Live data from Hacker News

Deterministic Pixels

darrennewton.com

1–10 of 13 posts

Re: Deterministic Pixels

#4
post #3

Nice. In 2005 the artist Allan McCollum made software to design 31 billion shapes, with the intention to have at least one shape per person for a long time. Not usable the way identicons are, but the shapes are pleasing. http://en.wikipedia.org/wiki/The_Shapes_Project

>made software

On the artist's webpage there's following note in the footer:

"Note: Contrary to some errors made in certain press articles, McCollum's Shapes are not "generated" in a computer with an invented or scripted "program." Every shape is laboriously created by the artist using Adobe Illustrator — a common, everyday graphics program — by drawing little parts, cutting and pasting the parts into bigger parts, then cutting and pasting those parts into even bigger parts, and so on, and keeping track according to a written protocol, to insure against repetitions. The first exhibition of the project, in 2006, took around two years to complete."

Anyway, thanks for a very interesting link.

EDIT: More info on the tedious labor involved: http://allanmccollum.net/amcnet2/album/shapes/shapesworkshee...

Re: Deterministic Pixels

#5
post #3

Nice. In 2005 the artist Allan McCollum made software to design 31 billion shapes, with the intention to have at least one shape per person for a long time. Not usable the way identicons are, but the shapes are pleasing. http://en.wikipedia.org/wiki/The_Shapes_Project

>made software On the artist's webpage there's following note in the footer: "Note: Contrary to some errors made in certain press articles, McCollum's Shapes are not "generated" in a computer with an invented or scripted "program." Every shape is laboriously created by the artist using Adobe Illustrator — a common, everyday graphics program — by drawing little parts, cutting and pasting the parts into bigger parts, t…

So, they are generated in a computer using an invented program, but the mid level graphics API is implemented on his hand, instead of wholly inside the computer.

Re: Deterministic Pixels

#6
I feel like I've been seeing Racket come up more often recently. It's interesting to me that it seems to be winning the scheme wars when there are other excellent distros out there (such as Chicken and Gambit). Has anyone shipped any software with Racket? I'd be curious to hear about it in real world scenarios (e.g., long-running processes and deployment).

Re: Deterministic Pixels

#7
I've always liked autometically generated avatars in theory. But in practice they never seem that useful because there's nothing recognizable in the image for me to subconsciously recognize.

Secret had an interesting take on this. They had a limited set of simple icons, like boats or rockets, and then dynamically colored them. They were easy enough to recognize so patterns became more apparent.

Re: Deterministic Pixels

#8
It would be nice seeing something like this at self-signed https sites generated from the public key of the given site (a visual public key fingerprint), instead of scaring me away with a huge warning.

Re: Deterministic Pixels

#9
post #3

Nice. In 2005 the artist Allan McCollum made software to design 31 billion shapes, with the intention to have at least one shape per person for a long time. Not usable the way identicons are, but the shapes are pleasing. http://en.wikipedia.org/wiki/The_Shapes_Project

>made software On the artist's webpage there's following note in the footer: "Note: Contrary to some errors made in certain press articles, McCollum's Shapes are not "generated" in a computer with an invented or scripted "program." Every shape is laboriously created by the artist using Adobe Illustrator — a common, everyday graphics program — by drawing little parts, cutting and pasting the parts into bigger parts, t…

Wow. That sure looks like a process ripe for automation.

Re: Deterministic Pixels

#10

I've always liked autometically generated avatars in theory. But in practice they never seem that useful because there's nothing recognizable in the image for me to subconsciously recognize. Secret had an interesting take on this. They had a limited set of simple icons, like boats or rockets, and then dynamically colored them. They were easy enough to recognize so patterns became more apparent.

We should just pull a random cat image from google, chances are very likely that each user will get a new one.
Post reply on HN