminor: I figure eventually we can push our changes upstream
00:58 < ewjordan> jminor: yeah, continuous collisions might work wrong with them off, at least for tiny triangles; we're hopefully going to figure out a better solution soon, though
00:58 < _sj|> jminor : can you add yourself and describe your work via projectdb.olpc.at ?
00:58 < bjordan> jminor: if you do anything cool with Physics and want to push it to git, we can get you set up to do so
00:58 < _sj|> in case you don't have hardware...
00:58 < jminor> bjordan: assembla vs git - no opinion so far - I haven't used either one enough to develop a preference.
00:59 < bjordan> _sj|: good call
00:59 < bjordan> jminor: do you have an XO currently?
00:59 < _sj|> also, demo scene contacts would rock
00:59 < jminor> _sj|: will do
00:59 < jminor> bjordan: yeah, get it working and then worry about pushing upstream
01:00 < cjb> bjordan: nice not to proliferate where stuff is hosted
01:00 < jminor> bjordan: yes I have 2 G1G1s (one for me, one for my daughter)
01:00 < cjb> I like that I can usually go to dev.l.o/git and find the right code
01:00 < bemasc> the olpc git server system is really very convenient
01:00 -!- alemartinazzo [n=alexandr@189-46-199-83.dsl.telesp.net.br] has joined #sugar
01:00 < bemasc> and ties in nicely into the translation server, etc.
01:01 < jminor> yeah, this was my first intro to git - I mostly like it.
01:01 < jminor> I should go.  It was great chatting - feel free to keep me in the loop via email.
01:01 < jminor> cheers!
01:01 < bjordan> _sj|: demo scene contacts meaning contacting people who can make good demo scnees?
01:01 < bjordan> jminor: thanks for coming!
01:01 < nteon> jminor: thanks, see ya
01:02 < _sj|> btw, please see
01:02 < _sj|> http://wiki.laptop.org/go/Physics_meetings/July_10%2C_2008
01:02 < _sj|> and add yourself as an attendee
01:02 < _sj|> and any extra notes/thoughts
01:02 < _sj|> bjordan : no
01:02 < _sj|> bjordan: there is a 'demo' community
01:02 < _sj|> that makes amazing things with say 10k of disk
01:02 < _sj|> beautiful generative graphics, audio, &c
01:03 < bemasc> generally speaking, they are incredibly good at making weird, beautiful graphics that animate fluidly on a Commodore 64
01:03 < ewjordan> bjordan: see http://scene.org
01:03 < bemasc> they do this by cutting every single corner, and coming up with awesome approximation algorithms
01:03 < bemasc> which are great for abstract visuals and terrible for physics
01:04 < bjordan> cjb: point taken
01:04 < bemasc> personally, I think we should find some to do the XO bootup sequence, but I don't know how useful they are for us
01:04 < _sj|> bjordan, thanks for reminding about the physics jam
01:04 < _sj|> consider the weekend of aug 23/24
01:04 < bjordan> brain, thanks for remembering of it last minute
01:05 -!- mchua is now known as mchua|away
01:05 < bjordan> ewjordan: what day does becket end
01:05 < ewjordan> bjordan: hm, not sure, I'll ask mom when I get home
01:05 < bjordan> bemasc, _sj|: I really like that idea... let's see what they can do with an XO :)
01:06 < _sj|> ewjordan, do you know anyone involved?
01:06 -!- unimauro [n=unimauro@200.121.162.155] has quit ["Saliendo"]
01:06 < ewjordan> (btw, if I get cut off, that means my 2 hours of Starbucks prepaid internet has run down, and it was good talking with all of you - I'll be in touch)
01:06 < ewjordan> _sj|: no, not in demoscene stuff, alas
01:06 < _sj|> hmm.  I bet _bernie does
01:07 < ewjordan> does OLPC have any bare-metal types on staff?
01:08 < bjordan> _sj|: did jminor indicate he might?
01:08 < bemasc> ewjordan: yeah, but they're busy doing things that are actually useufl
01:08 < ewjordan> haha, ok
01:09 < ewjordan> alright, I gotta run guys, good to talk to all of you
01:09 < bjordan> OFFICIAL MEETING ENDING LINE ----___--_- Physics -- Marvin Minsky tested, Marvin Minsky approved.
01:09 < bjordan> bye eric