# KAOS RFC

Tags:
Last response: in Video Games
Share
Anonymous
February 1, 2005 6:50:39 PM

I'm currently working on a HLL to CAOS compiler - 'KAOS'. The grammar isn't
finalized yet, so I'd like to get opinions on what it should look like. The
current incarnation looks like this (these ones are already compiling):

# examples/99bottles.kaos
local bottlesonrack;
bottlesonrack = 99;
do {
print("\n");
if (bottlesonrack == 1) {
print("One bottle of beer on the wall,\n");
print("One bottle of beer...\n");
print("Take it down, pass it around,\n");
print("No bottles of beer on the wall!\n");
} else {
print(bottlesonrack, " bottles of beer on the wall,\n");
print(bottlesonrack, " bottles of beer...\n");
print("Take it down, pass it around,\n");
if (bottlesonrack > 2) {
print(bottlesonrack - 1, " bottles of beer on the wall!\n");
} else {
print("One bottle of beer on the wall!\n");
}
}
bottlesonrack = bottlesonrack - 1;
} while (bottlesonrack != 0)
# End examples/99bottles.kaos

# examples/fib.kaos
# generate the first 10 fibbonacci numbers
local i;
local x;
local y;
i = 0;
x = 0;
y = 1;
do {
local temp;
print("fib(", i, ") = ", y, "\n");
temp = x + y;
x = y;
y = temp;
i = i + 1;
} while (i < 10)

# End examples/fib.kaos

# (x - 4)(x + 5) = x^2 + x - 20
local a;
local b;
local c;
local determinant;
local x1;
local x2;
a = 1.0;
b = 1.0;
c = -20.0;
determinant = sqrt(b * b - a * c * 4.0);
b = b * -1.0;
a = a * 2.0;
x1 = (b + determinant) / a;
x2 = (b - determinant) / a;

print("x1 = ", x1, "\n");
print("x2 = ", x2, "\n");

Anonymous
February 2, 2005 4:47:40 AM

> Interesting idea, though I really don't know what's the point of it.
> You'd have to limit your HLL (C) code to such an extend that you'd
> probably better off writing it in CAOS in the first place. YMMV.

Well, just from the example code, I'd think it's at least a lot easier
to *understand* the code he's writing, and it's certainly a bit easier
for me to bash it out (stuff like being able to print 'x - 1' is a vast
improvement already). But it's the agent/array/etc stuff which would
really show the power of a system of this kind.. no need for messing
with TARG etc, just let it work it out itself.

> Hm... you don't intent to implement Maple in the CAOS, are you :-?
> It would probably be nice to see the CAOS it generates as well,
saying
> "Look, it compiles *this*." isn't that helpful, IMHO.

Right now, this is what you get for the quadratic one..

setv va00 1.0
setv va01 1.0
setv va02 -20.0
setv va05 va01
mulv va05 va01
setv va04 va05
setv va06 4.0
setv va07 va00
mulv va07 va02
mulv va06 va07
subv va04 va06
setv va03 sqrt va04
setv va11 -1.0
mulv va11 va01
setv va10 va11
setv va09 va10
setv va12 2.0
mulv va12 va00
divv va09 va12
setv va08 va09
setv va16 -1.0
mulv va16 va01
setv va15 va16
subv va15 va03
setv va14 va15
setv va17 2.0
mulv va17 va00
divv va14 va17
setv va13 va14
outs "x1 = "
outv va08
outs "\n"
outs "x2 = "
outv va13
outs "\n"

... which is probably the one which gains the most understandability
from being in a HLL.
Anonymous
February 2, 2005 10:27:53 AM

On 2005-02-01, bd <bdonlan@gmail.com> wrote:
> I'm currently working on a HLL to CAOS compiler - 'KAOS'. The grammar isn't
> finalized yet, so I'd like to get opinions on what it should look like. The
> current incarnation looks like this (these ones are already compiling):

Interesting idea, though I really don't know what's the point of it.
You'd have to limit your HLL (C) code to such an extend that you'd
probably better off writing it in CAOS in the first place. YMMV.

> # examples/99bottles.kaos
> local bottlesonrack;
> bottlesonrack = 99;
> do {
> print("\n");
> if (bottlesonrack == 1) {
> print("One bottle of beer on the wall,\n");
> print("One bottle of beer...\n");
> print("Take it down, pass it around,\n");
> print("No bottles of beer on the wall!\n");
> } else {
> print(bottlesonrack, " bottles of beer on the wall,\n");
> print(bottlesonrack, " bottles of beer...\n");
> print("Take it down, pass it around,\n");
> if (bottlesonrack > 2) {
> print(bottlesonrack - 1, " bottles of beer on the wall!\n");
> } else {
> print("One bottle of beer on the wall!\n");
> }
> }
> bottlesonrack = bottlesonrack - 1;
> } while (bottlesonrack != 0)
> # End examples/99bottles.kaos
>
> # examples/fib.kaos
> # generate the first 10 fibbonacci numbers
> local i;
> local x;
> local y;
> i = 0;
> x = 0;
> y = 1;
> do {
> local temp;
> print("fib(", i, ") = ", y, "\n");
> temp = x + y;
> x = y;
> y = temp;
> i = i + 1;
> } while (i < 10)
>
> # End examples/fib.kaos
>
> # (x - 4)(x + 5) = x^2 + x - 20
> local a;
> local b;
> local c;
> local determinant;
> local x1;
> local x2;
> a = 1.0;
> b = 1.0;
> c = -20.0;
> determinant = sqrt(b * b - a * c * 4.0);
> b = b * -1.0;
> a = a * 2.0;
> x1 = (b + determinant) / a;
> x2 = (b - determinant) / a;
>
> print("x1 = ", x1, "\n");
> print("x2 = ", x2, "\n");

Hm... you don't intent to implement Maple in the CAOS, are you :-?
It would probably be nice to see the CAOS it generates as well, saying
"Look, it compiles *this*." isn't that helpful, IMHO.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Related resources
Anonymous
February 2, 2005 12:37:17 PM

> > for me to bash it out (stuff like being able to print 'x - 1' is a
vast
>
> I'm not sure how it's exactly handeled, but I guess it does nothing
else
> but converting it in a couple of OUTS and OUTVs - nothing you
actually
> use a lot in CAOS and not to hard to do in CAOS anyway.

