Scripting Salon Minutes, July 26, 2008

We had another salon at Think Coffee yesterday, so I am writing to summarize. This was the best-attended salon yet (by far). Present were Che-Wei, Adam, Frank, Sean, Caleb, Mike, Ken, Christine, and myself. Karen (ITP friend of Che-wei’s) came at the very end, when I think the salon had basically dissipated. Think worked pretty well as a venue; as the group grew we were able to grab one of the central couches and pull up enough chairs.

So, the minutes:

A recently-graduated scripter from Pratt: and also (this last is a great example of an online portfolio, it seems to me).

We discuss Frank’s need to prototype his bike rack design for NYCityRacks (he was one of ten finalists). Somebody suggests a fiberglass manufacturer called Seal. And Sean suggests looking into guys who do motorcycle customization (if I remember correctly).

We discuss what it would take to develop a Rhino plugin that does material mapping properly, since this is our #1 gripe with using Rhino as a primary rendering solution (with whatever render plugin you like — at Pratt it seems like it will be V-Ray this year). There is no clear answer to this yet — I have to look into it — but there’s a free Rhino SDK available for download so I will get on it.

We have an extensive discussion on polygon modeling vs. NURBS modeling. Rhino Grasshopper comes up a few times — as a way to address some of our concerns about Rhino’s slow, “one-way” (“fire and forget”, perhaps?) modeling capabilities. Well, Rhino has fabulous powers of precision, in geometry and gesture (creative actions, so to speak), but the lack of good history and editing, compared to the spontaneity of Max’s or Modo‘s polygon powers, makes it a pretty cruel thing to foist on students as their main app for 3D thinking. Or at least, I think so.  In response to Adam’s and my griping about Rhino, and my particular wondering as to whether there are not just certain ways of working in Rhino that are cleaner and more adaptable than others, Che-wei mentions that he always models in wireframe — i.e. drawing only curves — before generating any surfaces.  I am going to try to apply this rigorously in the coming semester.  Che-wei also mentioned an Audi TT tutorial that he thinks is especially effective for learning effective modeling.

As the evening progresses we realize that Maya does, and has done, basically everything we talk about. But Maya is out, for some reason. So sad!!!  Can we blame the perhaps overly obscure design processes of the late 1990′s studios at Columbia for this failure of the architecture community to just use the full powers of this tool, and get the guys who make Maya to add a few useful features that would have just made it click?

Thanks to Mike’s and Caleb’s presence, especially, there is a bunch of web-dev talk. Apparently some cool sites (which I will check out shortly) include FFFound (okay I am already hooked), Everything Everywhere (can’t find this, actually), and We also discussed an idea I’ve had for some time of creating a visual database and map of (primarily) digital drawing techniques. As references people mentioned RISD’s recent “web video interface” and the early GSAPP website by Corey Clarke which apparently had a kind of tree-like network view.

Next up, some hardcore software and hardware discussion, courtesy of Caleb. He mentions:

Datarush, by Sun Microsystems… a better kind of relational database, apparently, and very easy to use.  Still in development.

GPGPU… a framework for taking advantage of new multi-core graphics processors, if I understand correctly.  Apparently (perhaps this is a separate point), the upcoming Flash version will actually be able to use some graphics card capabilities, thus permitting fairly hardcore 3D work in Flash.

OpenGL3 is coming soon, apparently, and will re-legitimize this oh-so-sweetly open graphics library.

Caleb is excited about COLLADA, an open graphics file format that is apparently becoming more and more feature-rich. E.g. there’s a Flash library, PaperVision, which has a COLLADA parser.  Ummm, check out this Papervision site, actually.  Wow.

We discuss briefly how there’s not a lot of attention being paid in the geek-architect community to scripting at the level of the rendering itself (e.g. shaders, raytracing).  I.e. we make funky geometry, and we also do a lot of pixel-based manipulation of 2D images, but we don’t play that much with the magic that creates specific patterns of pixels from the geometry and lights, etc.

Answers:  Sunflow, including P5 Sunflow for Processing, and Zbrush.  But I’m not satisfied.

On the fuzzier, friendlier end of graphics work, apparently Che-wei saw Gil Akos do a project where max/msp and Processing were running in parallel and talking to each other. But (and I was surprised to hear this!), Che-wei is moving away from Processing to OpenFrameworks, an apparently amazing set of libraries for C++.  Yes, C++!  Holy smokes!  He says the OpenCV library is especially amazing (that’s “computer vision”, not “control vertex”).

Adam is going to Siggraph in a couple of weeks (or whenever it is), and intends to attend a “design and computation” session they’re going to have there.  There will be discussion of folding, apparently.  Sounds fun — please report back Adam!

Lastly, for Che-wei (and anybody who likes to hike around NYC), I present to you Breakneck Ridge.  It’s best to get a topo map from these guys, I have one and it’s very helpful.

Oh, and Caleb thought you guys would dig this recent Radiohead video made with LIDAR, and the making of it.  IN fact to really get the beauty of the lazer-survey technology in the video you should download the 86MB version from Radiohead’s website.

Okay, that’s all.  Attendees, please comment with links if you have other stuff you wanted to share!

2 comments to Scripting Salon Minutes, July 26, 2008

Leave a Reply




You can use these HTML tags

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>