what is the name for this type of work?

seed

somewhere in between sense 4 & 5 in Wordnet 2.0:
The noun seed has 5 senses (first 2 from tagged texts)

1. (12) seed – (a small hard fruit)
2. (5) seed – (a mature fertilized plant ovule consisting of an embryo and
its food source and having a protective coat or testa)
3. seeded player, seed – (one of the outstanding players in a tournament)
4. source, seed, germ – (anything that provides inspiration for later work)
5. semen, seed, seminal fluid, ejaculate, cum – (the thick white fluid
containing spermatozoa that is ejaculated by the male genital tract)

dv @ Neue Kathedrale des erotischen Elends
http://www.vilt.net/nkdee

> —–Oorspronkelijk bericht—–
> Van: [email protected] [mailto:[email protected]]
> Namens Jim Andrews
> Verzonden: maandag 14 november 2005 12:06
> Aan: [email protected]
> Onderwerp: RHIZOME_RAW: what is the name for this type of work?
>
> I am wondering if there is a name for a type of work that is
> aleatorically generative (the entities generated are, in
> part, the result of random
> processes) and the idea is to select among the entities as to
> which are meaningful and which aren't. In other words, the
> generation is not intended to produce nothing but gems but,
> instead, invites selectivity amongst the generated entities.
>
> ja
> http://vispo.com
>
>
> +
> -> post: [email protected]
> -> questions: [email protected]
> -> subscribe/unsubscribe:
> http://rhizome.org/preferences/subscribe.rhiz
> -> give: http://rhizome.org/support
> +
> Subscribers to Rhizome are subject to the terms set out in
> the Membership Agreement available online at
> http://rhizome.org/info/29.php
>

Jim Andrews wrote:

> I am wondering if there is a name for a type of work that is
> aleatorically
> generative (the entities generated are, in part, the result of random
> processes) and the idea is to select among the entities as to which
> are
> meaningful and which aren't. In other words, the generation is not
> intended
> to produce nothing but gems but, instead, invites selectivity amongst
> the
> generated entities.
>
> ja
> http://vispo.com
>
>

Jim:

i dont know, you may be have a genetic algorithm in mind:
http://en.wikipedia.org/wiki/Genetic_algorithm

the selection scheme is also called sometimes "survival of the fittest".

******
Dirk:

You may be had a "random seed" in mind. this is usually
an integer number, which initializes an algorithm to generate a certain
sequence of numbers. the generated numbers look like random numbers, but
they are not, since given the random seed you can reproduce them.
so they are called pseudorandom numbers. hence the choice of
a random seed is important for producing "randomness" with a computer.

******

let me know if i misunderstood.

nad

This idea of yours has kept me occupied – do you have a specific
application in mind? Or are you actually working on something similar??

And – at least if you mean to "produce" "art" (why don't just quote
EVERYTHING, Geert) – you seem to imply that art should be
meaningful. That the art object should somehow "carry" meaning.

A parable – the French salons at the turn of the 19th/20th century
– invaded from WITHIN by Duchamps by that which was meaningless to
the point of banality – the pissoir – again at the turn of the next
(I know, I'm pushing this a bit) – Jeff Koons – "Ushering In
Banality".

Isn't that how it goes? The art that really makes a difference
invades our stockades – with meaninglessness??

Geert
http://nznl.com



On 14-nov-2005, at 12:06, Jim Andrews wrote:

> I am wondering if there is a name for a type of work that is
> aleatorically
> generative (the entities generated are, in part, the result of random
> processes) and the idea is to select among the entities as to which
> are
> meaningful and which aren't. In other words, the generation is not
> intended
> to produce nothing but gems but, instead, invites selectivity
> amongst the
> generated entities.
>
> ja
> http://vispo.com
>
>
> +
> -> post: [email protected]
> -> questions: [email protected]
> -> subscribe/unsubscribe: http://rhizome.org/preferences/
> subscribe.rhiz
> -> give: http://rhizome.org/support
> +
> Subscribers to Rhizome are subject to the terms set out in the
> Membership Agreement available online at http://rhizome.org/info/
> 29.php

> This idea of yours has kept me occupied – do you have a specific
> application in mind? Or are you actually working on something similar??

I was writing about work by someone else that fit the description i offered
and wasn't sure if there was a word to describe that sort of work. Am glad
no one wrote back that it was type whatever. Was thinking of Grafik Dynamo
by Kate Armstrong and Michael Tippett (
http://turbulence.org/Works/dynamo ).

> And – at least if you mean to "produce" "art" (why don't just quote
> EVERYTHING, Geert) – you seem to imply that art should be
> meaningful. That the art object should somehow "carry" meaning.