Well, it has to do the SUBV first (and probably a SETV, given you
generally have to preserve x), which makes it less simple .. not very
useful for 'print', admittedly, but could be very useful for other,
more useful, functions if you're doing more sophisticated calculations.
Anonymous
February 2, 2005 1:50:37 PM

Alyssa Milburn wrote:

>> Interesting idea, though I really don't know what's the point of it.
>> You'd have to limit your HLL (C) code to such an extend that you'd
>> probably better off writing it in CAOS in the first place. YMMV.
>
> Well, just from the example code, I'd think it's at least a lot easier
> to *understand* the code he's writing, and it's certainly a bit easier
> for me to bash it out (stuff like being able to print 'x - 1' is a vast
> improvement already). But it's the agent/array/etc stuff which would
> really show the power of a system of this kind.. no need for messing
> with TARG etc, just let it work it out itself.

I do intend to use ovxx as an array type (obj[42]) and NAME variables for
associative arrays (obj["hello, world!"]). And eventually it'll be able to
work out TARG on its own.
Anonymous
February 2, 2005 2:43:16 PM

On 2005-02-02, Alyssa Milburn <fuzzie@gmail.com> wrote:
>> Interesting idea, though I really don't know what's the point of it.
>> You'd have to limit your HLL (C) code to such an extend that you'd
>> probably better off writing it in CAOS in the first place. YMMV.
>
> Well, just from the example code, I'd think it's at least a lot easier
> to *understand* the code he's writing, and it's certainly a bit easier

Probably.

> for me to bash it out (stuff like being able to print 'x - 1' is a vast

I'm not sure how it's exactly handeled, but I guess it does nothing else
but converting it in a couple of OUTS and OUTVs - nothing you actually
use a lot in CAOS and not to hard to do in CAOS anyway.

> improvement already). But it's the agent/array/etc stuff which would
> really show the power of a system of this kind.. no need for messing
> with TARG etc, just let it work it out itself.

Interesting thought...

>> Hm... you don't intent to implement Maple in the CAOS, are you :-?
>> It would probably be nice to see the CAOS it generates as well,
> saying
>> "Look, it compiles *this*." isn't that helpful, IMHO.
>
> Right now, this is what you get for the quadratic one..
>
> setv va00 1.0
> setv va01 1.0
> setv va02 -20.0
> setv va05 va01
> mulv va05 va01
> setv va04 va05
> setv va06 4.0
> setv va07 va00
> mulv va07 va02
> mulv va06 va07
> subv va04 va06
> setv va03 sqrt va04
> setv va11 -1.0
> mulv va11 va01
> setv va10 va11
> setv va09 va10
> setv va12 2.0
> mulv va12 va00
> divv va09 va12
> setv va08 va09
> setv va16 -1.0
> mulv va16 va01
> setv va15 va16
> subv va15 va03
> setv va14 va15
> setv va17 2.0
> mulv va17 va00
> divv va14 va17
> setv va13 va14
> outs "x1 = "
> outv va08
> outs "\n"
> outs "x2 = "
> outv va13
> outs "\n"
>
> .. which is probably the one which gains the most understandability
> from being in a HLL.

You definitely have a point here.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 3, 2005 12:05:22 PM

On 2005-02-02, Alyssa Milburn <fuzzie@gmail.com> wrote:
>> > for me to bash it out (stuff like being able to print 'x - 1' is a
> vast
>>
>> I'm not sure how it's exactly handeled, but I guess it does nothing
> else
>> but converting it in a couple of OUTS and OUTVs - nothing you
> actually
>> use a lot in CAOS and not to hard to do in CAOS anyway.
>
> Well, it has to do the SUBV first (and probably a SETV, given you
> generally have to preserve x), which makes it less simple .. not very
> useful for 'print', admittedly, but could be very useful for other,
> more useful, functions if you're doing more sophisticated calculations.

Yes, I think so. Still - what's the *point*. Except that it can be done,
of course?
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 3, 2005 1:03:19 PM

> I wonder what the CAOS engine actually looks like. Perl plugins would
> have been an interesting alternative to CAOS IMHO.

Well, at the time of C1 (and even C2), it wasn't particularly viable to
use perl/python/whatever .. plus the open source stuff was only just
taking off and not a lot of people were aware of it all.

And I expect they didn't see much point changing it for C3, considering
that their agent developers were presumably accustomed to CAOS and they

And, 'looks like'? If you mean code-wise, I have an unfinished (as of
yet) C3 engine clone here which seems pretty similar in behaviour to
the proper C3 one (error handling in the same places, etc); the
language is very simple, so you can just split it up into groups of
tokens representing commands (looking up the functions/commands in a
hash table) and run those through the runtime.
Anonymous
February 4, 2005 1:51:24 AM

emmel wrote:

> On 2005-02-02, Alyssa Milburn <fuzzie@gmail.com> wrote:
>>> Interesting idea, though I really don't know what's the point of it.
>>> You'd have to limit your HLL (C) code to such an extend that you'd
>>> probably better off writing it in CAOS in the first place. YMMV.
>>
>> Well, just from the example code, I'd think it's at least a lot easier
>> to *understand* the code he's writing, and it's certainly a bit easier
>
> Probably.
>
>> for me to bash it out (stuff like being able to print 'x - 1' is a vast
>
> I'm not sure how it's exactly handeled, but I guess it does nothing else
> but converting it in a couple of OUTS and OUTVs - nothing you actually
> use a lot in CAOS and not to hard to do in CAOS anyway.
>

We intend to add user-defined functions with return values and stuff, that
should be a big improvement. You might think it's useless, but that's fine,
'cause REVELATION won't show the KAOS anyway :3

>> improvement already). But it's the agent/array/etc stuff which would
>> really show the power of a system of this kind.. no need for messing
>> with TARG etc, just let it work it out itself.
>
> Interesting thought...

