Blog archive: Division (2009)

From my very first steps with processing, I’ve always been fas­ci­nated by the divi­sion of space, vol­umes and sur­faces. In ret­ro­spect, many con­structs can be seen as explo­rations of this theme. To my delight, a lot of the con­cepts I dab­bled with in igno­rance are actively explored in the field of algorithmic architecture, or to use a term coined by Kostas Terzidis, algo­tec­ture. On a whim, I bought a pam­flet by Benjamin Aranda & Chris Lasch, Tooling. It was like lift­ing a peb­ble and find­ing a mountain.

I had this feel­ing only once before. When we moved to another house at age 10, I dis­cov­ered an aban­doned chem­istry set in the garage. I hap­pily wasted hours with the strange mate­ri­als inside and as a high­light cre­ated a pur­ple froth­ing expanse of foam stain­ing the skin a sickly brown for days after. Surprisingly many ran­dom com­bi­na­tions resulted in mal­odor­ous vapors of amaz­ing vari­ety to delight fam­ily and neigh­bors. Fortunately I ran out of chem­i­cals before I ran out of ideas. With their usual unyield­ing com­mon sense my par­ents declined to restock the var­i­ous vials and pow­ders. My chem­i­cal days over, I resumed my favorite passtime, sit­ting in the mid­dle of a big pile of Lego mak­ing ping-ping noises.(Much like a mete­orite cool­ing in its impact crater after a super­heated entry).

Anyway, yes, Processing, exactly… Years after my short stint in chem­i­cal horse­play, I was aston­ished to dis­cover that chem­istry was not merely a toy, it was a sci­ence, an indus­try, a pos­si­ble life­time devo­tion. It is a path I never fol­lowed, I went into physics instead.

Processing is my sec­ond chem­istry set. With its low thresh­old it helps me imag­ine math­e­mat­i­cal and geo­met­ri­cal what­ifs. What if I take a cube and cut it up in smaller cubes, and do it again and again? What if I repeat­edly divide a sur­face? What if… We play end­lessly on the crowded beach of com­plex­ity, self-similarity and other fash­ion­able con­cepts. The sand­cas­tles we build are but tiny, pale reflec­tions of the mastodonts dreamt of by others.

I have fun build­ing my sand­cas­tles but real­ize all too well it’s just play. Complexity is easy and inter­est­ing but rarely usable. And nowhere is this more evi­dent than in mod­ern, parametric/algorithmic archi­tec­ture. A design con­cept is one thing, get­ting a viable build­ing out of it is evi­dently some­thing else alto­gether. Somewhat iron­i­cally I feel, this is referred to as the prob­lem of sim­plex­ity, in this con­text the reduc­tion of an aes­thet­i­cally pleas­ing form to a struc­tural model. E.g. the ran­dom sub­di­vi­sion of cubes as explored in the early D5LV results in a self-similar swarm-like col­lec­tion of boxes. However, these boxes often inter­sect, float freely, have no eas­ily extractable rela­tion to each other. Translating this into some­thing real, like try­ing to get a STL out of it, would require exten­sive work, far more than the orig­i­nal piece. Complexity attracts all the atten­tion, with spec­tac­u­lar imagery, huge promises and boast­ful lines. In the back­ground sim­plex­ity qui­etly toils away, its achieve­ments wrongly attrib­uted to its loud brother.

In a way, there’s a dan­ger inher­ent to Processing and the gen­er­a­tive con­cepts it so read­ily allows access to: we might come to believe that it’s easy. Judging from ques­tions I get, many already believe this. From absurdly com­plex projects to hun­dred­fold increases in scale, the easy ques­tions are often impos­si­ble to answer. (The hard ques­tions are often quite easy.) I see sim­i­lar­i­ties with the pro­fil­er­a­tion of CGI indus­trial design mod­els on the inter­net. Whole blogs thrive on pretty images of impos­si­ble to real­ize but beau­ti­ful con­cepts, reduc­ing design to illustration.

Anyway, enough with the ram­blings, I’ve got two small con­structs, some­what related to this post. My cur­rent fas­ci­na­tion with the more prac­ti­cal side of gen­er­a­tive algo­rithms leads me to rethink much of my past con­structs. The afore­men­tioned cube sub­di­vi­sion is fun but lacks a clear rela­tion­ship between its ele­ments, beyond the obvi­ous self-similarity. Using my mesh code-in-progress, I sought an iter­a­tive sub­di­vi­sion algo­rithm that wouldn’t suf­fer from this kind of dis­so­cia­tive behav­ior. The end result should be inter­est­ingly com­plex but still be a sin­gle con­nected solid with­out self­in­ter­sec­tion. I came up with two entries for this self-contest (actu­ally three but I’m keep­ing the 3D Voronoi frac­tal for a bit later). They’re con­cep­tu­ally very sim­ple so don’t expect any­thing rev­o­lu­tion­ary (or new), click on the images to go to the applet.

Note, this was in 2009, in 2022 applet means tiny apple and this type of code is not runnable online without rewriting it in something that is, so instead the links point to zip archives of offline code for processing 3.