Meaning. Ya.

Well, "meaning is what an explanation of meaning explains," I gather I hold
I carry I hew.

Perhaps meaning is what we construct when we experience things. An ontology
would, then, be a system that permits construction of meaning.

It is a constructed thing, isn't it.

Yet while it's constructed, there are often erm guidelines or rules that we
use in the construction. So that, for instance, when different people read
this sentence, what they think it means will not be identical, but will
share a great deal because language definitions and guidelines and rules are
shared amongst people–though, once again, two people's notions of what
those are aren't identical.

Do I think "art should be meaningful"? Was Duchamp's pissoir meaningful? In
a very unusual way, yes. The bowl is full of it by now, is it not? But the
meaning was not ready-made. It has accumulated over the last century of art
history. In its time, the world was apparently ready for a change, and
Duchamp's pissoir helped flush a few preconceptions about art down the
drain.

It's not so much that I think art itself should be meaningful as this: there
is no type of object to which the activity of constructing meaning is more
intensely directed than the object of art.

> A parable – the French salons at the turn of the 19th/20th century
> – invaded from WITHIN by Duchamps by that which was meaningless to
> the point of banality – the pissoir – again at the turn of the next
> (I know, I'm pushing this a bit) – Jeff Koons – "Ushering In
> Banality".
>
> Isn't that how it goes? The art that really makes a difference
> invades our stockades – with meaninglessness??

Apparently. Just gotta keep em guessing.

Of course "the art that really makes a difference" is many things. Terry
Eagleton called it "a history of barbarism." In the sense that it usually
validates world views from the dominant cultures.

But not always.

My dad used to call the pissoir the throne.

ja
http://vispo.com

> I was writing about work by someone else that fit the description i
> offered
> and wasn't sure if there was a word to describe that sort of work. Am
> glad
> no one wrote back that it was type whatever. Was thinking of Grafik
> Dynamo
> by Kate Armstrong and Michael Tippett (
> http://turbulence.org/Works/dynamo ).

> Yet while it's constructed, there are often erm guidelines or rules
> that we
> use in the construction. So that, for instance, when different people
> read
> this sentence, what they think it means will not be identical, but
> will
> share a great deal because language definitions and guidelines and
> rules are
> shared amongst people–though, once again, two people's notions of
> what
> those are aren't identical.
>

very true. :-) . language can be terribly redundant. I thought with the term "work" you meant programming
work not art work….:-)

but still also if you talk about art work, from your description
one could easily get the picture of an evolutionary or genetic art work,
especially if one interpretes the term "meaningful", as
"having a value, which allows for selection".
i know i am on very slippery terrain
now, but even for kate and michaels work (which is
i guess definitely not evolutionary art in the usual sense) the following is true: if their selection (like of the feeded sentences) is
not completely random (IS IT?), then they must have had some RULES
specifying what they thought was and had a meaning. and they
must have assigned values to what they thought was more or less
meaningful.

so what i mean is that their work can in some sense be interpreted
as an evolutionary artwork, where evolution stops after the
first generation.

but then one could be interested in getting on in evolution, i.e.
like by letting the viewers select their favorites and store
them again in some datbase. So this time the viewers would
decide about whats been meaningful to them. a.s.o.
..and finally one gets the top of evolution…the ultimate
meaningful comic…..:-O

nad

Hi Nad,

Yes, I see that evolutionary or genetic algorithms/works are related to the type of work i mentioned (ie, generative, aleatorical, where each thing generated is not presumed to be a gem, and one is explicitly or implicitly invited to select concerning which are meaningful (to oneself)).

I don't know much about genetic/evolutionary algorithms. Is it possible to give a brief definition?

Concerning the method of selection in Kate & Michael's piece of the images, I don't know whether it's totally random.

I see you're pretty seriously into art, programming, and math at http://daytar.de . Interesting work!

ja
http://vispo.com

Jim Andrews wrote:

> Hi Nad,
>
> Yes, I see that evolutionary or genetic algorithms/works are related
> to the type of work i mentioned (ie, generative, aleatorical, where
> each thing generated is not presumed to be a gem, and one is
> explicitly or implicitly invited to select concerning which are
> meaningful (to oneself)).
>
> I don't know much about genetic/evolutionary algorithms. Is it
> possible to give a brief definition?
>

a genetic algorithm is basically a mathematical model of evolution
(not the kansas schoolbook one, but the darwin one….)
you start out with a giving (parent) set of data, lets say
a set of strings (the strings are bearing the "genetic information"
like in the DNA). then you reshuffle in order to get a new string.
There are several variants of reshuffling (for details see
crossover etc. in wikipedia), like reversing the order of
letters in one string or mixing two different strings (thats
how humans do it…). in this way you get NEW strings.
now you have to choose which one should be your new parent strings.
this choice is often based on rules (choice of the fittest)
sometimes its up to you like e.g. in the genetic art program
Kandid:
http://kandid.sourceforge.net/