I still don't know how that will work. ^^;;

>
>>> Hm... you don't intent to implement Maple in the CAOS, are you :-?

What's Maple?

>>> It would probably be nice to see the CAOS it generates as well,
>> saying
>>> "Look, it compiles *this*." isn't that helpful, IMHO.
>>
>> Right now, this is what you get for the quadratic one..
>>
>> <snip>
>>
>> .. which is probably the one which gains the most understandability
>> from being in a HLL.
>
> You definitely have a point here.

<nod> the KAOS is much easier to understand than the CAOS

--
- nornagon
http://www.nornrock.com - Still In Development
mailto: nornagon@nornrock.com and mailto: nornagon@gmail.com
Species range 10001-10100
Anonymous
February 4, 2005 1:51:25 AM

On 2005-02-03, nornagon <nornagon@gmail.com> wrote:
> emmel wrote:
>
>> On 2005-02-02, Alyssa Milburn <fuzzie@gmail.com> wrote:
>>>> Interesting idea, though I really don't know what's the point of it.
>>>> You'd have to limit your HLL (C) code to such an extend that you'd
>>>> probably better off writing it in CAOS in the first place. YMMV.
>>>
>>> Well, just from the example code, I'd think it's at least a lot easier
>>> to *understand* the code he's writing, and it's certainly a bit easier
>>
>> Probably.
>>
>>> for me to bash it out (stuff like being able to print 'x - 1' is a vast
>>
>> I'm not sure how it's exactly handeled, but I guess it does nothing else
>> but converting it in a couple of OUTS and OUTVs - nothing you actually
>> use a lot in CAOS and not to hard to do in CAOS anyway.
>>
>
> We intend to add user-defined functions with return values and stuff, that
> should be a big improvement. You might think it's useless, but that's fine,
> 'cause REVELATION won't show the KAOS anyway :3

Never actually used REVELATION. Anyway, I was just saying what it
appeared to me. I never said it would be generaly useless.

>>> improvement already). But it's the agent/array/etc stuff which would
>>> really show the power of a system of this kind.. no need for messing
>>> with TARG etc, just let it work it out itself.
>>
>> Interesting thought...
>
> I still don't know how that will work. ^^;;

I wonder what the CAOS engine actually looks like. Perl plugins would
have been an interesting alternative to CAOS IMHO.

>>
>>>> Hm... you don't intent to implement Maple in the CAOS, are you :-?
>
> What's Maple?

Very popular (and powerful) Computer Algebra System. A bit on the
expesive side, though (so I don't actually use it).

>>>> It would probably be nice to see the CAOS it generates as well,
>>> saying
>>>> "Look, it compiles *this*." isn't that helpful, IMHO.
>>>
>>> Right now, this is what you get for the quadratic one..
>>>
>>> <snip>
>>>
>>> .. which is probably the one which gains the most understandability
>>> from being in a HLL.
>>
>> You definitely have a point here.
>
><nod> the KAOS is much easier to understand than the CAOS

But chaos is much harder to mispell for 'kaos'...
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 5, 2005 6:03:10 AM

> Yes, you are right. I remember (was it Lisa?) saying that the C3
engine
> was pretty close to the C2 codebase alhough it didn't look quite like
> it.

Ah, interesting, I'd never heard that.

> Freetures engine?

Maybe. It might be better to build a Freetures engine around something
like perl/python, but certainly it has code which would be useful for
Freetures..

> The point I was trying to make was that it should be pretty similar
to
> the KAOS compiler.

Not really. The KAOS compiler is basically just parsing a quite complex
grammar and transforming it to bytecode-ish-ness (which is what CAOS is
a representation of, really), whereas the CAOS runtime is parsing a
simpler grammar into internal bytecode. You could *add* a HLL compiler
to the engine (assuming you had the source), but it'd just add
complexity and slowness which isn't really needed.

> And wheather it was easy to put something like perl in.

I think it'd be quite difficult, although I'm not sure about perl
embedding .. one important thing CAOS provides is the ability to run a
bunch of scripts at once (ie, to be able to execute 10 bytecodes of a
script and then switch to the next script) whereas all the embedding
interfaces for scripting languages seem tailored to the "execute the
whole script now!" model.

Plus, CAOS is rather fast to execute (grab line of parsed tokens, call
CAOS function referenced in the first token, function pops tokens while
executing code) and even then enough CPU time is used up by the CAOS
scripts as it is...
Anonymous
February 5, 2005 12:33:59 PM

On 2005-02-03, Alyssa Milburn <fuzzie@gmail.com> wrote:
>> I wonder what the CAOS engine actually looks like. Perl plugins would
>> have been an interesting alternative to CAOS IMHO.
>
> Well, at the time of C1 (and even C2), it wasn't particularly viable to
> use perl/python/whatever .. plus the open source stuff was only just
> taking off and not a lot of people were aware of it all.

Hm. Didn't think of that.

