Juhi, Marcus, Shuang, Nicole

From KokkugiaWiki

latest pin-up

Image:Skywalk-Page_1.jpg Image:Skywalk-Page_2.jpg Image:Skywalk-Page_3.jpg Image:Skywalk-Page_4.jpg

Some of Roland's main comments were: - 1. To use more agents in the path generation code, as it is lacking complexity right now 2. If components form these kinds of bridge structures to rethink the nomenclature of the project as it is no longer a skywalk 3. Also that the scale of the components is too large at present


COMMENTS DEC 8
I agree with Roland's comments on the project. It is good to see the level of control you now have over the agents, however I think you need a higher population of strands (agents) to generate sufficient complexity (probably 5 times as many strands are required). It would add a level of complexity to your project if your components were able to respond to each other (eg respond to the other components or trajectories) so that they are not simply being arrayed along the curve, but instead generate a non-linear relationship. Likewise the regularity of the columns seems a little artificial, perhaps these should respond to context or density of the strands rather than being entirely periodic.



project development

Nov 1st Uploads Image:SkywalkNov1_1.jpg Image:SkywalkNov1_2.jpg COMMENTS NOV 5
to differentiate the mesh object i recommend using a very low polygon model which you control through the manipulation of its control points - if this is carefully setout it should be effective. but note that you should use the least possible number of faces knowing that you are able to manipulate the geometry later such as smoothing and refining the mesh once the geometrical manipulations are complete. you will need to work out the index of the mesh vertex or grips so you can have precise control - perhaps write a script to label an object with text dots so you can quickly visualize what is happening when you are initially designing the behavior of the component.

to find a neighbor segment without refering to the agents own line: i recommend joining all the line segments for a each agent into a polyline ie. one polyline per agent. then loop through each cv and check for the proximity of cvs on other curves. you just need a conditional statement to ensure that you are not checking against the current curve - this will end up looking something like:

if Not j = k Then

if j is the counter for looping through the current curves and k is the counter for looping through the curves to check against. i think this sort of conditional response is potentially very useful, however i think you need to explore the full potential of this local interaction - currently it is overly simplified.

i am concerned about the geometry you are rending of the bridge. i think you need to be more serious about establishing a geometry which responds to and emerges from your generative (agent) system - you have reverted to a very normative geometrical instantiation of a complex process. i would urge you to return to your earlier ambitions of generating an intensity of articulation from the formative process.


Scripting the Skywalk: scripting sequence- Image:Skywalk3a.jpg Image:skywalk-4.jpg

project proposal

Stage 1: site Analysis Image:skywalk1.jpg

Stage 2: case studies and proposal Image:Skywalk2a.jpg

Views