so the term genetic ALGORITHM (GA) is more or less settled, the term
genetic ART however not. some people denote with this term
artwork which involves genetic algorithms or at least some
evolutionary principles, in this case they call these kind of
artwork also evolutionary. a very nice example is e.g.
Golan Levins alphabeth synthesis machine:
http://alphabet.tmema.org/

other people mean with genetic art artwork where you fumble around
with real genes, like Eduardo Kac and his green bunny. Here the
evolutionary aspect is not important, the remixing of genes
however is.


> Concerning the method of selection in Kate & Michael's piece of the
> images, I don't know whether it's totally random.
>

i think its random, it would be interesting to implement a rule base GA.

> I see you're pretty seriously into art, programming, and math at
> http://daytar.de . Interesting work!
>
>
yes may be i am a bit too serious about everything…:-)

you have also very nice stuff on your webpage. unfortunately
i couldnt see the shockwave parts, since i am breeding on linux.
i am skeptical wether this problem will be resolved with
adobes new shopping spree. ..


Nad

> a genetic algorithm is basically a mathematical model of evolution
> (not the kansas schoolbook one, but the darwin one….)
> you start out with giving a (parent) a set of data, lets say
> a set of strings (the strings are bearing the "genetic information"
> like in the DNA). then you reshuffle in order to get a new string.
> There are several variants of reshuffling (for details see
> crossover etc. in wikipedia), like reversing the order of
> letters in one string or mixing two different strings (thats
> how humans do it…). in this way you get NEW strings.
> now you have to choose which one should be your new parent strings.
> this choice is often based on rules (choice of the fittest)
> sometimes its up to you like e.g. in the genetic art program
> Kandid:
> http://kandid.sourceforge.net/

In this project, is the "genetic information" the graphics themselves or properties of the graphics?

> so the term genetic ALGORITHM (GA) is more or less settled, the term
> genetic ART however not. some people denote with this term
> artwork which involves genetic algorithms or at least some
> evolutionary principles, in this case they call these kind of
> artwork also evolutionary. a very nice example is e.g.
> Golan Levins alphabeth synthesis machine:
> http://alphabet.tmema.org/
>
> other people mean with genetic art artwork where you fumble around
> with real genes, like Eduardo Kac and his green bunny. Here the
> evolutionary aspect is not important, the remixing of genes
> however is.
>
>
> > Concerning the method of selection in Kate & Michael's piece of the
> > images, I don't know whether it's totally random.
> >
>
> i think its random, it would be interesting to implement a rule base GA.

would it be accurate to say that they leave the 'evolution' to the livejournal images (they draw the images from that database)? the database simply changes over time, though (rather than 'evolving'), so maybe not. i guess the crucial thing, if i understand correctly, is that the change be based on the previous iteration?

> > I see you're pretty seriously into art, programming, and math at
> > http://daytar.de . Interesting work!
> >
> >
> yes may be i am a bit too serious about everything…:-)
>
> you have also very nice stuff on your webpage. unfortunately
> i couldnt see the shockwave parts, since i am breeding on linux.
> i am skeptical wether this problem will be resolved with
> adobes new shopping spree. ..

Someone on this list, perhaps a year ago, pointed out that it is possible to run Shockwave under Linux. I could be wrong, but I think it was via Crossover. You are correct, however, that there is no official Linux version of Shockwave from Macromedia.

ja

Jim Andrews wrote:



> > sometimes its up to you like e.g. in the genetic art program
> > Kandid:
> > http://kandid.sourceforge.net/
>
> In this project, is the "genetic information" the graphics themselves
> or properties of the graphics?
>

the image data is the "flesh" i.e. that what grows from
your genes. so it has to be specified how this "growing"
should work. in kandid you can do this by hand (see documentation)
however in the given examples this has been done already.
the genes or chromosomes are depending on our example they
are for example parameters.

take for example the black and white images in the top row.
they are the attractor of an IFS. which IFS ou take depends
on your parameters (you can actually edit these in the population
panel)


>
> > > Concerning the method of selection in Kate & Michael's piece of
> the
> > > images, I don't know whether it's totally random.
> > >
> >
> > i think its random, it would be interesting to implement a rule base
> GA.
>
> would it be accurate to say that they leave the 'evolution' to the
> livejournal images (they draw the images from that database)? the
> database simply changes over time, though (rather than 'evolving'), so
> maybe not.