> And I expect they didn't see much point changing it for C3, considering
> that their agent developers were presumably accustomed to CAOS and they

Yes, you are right. I remember (was it Lisa?) saying that the C3 engine
was pretty close to the C2 codebase alhough it didn't look quite like
it.

> And, 'looks like'? If you mean code-wise, I have an unfinished (as of

Yup.

> yet) C3 engine clone here which seems pretty similar in behaviour to
> the proper C3 one (error handling in the same places, etc); the

Freetures engine?

> language is very simple, so you can just split it up into groups of
> tokens representing commands (looking up the functions/commands in a
> hash table) and run those through the runtime.

The point I was trying to make was that it should be pretty similar to
the KAOS compiler. And wheather it was easy to put something like perl
in.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 5, 2005 12:38:14 PM

> Yes, but both interpret the CAOS to a certain degree. Of course the
one
> executes it and the other _just_ translates it, but you get my point.
> You do, don't you?

KAOS isn't interpreting the CAOS at all, it just produces CAOS. It
doesn't actually have any knowledge beyond 'this code in the HLL
produces this code in CAOS'.

> sharing shouldn't be much of a problem. There's this pseudo game out
> there where you program robots to fight in a virtual arena - that
kind
> of thing (nothing to do with those stupid robot battle shows, I'm

It doesn't have thousands and thousands of robots which need updating
often, though. If you look back to the Freetures thread Chris Double
points out that the need for interruptable scripting is something which
is rather important.

There *is* the possibility of writing a differently designed engine
where agents do things in different ways and the engine takes a lot
more responsibility for things, but it'd involve reworking an awful
lot, I'm just discussing a Creatures-engine-like model here.

> I don't think perl would be much slower. It's executeable code just
as
> everything else, though there might be a 'slight' overhead through
the
> supporting libraries use to actualle *get* all that HLL stuff done -
> like advanced variable handling and the like.

Perl is really not fast - remember it's a scripting language, and
designed for string manipulation at heart - although it's not bad.
Parrot (the VM for perl 6) is a lot more interesting for this sort of
thing, but not an awful lot of use while it's still in heavy
development.

Lua might be a good idea. It's what Gameware have been using for their
recent projects. I hate it as a language, though.
Anonymous
February 5, 2005 7:55:21 PM

On 2005-02-05, Alyssa Milburn <fuzzie@gmail.com> wrote:
>> Freetures engine?
>
> Maybe. It might be better to build a Freetures engine around something
> like perl/python, but certainly it has code which would be useful for
> Freetures..

::nods::

>> The point I was trying to make was that it should be pretty similar
> to
>> the KAOS compiler.
>
> Not really. The KAOS compiler is basically just parsing a quite complex
> grammar and transforming it to bytecode-ish-ness (which is what CAOS is
> a representation of, really), whereas the CAOS runtime is parsing a
> simpler grammar into internal bytecode. You could *add* a HLL compiler
> to the engine (assuming you had the source), but it'd just add
> complexity and slowness which isn't really needed.

Yes, but both interpret the CAOS to a certain degree. Of course the one
executes it and the other _just_ translates it, but you get my point.
You do, don't you?

>> And wheather it was easy to put something like perl in.
>
> I think it'd be quite difficult, although I'm not sure about perl
> embedding .. one important thing CAOS provides is the ability to run a
> bunch of scripts at once (ie, to be able to execute 10 bytecodes of a
> script and then switch to the next script) whereas all the embedding
> interfaces for scripting languages seem tailored to the "execute the
> whole script now!" model.

Interesting aspect. Though of course if the agents were actually
programs by themselves executed through the engine implementing time
sharing shouldn't be much of a problem. There's this pseudo game out
there where you program robots to fight in a virtual arena - that kind
of thing (nothing to do with those stupid robot battle shows, I'm

> Plus, CAOS is rather fast to execute (grab line of parsed tokens, call
> CAOS function referenced in the first token, function pops tokens while
> executing code) and even then enough CPU time is used up by the CAOS
> scripts as it is...

I don't think perl would be much slower. It's executeable code just as
everything else, though there might be a 'slight' overhead through the
supporting libraries use to actualle *get* all that HLL stuff done -
like advanced variable handling and the like.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 6, 2005 10:56:36 AM

On 2005-02-05, Alyssa Milburn <fuzzie@gmail.com> wrote:
>> Yes, but both interpret the CAOS to a certain degree. Of course the
> one
>> executes it and the other _just_ translates it, but you get my point.
>> You do, don't you?
>
> KAOS isn't interpreting the CAOS at all, it just produces CAOS. It
> doesn't actually have any knowledge beyond 'this code in the HLL
> produces this code in CAOS'.

Yeah, yeah, yeah... Look, I'd call that interpreting - in a way. YMMV

>> sharing shouldn't be much of a problem. There's this pseudo game out
>> there where you program robots to fight in a virtual arena - that
> kind
>> of thing (nothing to do with those stupid robot battle shows, I'm
>
> It doesn't have thousands and thousands of robots which need updating
> often, though. If you look back to the Freetures thread Chris Double
> points out that the need for interruptable scripting is something which
> is rather important.

I only wonder if that couldn't be done with mupltiple processes.

> There *is* the possibility of writing a differently designed engine
> where agents do things in different ways and the engine takes a lot
> more responsibility for things, but it'd involve reworking an awful
> lot, I'm just discussing a Creatures-engine-like model here.

Well, I thought of getting the engine to do less. Just kicking the CAOS
interpreter out and replacing it by some other interface. I have to
admit that that'd mean changing most of the engine and it'd probably
wouldn't be that safe (viri anyone?), but it sure would be nice. If it
gave the performance. Using HLL directly, you see. But I know that
that's unlikely to happen.

