Re: breeder approaches to iDrama

["Mark 'Kamikaze' Hughes" <kamikaze@kuoi.asui.uidaho.edu>]

Mon, Jun 06, 2005 at 01:21:13PM -0700 in <42A4B039.6020808@indiegamedesign.com>,
Brandon J. Van Every <vanevery@indiegamedesign.com> spake:
>Mark 'Kamikaze' Hughes wrote:
>> There are software-generated poems
>>and stories, and I've been entertained by them in a few cases, which is
>>more than I can say for many meat-based writers.
>Any URLs?

  The one I liked best was a random mystery story generator, but it
doesn't seem to be on the web anymore.  It used madlibs and essentially
an "elimination game" (those things where John, Sue, and Bill all had
soup, but only one had tomato...) to assemble the story, usually a page
or two long.  It was certainly no threat to James Ellroy, but was more
advanced than a Clue game, and more creative than some published mystery
short stories.

  Huh.  I bet I could recreate it as IF in GameScroll.  That's actually
a really good idea...

>> Consider <http://www.random-art.org/>, which is often surprisingly
>>good.  I've used many of them as desktop backgrounds over the years.
>That's interesting!  I'll have to try out something like this for 3D models.
>> All of these still have an artist, of sorts: the person who wrote the
>>program.
>I would also choose to function as the breeder / selector of "good" 
>artworks.

  That, too.  The earliest runs of random art were not very good, but
once they had been selected on for a while, some really interesting
pieces appeared.

>> But your belief that every single work of art must be manually
>>generated by an *ARTISTE* is clearly false.  The assertion that modern
>>impressionists are not artists is more than just false, it's openly
>>insulting, and you need to think long and hard about why you'd say such
>>a horrible and monstrous thing, why the notion of people making art that
>>is simply there to be pleasant to look at is offensive to you.  I can't
>>stand the work of Neil Diamond, yet I'd never assert that he isn't an
>>artist.
>I must admit, I had that reaction too, I just didn't choose to express 
>it quite that way.  Many artists develop a thick skin to other people's 
>baloney, and I'm no exception.  I think some points of view about 'Art' 
>are not worthy of much comment, being as prejudiced and removed from 
>practical reality as they are.  It goes with the territory.  Critics 
>hated the Impressionists once upon a time; now they're canon.

  I have no problem with people who self-deprecate themselves with "oh,
that's not art...", but given my definition of art, they're incorrect.
Saying someone else's art is "not art" is a massive insult.  It's most
commonly used by people who are very insecure about their own artistic
abilities, and want to attack anyone they perceive as a threat for being
better than them.

  Just because the art world is full of venemous little toads who hate
each other, is no reason for civilized people to tolerate that behavior
in public.

  Perhaps it's related to the academic politics paradox: The smaller and
more meaningless the stakes, the dirtier the politics.

>> The goal of the interactive fiction/drama/storytelling movement (which
>>this list exists to serve) is to produce storytelling entertainment
>>generated or mediated by software.  If you're not interested in that,
>>it's probably not going to be of any use to you, and assertions that
>>it's impossible and not art are probably not going to be of any use to
>>anyone here.
>Hey, if it's not so tough to breed 2D and 3D images, can we breed 
>stories as well?  In the same software framework!  The UI would present 
>the breeded story snippets, the breeded artwork, the breeded audio 
>clips.  It would probably need to work at the level of story elements, 
>and patch 'em together ala King Of Dragon Pass or something.
>Anyone already doing this, that we know about and can publically observe?

  It occurs to me that I've already done the adventure RPG version of
this, I just didn't automate the decision process.  Umbra (both the
Python version and the Hephaestus adventure) generate a random world.
When I'm playtesting, I'll look at the world and say "No, not like
that".

  To design it initially, I made a framework for generating random maps,
and then altered my algorithms or the algorithm-specific parameters,
reloaded, and generated a bunch of maps to test it.  It'd be possible to
randomize the parameters and "vote" on which ones were more like what I
wanted.

>> Oh, great Cthulhu.  Please, not another "Let's take a bunch of
>>strangers with radically different tastes and make a game!  Online!
>>I've got an idea, and you can all work for me for free!" thing.  Those,
>>we know to be failures, every single time they've been tried.
>Well, they teach the participants a lot of things about software 
>development and project management.

  True.  You can make a successful open-source project if there's one
very strong-willed leader with a clear vision who can farm out work to
junior programmers.  Money is about the only way you'll ever find
willing flunkies with any talent, though.

-- 
 <a href="http://kuoi.asui.uidaho.edu/~kamikaze/";> Mark Hughes </a>
"I think [Robert Heinlein] would take it kindly if we were all to refrain from
 abandoning civilization as a failed experiment that requires too much hard
 work." -_Rah, Rah, RAH!_, by Spider Robinson