Rhea Myers

Tag: Generative Art

  • draw-something: almost there

    1987 images since Friday evening. 2048 shortly…

    I’ll put some of the good ones on Flickr when it’s finished.

    Technorati Tags: , ,

  • draw something In Progress

    200709082343

    It’s up to 700 images. A 400 Mhz G4 takes a couple of minutes to create a drawing. I may just run it to 1000 rather than 2048, which will probably take until some time tomorrow.

    So far the images that work are reasonably diverse. The images that don’t work tend to be just a few dots or lines on a plain background. The colours aren’t working as well as they seemed to in the test application so I’ll have to find out why that is.

    Technorati Tags: , ,

  • draw-something 2048 times

    I’ve just started a version of draw-something running that is going to generate 2048 images. I wonder if it will be finished by Monday?

    I’m going to go through the images and see what works and what doesn’t. I’ll modify draw-something to make it more likely that it will produce images that work and less likely to produce images that don’t.

    The first results indicate that just dots on a plain background really doesn’t work. 😉

    Technorati Tags: , ,

  • Cold Reading, Misdirection and NLP

    ….are all potentially fertile areas for Artificial Intelligence research.

    Mediumship, conjuring and marketing all require an understanding and manipulation of human beings that contains a great degree of knowledge and understanding of the human condition. AI implicitly requires this as well, at least in its strong form or in forms that would recognize the Turing test.

    So if you would make a machine fool a person into accepting that it has knowledge, look at how people fool people into accepting that thet have knowledge…

    Technorati Tags:

  • Like That: Arrow Sketches

    200708250010

    200708250013

    200708250017

    Back to “Like That” after the CC Ironies.

  • Dynamic Artistic Practice Systems

    To produce “dynamic artistic practice” AARON would have to create AARONs. It would have to do what Harold Cohen does; set immediate and long term goals, create systems, evaluate results, reflect on its achievements and work this meta-knowledge into its goals and systems.

    This sounds like the sort of goalpost-moving that has always plagued AI. If a computer can play chess then chess playing must not be a sign of intelligence… But “dynamic artistic practice” would be relatively simple to achieve given a system of axiomatic graphical microdomains and a Lisp-style function composition system.

    The microdomain approach is the one that Douglas Hofstadter and FARG took: find a simple task that is representative of an aspect of creative behaviour and then analyze it in depth.

    AARON, like any painter, just pushes pigments around in 2D. It's a historically sufficient aesthetic domain in which to create and evaluate constraints or axioms.

    It may even generate historically novel techical practice, Margaret Boden's “h-creativity” from The Creative Mind. But this wouldn't be necessary to satisfy the requirements of “dynamic artistic production”, as Damien Hirst's appropriations show.

    So I'm not making a ridiculous demand of an AARON-like system. This behaviour could be shown by a fairly simple system (like Copycat), although to be aesthetically as well as conceptually satisfying it would need to be more complex.

    (From an answer to a comment by Yaxu below.)

    Technorati Tags: ,

  • draw-something fun

    One of the motivations for draw-something was my memory of my younger self’s desire to make a program that was better at art than they were. This was an ambition born of cyberpunk literature and media and of the utter incomprehensibility of perspective instruction books to me at that time.

    draw-something is not better at art than I am. Art Computing courses should teach their students on day one the difference between static technical competence and dynamic artistic practice. AARON is part of a dynamic artistic practice but is not a producer of dynamic artistic practice, Harold Cohen is. As a child I had a book featuring an illustration of AARON at work. It made a lasting impression.

    draw-something is not better at drawing than I am. draw-something does imitate my drawing style at the micro level very well, it captures my “hand”. Or I have captured my hand in it. I have tried to do so in a way that models the physical failings of my drawing style rather than just wobbling lines after they have been generated.

    What draw-something is much better than me at is just have fun with shape and colour. Opponents of Artificial Intelligence will You don’t have to be an opponent of Artificial Intelligence to will point out that draw-something isn’t actually having fun. But performatively speaking something is. The fun is there. Possibly it’s me, sublimating this into draw-something. But then I wouldn’t have that precise kind of fun otherwise, and I’m not having it as draw-something is not-having it.

    Both proponents and opponents of AI, and most right-thinking artists and critics, will cringe at the mention of something as nebulous as “fun”. Which makes me think I might be on to something. 😉

    Technorati Tags: , , , ,

  • minara And GPL 3

    minara uses a few small but very useful utility functions from the ttn-pers-scheme project, which seems to be GPL 2 only with no timescale set for upgrading to GPL 3.

    I’ll email the author to ask if the next release will be GPL 3. If so, I will hold off on making minara GPL 3 until then. If not, I will re-implement the functions. I know that’s not very “clean room”, but hopefully it will be OK.

    I’m not meant to be doing any more work on minara but I really would like to fix the outstanding bugs and do a 0.0.1a release at some point.

    Technorati Tags: ,

  • draw-something Project Plan

    Here are my current plans for draw-something. They are subject to change.

    1. Relicence as GPL 3. Done.

    2. Ensure that colouring code works properly.
    – Check that it matches Cohen’s description.
    – Do this in chordinator?

    3. Tune up “Interesting Points” system.
    – Make 100 images and try to code away unpleasant features.

    4. Finish basic simple & large -> complex & small figure plane system.

    5. Move to more complex “figures & developments” system.

    6. More figure types.
    – Currently there are polylines, polygons and hulls.
    – Add points, circles, arcs, linesets, circlesets.

    7. Switch to integer maths or at least whole-pixel offsets.
    – Error based line drawing?

    9. Relational groups.
    – Literal mirror, scale, rotate of core skeleton.
    – Conceptual groups e.g. “5 points”, “all same size/area”.

    10. Tagging scheme.
    – Tag interesting points to get clusters, relations.
    – Tag each development to guide later developments.
    – Basic implimenteation e.g. plist of symbols.

    A. Solve presentation problem.
    – Do it like current AARON or like Copycat, drawing screen & log screen.
    – Ltk seems best bet.
    – Dump drawing and commentary as event log to (e.g.) Java based player.

    B. Make sketch program & development program & colour program?
    – UNIX pipeline.
    – Make each component simpler.
    – Like Kazushi’s systems.

    C. Representational imagery?

    D. Pixmap/crop-based system or faster vector-based hit testing? (See 7.)

    E. Stop using CVS? Just release tarballs?

    Technorati Tags:

  • rheart is now GPL 3 licenced

    rheart (including draw-something) is now GPL 3 licenced in SourceForge CVS.

    draw-somethng-carbon has been removed due to bitrot and also because I cannot work out how the LLGPL interacts with the GPL. The old version is still available in the older release tarballs, and may even work. 😉

    Technorati Tags: ,