>> I don't think perl would be much slower. It's executeable code just
> as
>> everything else, though there might be a 'slight' overhead through
> the
>> supporting libraries use to actualle *get* all that HLL stuff done -
>> like advanced variable handling and the like.
>
> Perl is really not fast - remember it's a scripting language, and
> designed for string manipulation at heart - although it's not bad.
> Parrot (the VM for perl 6) is a lot more interesting for this sort of
> thing, but not an awful lot of use while it's still in heavy
> development.

Well, Perl isn't interpreted. Any overhead generated would be because of
the run time libraries. Maybe it'd be possible to strip those down to a
minium? Probably not very practical approach (read: time intensive), but
think of the possibilities.

> Lua might be a good idea. It's what Gameware have been using for their
> recent projects. I hate it as a language, though.

Frankly I haven't really looked at Lua, so I can't judge what it'd be
like as a language.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 7, 2005 9:26:34 PM

emmel wrote:

> On 2005-02-05, Alyssa Milburn <fuzzie@gmail.com> wrote:
>>> Yes, but both interpret the CAOS to a certain degree. Of course the
>> one
>>> executes it and the other _just_ translates it, but you get my point.
>>> You do, don't you?
>>
>> KAOS isn't interpreting the CAOS at all, it just produces CAOS. It
>> doesn't actually have any knowledge beyond 'this code in the HLL
>> produces this code in CAOS'.
>
> Yeah, yeah, yeah... Look, I'd call that interpreting - in a way. YMMV
>
>>> sharing shouldn't be much of a problem. There's this pseudo game out
>>> there where you program robots to fight in a virtual arena - that
>> kind
>>> of thing (nothing to do with those stupid robot battle shows, I'm
>>
>> It doesn't have thousands and thousands of robots which need updating
>> often, though. If you look back to the Freetures thread Chris Double
>> points out that the need for interruptable scripting is something which
>> is rather important.
>
> I only wonder if that couldn't be done with mupltiple processes.
>
>> There *is* the possibility of writing a differently designed engine
>> where agents do things in different ways and the engine takes a lot
>> more responsibility for things, but it'd involve reworking an awful
>> lot, I'm just discussing a Creatures-engine-like model here.
>
> Well, I thought of getting the engine to do less. Just kicking the CAOS
> interpreter out and replacing it by some other interface. I have to
> admit that that'd mean changing most of the engine and it'd probably
> wouldn't be that safe (viri anyone?), but it sure would be nice. If it
> gave the performance. Using HLL directly, you see. But I know that
> that's unlikely to happen.
>
>>> I don't think perl would be much slower. It's executeable code just
>> as
>>> everything else, though there might be a 'slight' overhead through
>> the
>>> supporting libraries use to actualle *get* all that HLL stuff done -
>>> like advanced variable handling and the like.
>>
>> Perl is really not fast - remember it's a scripting language, and
>> designed for string manipulation at heart - although it's not bad.
>> Parrot (the VM for perl 6) is a lot more interesting for this sort of
>> thing, but not an awful lot of use while it's still in heavy
>> development.
>
> Well, Perl isn't interpreted. Any overhead generated would be because of

On the contrary, I believe Perl _is_ interpreted.

> the run time libraries. Maybe it'd be possible to strip those down to a
> minium? Probably not very practical approach (read: time intensive), but
> think of the possibilities.
>
>> Lua might be a good idea. It's what Gameware have been using for their
>> recent projects. I hate it as a language, though.
>
> Frankly I haven't really looked at Lua, so I can't judge what it'd be
> like as a language.

advanced features, such as classes, etc. And, of course, it wouldn't need
to compile down to CAOS.

--
- nornagon
http://www.nornrock.com - Still In Development
mailto: nornagon@nornrock.com and mailto: nornagon@gmail.com
Species range 10001-10100
Anonymous
February 7, 2005 9:26:35 PM

On 2005-02-07, nornagon <nornagon@gmail.com> wrote:
> emmel wrote:
>
>>> Perl is really not fast - remember it's a scripting language, and
>>> designed for string manipulation at heart - although it's not bad.
>>> Parrot (the VM for perl 6) is a lot more interesting for this sort of
>>> thing, but not an awful lot of use while it's still in heavy
>>> development.
>>
>> Well, Perl isn't interpreted. Any overhead generated would be because of
>
> On the contrary, I believe Perl _is_ interpreted.

Isn't. Perl is compiled in byte code and then executed. At least that's
what my perl documentation says.

>> the run time libraries. Maybe it'd be possible to strip those down to a
>> minium? Probably not very practical approach (read: time intensive), but
>> think of the possibilities.
>>
>>> Lua might be a good idea. It's what Gameware have been using for their
>>> recent projects. I hate it as a language, though.
>>
>> Frankly I haven't really looked at Lua, so I can't judge what it'd be
>> like as a language.
>
> For Freetures, I'd suggest a KAOS-like language, though with some more
> advanced features, such as classes, etc. And, of course, it wouldn't need
> to compile down to CAOS.

Hm, nice idea, although I'm not sure what you actually want to use the
classes for. Who's actually *working* on Freatures?
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 7, 2005 9:26:36 PM

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

emmel wrote:
| On 2005-02-07, nornagon <nornagon@gmail.com> wrote:
|
|>emmel wrote:
|>
|>
|>>>Perl is really not fast - remember it's a scripting language, and
|>>>designed for string manipulation at heart - although it's not bad.
|>>>Parrot (the VM for perl 6) is a lot more interesting for this sort of
|>>>thing, but not an awful lot of use while it's still in heavy
|>>>development.
|>>
|>>Well, Perl isn't interpreted. Any overhead generated would be because of
|>
|>On the contrary, I believe Perl _is_ interpreted.
|
|
| Isn't. Perl is compiled in byte code and then executed. At least that's
| what my perl documentation says.

