value ranges in human communication

["Brandon J. Van Every" <vanevery@3DProgrammer.com>]

> What? First you say it's already been done, then you doubt that
> it's possible to do it at all.

Not true.  You're talking about two distinct *it*s.

I'm having a general communication problem with people lately.  This isn't
the first time recently that someone has said "You said X, but you also said
Y.  Those are contradictory!"  But they aren't, they're just on a range of
values rather than being binary YES | NO.  The default way people react to
arguments is "No!  Yes!  No!"  It's unfortunate, because so much of human
communication can be understood as jockeying up and down a normalized range
of values from 0.0 to 1.0.  All anyone is ever doing in a debate is
squeezing in on some value, higher, lower, higher, lower, until the epsilon
is close enough to satisfy multiple parties.

You've marked me as "Data Driven" in my outlook on IF.  If I understand the
term, then I agree.  I think human beings have to provide the bulk of what's
compelling about content.  I don't think computers can processually
rearrange the content in ways that are as fundamentally interesting and
compelling as the best novelists.  Any more than I think 3D graphics
optimization code can be automated in the chaos of real-world engineering.
It takes human skill to provide high-quality integration of elements, to
observe global relationships between local elements, and to smooth the
boundaries of local elements in palatable ways.  Really, the only thing I
think a computer can do is provide "glue," and it's a question of whether it
provides "good glue" or "bad glue."

Let's map "Content Automation" to the value range 0.0..1.0.
0.0 = completely data driven.
1.0 = completely process driven.
Let P = the most difficult IF authoring problem that is possible to solve
My belief is 0.0 < P <= 0.5
Let S = IF authoring problems we've already solved
My belief is 0.0 < S < P <= 0.5

Now, we can argue about where S and P really are, but those are my bounds on
the problem.  This summarizes my 2 earlier statements on the nature of S and
P, which you thought were contradictory.

> Perhaps we could split the difference and agree that the problem
> I've defined is extraordinarily difficult,

Indeed, my objection is you try to define 0.5 < P.  I'll never agree to
that, I consider it overly optimistic.  Even if you posit strong AI, why
should the equivalent of human intellect count as an automated process?  No,
you'd have to come up with an automated process that's far simpler than
human intellect, yet produces the high quality results of the best human
intellects.  I say it is impossible.

> and that the existing corpus contains good
> creative work that skirts that very difficult problem in clever
> ways but has
> not shown any tendency to converge on a solution to it. Which was
> my point in the first place.

I have no reason to agree to that.  I think you are being unnecessarily
pessimistic about the relations 0.0 < S < P <= 0.5.  I think your whole
notion of attractors proves that there *are* convergences, which have been
hinted at by various people's more data driven IF.  I also think my own Rule
Of Three is such an attractor.

> Since you seem to understand my point that the problem I've defined
> _requires_ analyzing, I think what you really mean is, work on a proper
> problem.

Yes, IMHO quit trying to pick the impossible 0.5 < P.  It's a waste of your
time.

> Some of us find that those standard models allow us to meet our creative
> goals. Because our creative goals differ, others of us find that
> they do not.
> This splits discussion of interactive storytelling into two camps: those
> seeking to get the most, technically and artistically, out of
> standard-model
> techniques, and those seeking to develop new better techniques.

That IMHO is overly simplistic because even within 0.0 < S < P <= 0.5 the
"standards" are at various points, and until you reach 0.5 one can always
implement "better" automation.  Again, the problem is people polarize their
arguments into YES | NO rather than recognize that they're jockeying and
squeezing in on numerical ranges.

> And yet, at the very least,
> neither side has much patience with the other.

True, when one of those sides is 0.5 < P.

Some impatience when one of the sides is actually within 0.0 < S < P <= 0.5,
but doesn't realize it and strenuously objects to the "seeming
contradictions" when there aren't any.  I'm not thinking of present company,
more like the rec.arts.int-fiction crowd, who at times wear their
data-drivenness like a badge.

So, the question becomes "What is worth debating?"  Do people feel the
details of 0.0 < S < P <= 0.5 are fundamentally important and in need of
refinement?  If so, I find debate fairly productive once terms of
communication are established.  I wish more people would use value ranges
instead of YES | NO.  I intend to try that myself from now on, to see if it
has any positive effect on mutual understanding.

Or do people feel 0.5 < P is the fundamental question?  If so, I'll be
talking about universal meaninglessness, nihilism, Existential angst,
crawling out of the primordial soup, random numbers, the existence of
difference in the universe, and the perplexing question of why anything
exists at all.  I really just think it's totally the wrong question to ask,
and that's a philosophical claim.  I'd be surprised if anyone can convert me
on that subject, but I'm happy to have such a debate on the oft chance that
I'm surprised.

Well, I'll debate it within reason.  I do think people have to produce,
that's the most important imperative.  No production, no Art.  Nothing to
theorize about.


Cheers,                                     www.3DProgrammer.com
Brandon Van Every                           Seattle, WA

For plot and pace, writers use words; game designers use numbers.
Anything understood over time has plot and pace.