Sunday, December 14, 2003

godel, escher, bach

Monday, December 01, 2003

center of mass and density function is something i read, perhaps not useful at all for the project.
but at the moment i havent got more info on that.

i am compiling the voronoi paper, its really hard, especially last week with end of fasting month but we all got sick. it has been about a month since came back from jakarta, and all the time we has coughs. last thurs we brought haniya to hospital and she'd been diagnosed chest infection, although mild but caused by virus. that explained why we all got the coughs. so husband and i we took antibiotics too. although the coughs are better, but i sneeze alot already 2 days now and not feeling any better, abit dizzy because after sneeze always has to blow the nose. apparently those virus want to go out now.

so i suggest, if you feel not better today, should go to GP. its this time of year when virus everywhere.

back to my voronoi paper, considering the health, i should finish mid of Jan if i want to present poster for MIT conf. and your bubble story should fit the first part of the paper, since what i got now abit boring with voronoi definition / the program. i'm planning to put frei otto in hillier's framework of spatial configuration.

so perhaps the paper completed will have headings like this:
1. real bubble and digital bubble
2. representation: bubbles and spatial configuration
3. to grow configuration; explain the role of voronoi (why voronoi?) --> we initially try to find free-form generator
4. how to (voronoi recipe)

by the way, i read Zachary Mason Programming with Stigmergy: Using Swarm for Construction (Artificial Life VIII, MIT Press 2002.
this literally looking for a program so swarm can build (hence word Construction), presented like a recipe. only 5 pages pdf, abit hard to imagine if not run it because all the comments already summarised what happened (author's opinion/observer point of view?).
quote from the paper:
Pheromone driven-rules implement the kind of continuous cellular automaton, defined by changing reaction of potentials, on top of a discrete one, defined by the motion of the agent swarm.

so should we skip pheromone-like condition and straight for CA?

perhaps the paper completed will have headings like this:
1. real bubble and digital bubble
2. representation: bubbles and spatial configuration
3. to grow configuration; explain the role of voronoi (why voronoi?) --> we initially try to find free-form generator
4. how to (voronoi recipe)
ramas thesis is on the swiki. you should read IL 39 (institute of lightweight
structures - its all about unplanned settlements and bubbles


the point about bubbles is that they want to equalise the pressure . bibber
bubbles have more pressure because they have more air in. the optimal close
packing of deformable spheres is dodecahedral is all spheres have the same
amount of air in. this close packing is of course a 3d voronoi diagram.

the question is what is the air modelling? ie why are some bubbles bigger
than others. rama has a 3d voronoi that you can play with i think.

as for your other q. i dont get why we need to calculate the centre of
bubbles ( i thought that the voronoi cells were based on the agents? anyway
there are a range of algorithms to determine the centroid of a polygon, the
simplest one being a kind of averaging system - same for 3d. much like
calculating the centre of gravity of an irrecular body - just calculate
moments . i would need to look this up.

BTW i was hunting about for images for the mit paper and came across several
papers on using agents to assess spatial organisation not all of them from
CASA. have a look on google for "ISOVISTS" and other topics (mr peponis is
still working away you will find)

hope your feast and congregation went ok - i am currently staying at home
because of a 'fluey throat so our todays meeting would be off anyway.


p