[NetBehaviour] exist.pl and communication, what next?

Wed Jul 30 17:47:01 CEST 2008

Hi Pall

I'm glad I helped. I really love your project, because such things
have been very often in my thoughts and a few times in my text editor.
I've been thinking about similar experiments very much, but (as often
with me) these reflexions have quickly turned into too complex things,
so that most of the time, I have not even reached the code writing
step.
Your project took directly an experimental approach, which consists in
writing code quickly to see what happens, and then going little step
by little step further, and I think it's indeed very important to keep
going that way, and avoid to go into over-complicated steps.
That's why I've been trying to repress my thoughts here when they look
too "parasiting" for the sake of simplicity.
Now that you're warned, I guess I can say a couple more:
I think the "duplication" possibility could be at the same time very
simple, and very powerful for the future. You could start by writing a
simple "spreading" algorithm that would copy the source here and there
on your system, and maybe run it. If your script copies itself once
somewhere else, and runs this copy, then the copy would copy itself
once again, until a certain point.
This "process" will probably meet opposition by the system : it will
probably reach limits of memory, but even more probably will not be
able to spread to some locations, like system protected ones (on Mac
and Linux, hehehe). Its spreading may also be constrained by the OS in
other ways, or by other programs like antiviruses...
I think the "duplication" process is interresting in the fact that it
reveals those limits, which are the limits between the "being" (which
here is the spreading mass of programs like cells) and the world.

For very different reasons, I've also been interrested in
phenomenology, though not very deeply. I don't want to go too far in
that (I don't want to prove my ignorance, and it's a bit difficult for
me to explain in english), but "being" seem to be not only about "here
and now", but also about the continuity of being "here and now", and
thus about projecting oneself in the future, like in the word
"pre-sence".
Also, "being" has something to do with pointing limits between the
body and the outside.

Anyway, in this approach, I think the "awareness" should imply for the
program to be aware of the activity of its "brothers" in some way.
Each program should then "communicate" with the others in a (real)
"peer to peer" way: maybe via server/client relationships (each
programs listens to its own port, and attempts to communicate with
other ports if opened), or via other ways (list all processes and send
signals to them, I don't know)

All this looks "simple" enough to me, at list in terms of
implentation, but yet may allow you to experiment funny things. It
could also totally mess up your computer, but that's why "beings" are
there, no ?

By the way, it's not directly related but do you know the novel
Permutation City by Greg Egan.
http://en.wikipedia.org/wiki/Permutation_City
If not insightful for your research, I think you'll at least spend
good times reading it.
(What makes me think about your project is mainly the "Dust theory"
developped in it :
http://gregegan.customer.netspace.net.au/PERMUTATION/FAQ/FAQ.html )

About "Polymorphic Code" (which I didn't know), something very
interesting in that topic is Genetic programming :
http://en.wikipedia.org/wiki/Genetic_programming .
In fact you never heard of this, it consists (AFAIK) in creating a
"population" of programs with random code. You "test" the "efficiency"
of each program (this is probably the hard part), and generate the
next "generation" of programs by combining them in function of this
"efficiency", giving more probability for good programs to "duplicate"
themselves. After a while generations, you may get a program that
solves your problem. It's in fact the programming equivalent of the
Darwin theory.
It's quite difficult to actually do (I've been trying to generate
random code ... and "efficiency" testing is not always very clear),
but might at least inspire you.

+++++++
Clément

On Wed, Jul 30, 2008 at 4:01 PM, Pall Thayer <pallthay at gmail.com> wrote:

> Ah you know, these things are always crystal clear when you're not
> sitting in front of a computer. For instance when I'm waiting for the
> bus in the morning, I've got it all figured out. But give me a few
> minutes and let me try to recall... Ok (now I remember), effectively the
> program shouldn't be able to do anything about dying. It could attempt
> to avoid being killed by capturing signals as Rob had suggested, but if
> something has already killed it or voided its existence (deleted the
> file), in human terms it shouldn't be able to do anything about it.
> However, it's not human and it CAN in fact recreate itself in the face
> of a threat to its being. It recreates itself because it can. The reason
> it dies after recreating itself is that the running process is no longer
> a product of the existing file. So the "state of being" has been
> compromised. There's no longer the same relation between the "existence"
> (the file) and the "state of being" (the process). It gets a bit
> complicated here because of the fact that a computer program such as
> this one can run independently of the file from which it was started. In
> our case, although some like to believe otherwise, you can't maintain a
> "state of being" if someone removes your physical body.

> The viral issue is a touchy one. Not a decision to be taken lightly.
>
> Thanks a lot for the comments. They really help.
>
> best,
> Pall
>
>
>> ++++++++
>> Clément
>>
>> On Mon, Jul 28, 2008 at 11:36 PM, Pall Thayer <pallthay at gmail.com> wrote:
>>
>>> (ps. I'm now posting these to both Rhizome and Netbehaviour. My
>>> apologies to anyone who receives them twice.)
>>>
>>> The latest revision to exist.pl has opened a whole new can of beans.
>>> Since it is now capable of receiving communication from other
>>> processes it will inevitably have to respond and that's the tricky
>>> part. How does a process that is just beginning to experiment with an
>>> awareness of anything at all, respond to anything at all? It makes no
>>> attempt to understand the message being conveyed or even who it's
>>> coming from. It would be great to get some feedback on this. Of
>>> course, my first inclination is to just have it respond to anything
>>> with a full dump of its entire "awareness". Well, no. My first
>>> inclination was to have it respond to anything by outputting the full
>>> path to the file (its "existence") and its process ID (its "state of
>>> being") but when you think about it, there's really nothing to
>>> indicate to exist.pl that those two bits of information would mean
>>> anything to anyone else.
>>>
>>> Pall
>>>
>>> --
>>> *****************************
>>> Pall Thayer
>>> artist
>>> http://www.this.is/pallit
>>> *****************************
>>> _______________________________________________
>>> NetBehaviour mailing list
>>> NetBehaviour at netbehaviour.org
>>> http://www.netbehaviour.org/mailman/listinfo/netbehaviour
>>>
>>>
>>
>> _______________________________________________
>> NetBehaviour mailing list
>> NetBehaviour at netbehaviour.org
>> http://www.netbehaviour.org/mailman/listinfo/netbehaviour
>>
>
> _______________________________________________
> NetBehaviour mailing list
> NetBehaviour at netbehaviour.org
> http://www.netbehaviour.org/mailman/listinfo/netbehaviour
>