Actually, it's more like a binary parse tree than true bytecode (though
there is a backend to output bytecode, it is slow). See `perldoc B' for

|>>the run time libraries. Maybe it'd be possible to strip those down to a
|>>minium? Probably not very practical approach (read: time intensive), but
|>>think of the possibilities.
|>>
|>>
|>>>Lua might be a good idea. It's what Gameware have been using for their
|>>>recent projects. I hate it as a language, though.
|>>
|>>Frankly I haven't really looked at Lua, so I can't judge what it'd be
|>>like as a language.
|>
|>advanced features, such as classes, etc. And, of course, it wouldn't need
|>to compile down to CAOS.
|
|
| Hm, nice idea, although I'm not sure what you actually want to use the
| classes for. Who's actually *working* on Freatures?

fuzzie's working on a clone of CEE.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCB6Ge+hz2VlChukwRAoN4AKCjxs3TBHMINrHoi8kNF5lJwZHwbACgtwJk
H7YRTuK10OHYaWiMlKLw6g4=
=vJIT
-----END PGP SIGNATURE-----
Anonymous
February 8, 2005 2:15:27 PM

On 2005-02-07, bd <bdonlan@bd.beginyourfear.com> wrote:
>
> emmel wrote:
>| On 2005-02-07, nornagon <nornagon@gmail.com> wrote:
>|
>|>emmel wrote:
>|>
>|>
>|>>>Perl is really not fast - remember it's a scripting language, and
>|>>>designed for string manipulation at heart - although it's not bad.
>|>>>Parrot (the VM for perl 6) is a lot more interesting for this sort of
>|>>>thing, but not an awful lot of use while it's still in heavy
>|>>>development.
>|>>
>|>>Well, Perl isn't interpreted. Any overhead generated would be because of
>|>
>|>On the contrary, I believe Perl _is_ interpreted.
>|
>|
>| Isn't. Perl is compiled in byte code and then executed. At least that's
>| what my perl documentation says.
>
> Actually, it's more like a binary parse tree than true bytecode (though
> there is a backend to output bytecode, it is slow). See `perldoc B' for

Hm, I'll have a look at that some time...

>|>>the run time libraries. Maybe it'd be possible to strip those down to a
>|>>minium? Probably not very practical approach (read: time intensive), but
>|>>think of the possibilities.
>|>>
>|>>
>|>>>Lua might be a good idea. It's what Gameware have been using for their
>|>>>recent projects. I hate it as a language, though.
>|>>
>|>>Frankly I haven't really looked at Lua, so I can't judge what it'd be
>|>>like as a language.
>|>
>|>advanced features, such as classes, etc. And, of course, it wouldn't need
>|>to compile down to CAOS.
>|
>|
>| Hm, nice idea, although I'm not sure what you actually want to use the
>| classes for. Who's actually *working* on Freatures?
>
> fuzzie's working on a clone of CEE.

CEE? I might be stupid, but I have no clue...
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 8, 2005 2:15:28 PM

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

emmel wrote:
| On 2005-02-07, bd <bdonlan@bd.beginyourfear.com> wrote:
|
|>emmel wrote:
|>| On 2005-02-07, nornagon <nornagon@gmail.com> wrote:
|>|
|>|>emmel wrote:
|>|>
|>|>
|>|>>>Perl is really not fast - remember it's a scripting language, and
|>|>>>designed for string manipulation at heart - although it's not bad.
|>|>>>Parrot (the VM for perl 6) is a lot more interesting for this sort of
|>|>>>thing, but not an awful lot of use while it's still in heavy
|>|>>>development.
|>|>>
|>|>>Well, Perl isn't interpreted. Any overhead generated would be
because of
|>|>
|>|>On the contrary, I believe Perl _is_ interpreted.
|>|
|>|
|>| Isn't. Perl is compiled in byte code and then executed. At least that's
|>| what my perl documentation says.
|>
|>Actually, it's more like a binary parse tree than true bytecode (though
|>there is a backend to output bytecode, it is slow). See `perldoc B' for
|
|
| Hm, I'll have a look at that some time...
|
|
|>|>>the run time libraries. Maybe it'd be possible to strip those down to a
|>|>>minium? Probably not very practical approach (read: time
intensive), but
|>|>>think of the possibilities.
|>|>>
|>|>>
|>|>>>Lua might be a good idea. It's what Gameware have been using for their
|>|>>>recent projects. I hate it as a language, though.
|>|>>
|>|>>Frankly I haven't really looked at Lua, so I can't judge what it'd be
|>|>>like as a language.
|>|>
|>|>advanced features, such as classes, etc. And, of course, it wouldn't
need
|>|>to compile down to CAOS.
|>|
|>|
|>| Hm, nice idea, although I'm not sure what you actually want to use the
|>| classes for. Who's actually *working* on Freatures?
|>
|>fuzzie's working on a clone of CEE.
|
|
| CEE? I might be stupid, but I have no clue...

