Robert Cumming

TOP:  A screenshot of a scan of (what I believe to be) a photocopy of a photograph titled “Decorator Test” by Robert Cumming (1974).  Printed in LAICA Journal #1, found via Experiments in Print.

BOTTOM: Robert Cumming’s “Of 8 Balls Dropped From The Peak Roof, 2 Fell To The North, 6 Fell To The East”, also 1974.  Found via SFMoMA.

First Experiment With “Mechanical Turk”

The completed tasks, not 20 minutes after publishing them!  Click on the image for full-size.

An initial experiment (my first) with Amazon’s Mechanical Turk, a service that allows workers around the world to complete simple online tasks.  While most of the tasks are weird surveys or spamming attempts that pay between $0.05 – $0.25 per job, I decided that paying a reasonable fee would be fair and more likely to be completed.  I asked participants to visit this site and determine the color they saw when their eyes were closed looking at a bright white screen.  Prepaid for 100 people to complete my task, I assumed it would take a week or two.

I took a shower and came back to my computer to find… 100 responses!  In less than 20 minutes the whole project was finished, to my astonishment.

The first 100 responses – not all what I would call “accurate”

A few thoughts on the process:

  1. The results were almost entirely quality – about 93% took more than 90 seconds to complete the task and returned useable values.  The others apparently did not read the instructions very carefully and returned color names like “pink”, “black”, and “rainbow”.  While interesting, I wanted completely objective answers.  Mechanical Turk allows you to reject jobs that don’t meet your criteria so I did; a few sent back emails saying they were upset that I had rejected their one minute of work, but I think that’s likely par for the course with this system.
  2. It is clear that not all the values are “good” – I can’t imagine the scenario that someone sees bright blue.  In the future, I’ll likely proof the data first.
  3. Mechanical Turk will return your results as a CSV file, which is very useful.  It also includes lots of great “extra” data, including time spent finishing the task and the exact time the task was completed.
  4. I am very much thinking of setting out another batch to be completed.  While $100 isn’t cheap ($110 actually, since Amazon charges a fee), the data was fast and quality.  I have been considering a sliding scale so that the price goes down over time, making those who answer early get paid more, those that wait less.

If you are feeling egalitarian and want to help the project without getting paid, you can head over to this page and email me your results.

Mental objects

Great quote from an article in Frieze by Edward Allington.  First published in Issue 92, June-August 2005; the article contains notes which are not online.  Must find that McEvilley quote…

Most pieces of Conceptual art are, to quote Thomas McEvilley, objects: ‘Theories, of course, are things; they are what Edmund Husserl called Noemantic objects, that is, mental objects. Every thought or concept is an object, and every object has form or aesthetic presence (what does a centaur look like? an angel?). There is, in other words, an aesthetics of thought with its own styles and its own formalism.’

CubeSat NSF grant = art

A CubeSat high above the Earth
A CubeSat high above the Earth

The National Science Foundation has issued a RFP for researchers interested in the CubeSat.  According to Wikipedia, the CubeSat is “a type of miniaturized satellite for space research that usually has a volume of exactly one liter (10 cm cube), weighs no more than one kilogram, and typically uses commercial off-the-shelf electronics components”.  The satellites are then ganged onto existing rockets allowing for low-budget space research.

Needless to say, when I first read about the CubeSat a few years ago I thought that conceptual art and data-driven art projects truly need to get launched into space.  Anyone have a few spare centimeters?

Image via: Wikipedia