yes i understod this also in this way.

i guess the crucial thing, if i understand correctly, is
> that the change be based on the previous iteration?
>

i think they choose the comic images based on some keywords.
so they make a choice. like lets say for each keyword (or a set of)
there is an image. one could try to "optimize" this choice, like
by providing different choices (permuting the assignment of
keyword to comic image). this would give a (parent) set
of comic strips for a given document. now the viewer could
decide which strips are "best" (choice of the fittest)
and one could use this for evolving the choices in order to
get an "optimal cartoonization". it would be interesting to
see what comes out of this, since i could imageine that e.g.
if you choose a fighting scene for the keyword tension
or a electrical tension sign for the keyword tension, then
if you combine this with a comic image for "relationship"
(however this looks like…) then it would be iteresting to
see peoples choice. but may there is also no interesting outcome…

>
> Someone on this list, perhaps a year ago, pointed out that it is
> possible to run Shockwave under Linux. I could be wrong, but I think
> it was via Crossover. You are correct, however, that there is no
> official Linux version of Shockwave from Macromedia.
>
oh yes i didnt check the crossover or wine. there is
a commercial solution, but its 30$ and macromedia
was perpetually announcing linux support….
thanks for the hint :-)

nad

> i guess the crucial thing, if i understand correctly, is
> > that the change be based on the previous iteration?
> >
>
> i think they choose the comic images based on some keywords.
> so they make a choice. like lets say for each keyword (or a set of)
> there is an image. one could try to "optimize" this choice, like
> by providing different choices (permuting the assignment of
> keyword to comic image). this would give a (parent) set
> of comic strips for a given document. now the viewer could
> decide which strips are "best" (choice of the fittest)
> and one could use this for evolving the choices in order to
> get an "optimal cartoonization". it would be interesting to
> see what comes out of this, since i could imageine that e.g.
> if you choose a fighting scene for the keyword tension
> or a electrical tension sign for the keyword tension, then
> if you combine this with a comic image for "relationship"
> (however this looks like…) then it would be iteresting to
> see peoples choice. but may there is also no interesting outcome…

i see. yes, if the "crucial thing…is that the change be based on the previous iteration" then the system is merely stochastic. a stochastic process is one in which the next state depends on one or more previous states. so i guess an 'evolutionary process' is a type of stochastic process such as you have outlined in your last two posts.

and that does actually describe the type of work i originally suggested–though it doesn't fit kate and michael's piece, as you note.

ja
http://vispo.com

Jim Andrews wrote:

>
> > i guess the crucial thing, if i understand correctly, is
> > > that the change be based on the previous iteration?
> > >
> >
> > i think they choose the comic images based on some keywords.
> > so they make a choice. like lets say for each keyword (or a set of)
> > there is an image. one could try to "optimize" this choice, like
> > by providing different choices (permuting the assignment of
> > keyword to comic image). this would give a (parent) set
> > of comic strips for a given document. now the viewer could
> > decide which strips are "best" (choice of the fittest)
> > and one could use this for evolving the choices in order to
> > get an "optimal cartoonization". it would be interesting to
> > see what comes out of this, since i could imageine that e.g.
> > if you choose a fighting scene for the keyword tension
> > or a electrical tension sign for the keyword tension, then
> > if you combine this with a comic image for "relationship"
> > (however this looks like…) then it would be iteresting to
> > see peoples choice. but may there is also no interesting outcome…
>
> i see. yes, if the "crucial thing…is that the change be based on the
> previous iteration" then the system is merely stochastic.



> a stochastic
> process is one in which the next state depends on one or more previous
> states.
no. stochastic is another word for "random". you could have
independent random events. random basically means that you
cant tell which of all the possible outcomes of a process
(the outcomes are called events in probability theory)
will occur. in this case you only know the probability of an occurrence
of an event. (like if you throw a dice, an event would be
"get number six", but you know not for sure if you are going
to get number six and in fact this is INDEPENDENT of what you
had before.)

may be you got confused due to the word "aleatoric" (which you
used before).
it comes from latin: alea -the dice. it COULD mean RANDOM (in
fact in french it does - aleatoire), however in art it may
be something different. the term was used first by
Werner Meyer-Eppler in 1954 in connection with electronic music.
here aleatoric means something which is structured by the artist
in the big picture, but has random fine structure.

> so i guess an 'evolutionary process' is a type of stochastic
> process such as you have outlined in your last two posts.
>

yes you are right an evolutionary process is usually a
stochastic process, but in fact you could have a purely
deterministic evolution.(like if you mutate lets say only
the first two genes in a string this would be a fairly
boring evolution)


Nad