Creatures Evolution Engine. The game engine for CA, C3, DS, Sea monkeys,
etcetc.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCCMEP+hz2VlChukwRApbsAJoDZPQ6GbulCURV5fWE7Kii3YtK0wCfbkNp
nX/VcZCKNqIkfOYVJyk+B2E=
=hM4y
-----END PGP SIGNATURE-----
Anonymous
February 9, 2005 11:43:45 AM

On 2005-02-08, bd <bdonlan@bd.beginyourfear.com> wrote:
>
> emmel wrote:
>| On 2005-02-07, bd <bdonlan@bd.beginyourfear.com> wrote:
>|
>|>emmel wrote:
>|>| On 2005-02-07, nornagon <nornagon@gmail.com> wrote:
>|>|
>|>|>emmel wrote:
>|>|>
>|>|>
>|>|>>>Perl is really not fast - remember it's a scripting language, and
>|>|>>>designed for string manipulation at heart - although it's not bad.
>|>|>>>Parrot (the VM for perl 6) is a lot more interesting for this sort of
>|>|>>>thing, but not an awful lot of use while it's still in heavy
>|>|>>>development.
>|>|>>
>|>|>>Well, Perl isn't interpreted. Any overhead generated would be
> because of
>|>|>
>|>|>On the contrary, I believe Perl _is_ interpreted.
>|>|
>|>|
>|>| Isn't. Perl is compiled in byte code and then executed. At least that's
>|>| what my perl documentation says.
>|>
>|>Actually, it's more like a binary parse tree than true bytecode (though
>|>there is a backend to output bytecode, it is slow). See `perldoc B' for
>|
>|
>| Hm, I'll have a look at that some time...
>|
>|
>|>|>>the run time libraries. Maybe it'd be possible to strip those down to a
>|>|>>minium? Probably not very practical approach (read: time
> intensive), but
>|>|>>think of the possibilities.
>|>|>>
>|>|>>
>|>|>>>Lua might be a good idea. It's what Gameware have been using for their
>|>|>>>recent projects. I hate it as a language, though.
>|>|>>
>|>|>>Frankly I haven't really looked at Lua, so I can't judge what it'd be
>|>|>>like as a language.
>|>|>
>|>|>advanced features, such as classes, etc. And, of course, it wouldn't
> need
>|>|>to compile down to CAOS.
>|>|
>|>|
>|>| Hm, nice idea, although I'm not sure what you actually want to use the
>|>| classes for. Who's actually *working* on Freatures?
>|>
>|>fuzzie's working on a clone of CEE.
>|
>|
>| CEE? I might be stupid, but I have no clue...
>
> Creatures Evolution Engine. The game engine for CA, C3, DS, Sea monkeys,
> etcetc.

It's called that?
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 9, 2005 12:04:26 PM

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

emmel wrote:
| On 2005-02-08, bd <bdonlan@bd.beginyourfear.com> wrote:
|
|>emmel wrote:
|>| On 2005-02-07, bd <bdonlan@bd.beginyourfear.com> wrote:
|>|
|>|>emmel wrote:
|>|>| On 2005-02-07, nornagon <nornagon@gmail.com> wrote:
|>|>|
|>|>|>emmel wrote:
|>|>|>
|>|>|>
|>|>|>>>Perl is really not fast - remember it's a scripting language, and
|>|>|>>>designed for string manipulation at heart - although it's not bad.
|>|>|>>>Parrot (the VM for perl 6) is a lot more interesting for this
sort of
|>|>|>>>thing, but not an awful lot of use while it's still in heavy
|>|>|>>>development.
|>|>|>>
|>|>|>>Well, Perl isn't interpreted. Any overhead generated would be
|>because of
|>|>|>
|>|>|>On the contrary, I believe Perl _is_ interpreted.
|>|>|
|>|>|
|>|>| Isn't. Perl is compiled in byte code and then executed. At least
that's
|>|>| what my perl documentation says.
|>|>
|>|>Actually, it's more like a binary parse tree than true bytecode (though
|>|>there is a backend to output bytecode, it is slow). See `perldoc B' for
|>|
|>|
|>| Hm, I'll have a look at that some time...
|>|
|>|
|>|>|>>the run time libraries. Maybe it'd be possible to strip those
down to a
|>|>|>>minium? Probably not very practical approach (read: time
|>intensive), but
|>|>|>>think of the possibilities.
|>|>|>>
|>|>|>>
|>|>|>>>Lua might be a good idea. It's what Gameware have been using for
their
|>|>|>>>recent projects. I hate it as a language, though.
|>|>|>>
|>|>|>>Frankly I haven't really looked at Lua, so I can't judge what it'd be
|>|>|>>like as a language.
|>|>|>
|>|>|>advanced features, such as classes, etc. And, of course, it wouldn't
|>need
|>|>|>to compile down to CAOS.
|>|>|
|>|>|
|>|>| Hm, nice idea, although I'm not sure what you actually want to use the
|>|>| classes for. Who's actually *working* on Freatures?
|>|>
|>|>fuzzie's working on a clone of CEE.
|>|
|>|
|>| CEE? I might be stupid, but I have no clue...
|>
|>Creatures Evolution Engine. The game engine for CA, C3, DS, Sea monkeys,
|>etcetc.
|
|
| It's called that?

Yep. Sometimes C2E - hence fuzzie's "openc2e" project.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCChhp+hz2VlChukwRAo9SAJ9EEgNxECmyd3oeALqLbp2zvZ5y/gCglYPP
3h+jPQ2gXnuuCVSGIscJxTA=
=LqB9
-----END PGP SIGNATURE-----
Anonymous
February 11, 2005 3:26:22 PM

On 2005-02-09, bd <bdonlan@bd.beginyourfear.com> wrote:
>
> emmel wrote:
>| On 2005-02-08, bd <bdonlan@bd.beginyourfear.com> wrote:
>|>|>fuzzie's working on a clone of CEE.
>|>|
>|>|
>|>| CEE? I might be stupid, but I have no clue...
>|>
>|>Creatures Evolution Engine. The game engine for CA, C3, DS, Sea monkeys,
>|>etcetc.
>|
>|
>| It's called that?
>
> Yep. Sometimes C2E - hence fuzzie's "openc2e" project.

Interesting.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 11, 2005 3:26:23 PM

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

| emmel wrote:
|
|
|>>On 2005-02-09, bd <bdonlan@bd.beginyourfear.com> wrote:
|>>
|>>>emmel wrote:
|>>>| On 2005-02-08, bd <bdonlan@bd.beginyourfear.com> wrote:
|>>>|>|>fuzzie's working on a clone of CEE.
|>>>|>|
|>>>|>|
|>>>|>| CEE? I might be stupid, but I have no clue...
|>>>|>
|>>>|>Creatures Evolution Engine. The game engine for CA, C3, DS, Sea
monkeys,
|>>>|>etcetc.
|>>>|
|>>>|
|>>>| It's called that?
|>>>
|>>>Yep. Sometimes C2E - hence fuzzie's "openc2e" project.
|>>
|>>Interesting.
|
|
| In Linux I think the binary was called lc2e.
|
| I don't like it much, myself, so I'm rolling my own. Nothing very
impressive
| is done yet, but I'm getting there :-)
|
| It will be interesting to see what comes out of it. Fuzzie's goal is
to make
| a clone, mine is to make something new and hopefully better AL-wise.
| a very interesting discussion about how to make a better biochemistry the
| other day in IRC.
|

Don't forget a link or two:

[beware, slow server]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCDOYB+hz2VlChukwRAnqUAJ95z8yqNRQxV60xDF9ah+MnJK7v5ACeIaXk
LM62yOByyZhM7wny1XAF49U=
=xltk
-----END PGP SIGNATURE-----
Anonymous
February 15, 2005 3:12:16 PM

>
> emmel wrote:
>
>> On 2005-02-09, bd <bdonlan@bd.beginyourfear.com> wrote:
>>>
>>> emmel wrote:
>>>| On 2005-02-08, bd <bdonlan@bd.beginyourfear.com> wrote:
>>>|>|>fuzzie's working on a clone of CEE.
>>>|>|
>>>|>|
>>>|>| CEE? I might be stupid, but I have no clue...
>>>|>
>>>|>Creatures Evolution Engine. The game engine for CA, C3, DS, Sea monkeys,
>>>|>etcetc.
>>>|
>>>|
>>>| It's called that?
>>>
>>> Yep. Sometimes C2E - hence fuzzie's "openc2e" project.
>>
>> Interesting.
>
> In Linux I think the binary was called lc2e.

Yeah, I recently checked. I'm feeling stupid now.

> I don't like it much, myself, so I'm rolling my own. Nothing very impressive
> is done yet, but I'm getting there :-)

Keep us updated.

> It will be interesting to see what comes out of it. Fuzzie's goal is to make
> a clone, mine is to make something new and hopefully better AL-wise. We had
> a very interesting discussion about how to make a better biochemistry the
> other day in IRC.

And someday we'll even have the computing power to actually run it...
;-)
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 16, 2005 3:14:38 AM

emmel InSaNiTised:

>>
>> emmel wrote:
>>
>>> On 2005-02-09, bd <bdonlan@bd.beginyourfear.com> wrote:
>>>>
>>>> emmel wrote:
>>>>| On 2005-02-08, bd <bdonlan@bd.beginyourfear.com> wrote:
>>>>|>|>fuzzie's working on a clone of CEE.
>>>>|>|
>>>>|>|
>>>>|>| CEE? I might be stupid, but I have no clue...
>>>>|>
>>>>|>Creatures Evolution Engine. The game engine for CA, C3, DS, Sea
>>>>|>monkeys, etcetc.
>>>>|
>>>>|
>>>>| It's called that?
>>>>
>>>> Yep. Sometimes C2E - hence fuzzie's "openc2e" project.
>>>
>>> Interesting.
>>
>> In Linux I think the binary was called lc2e.
>
> Yeah, I recently checked. I'm feeling stupid now.
>
>> I don't like it much, myself, so I'm rolling my own. Nothing very
>> impressive is done yet, but I'm getting there :-)
>
> Keep us updated.
>
>> It will be interesting to see what comes out of it. Fuzzie's goal is to
>> make a clone, mine is to make something new and hopefully better AL-wise.
>> We had a very interesting discussion about how to make a better
>> biochemistry the other day in IRC.
>
> And someday we'll even have the computing power to actually run it...
> ;-)

Hehe. Actually it was based on binary strings for chemicals that define what
it can and can't bind with, and how well.

--
- nornagon
http://www.nornrock.com
mailto: nornagon@gmail.com
DS Species range: 10001-10100
Anonymous
February 16, 2005 8:11:57 PM

<snip>
> Well, bd came up with a very nice idea: Simulate chemicals with bit
> fields.
> Perhaps we could simulate the interactions between chemicals without
> accurately modelling anything.
>
> For the example, I tried to describe carbon monoxide poisoning in this
> model. The goal is to make it close to the way it happens for humans:
>
> CO binds with hemoglobin, but it binds so strongly it makes the red cell
> useless for transporting oxygen. Death happens wen enough red cells are
> taken out of circulation, and the organism suffocates.
>
> First, we define the chemicals in terms of binary strings:
> Oxygen: 110100
> CO: 110110
> hemoglobin: 001000
<snip>

For the record: haemoglobin/hemoglobin isn't actually a chemical. However,
the idea(s) is/are... definitely interesting.

--
User of 'Thingamajig!'
Refractor Dragon -=(UDIC)=-
Anonymous
February 17, 2005 8:05:12 PM

news:ZFOQd.438825\$A7.623607@telenews.teleline.es...
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
>
>>
>> For the record: haemoglobin/hemoglobin isn't actually a chemical.
> Well, I'd say conceptually it's about the same thing. Of course it's
> *huge*
> compared to O2, but both are groups of atoms, so I think it's quite fair
> to
> include them in the same category.

Hmm. Conceptually... they work in the same sorts of ways, yes. (Note to
selves: Find a good definition for 'chemical'.)

> Now, I never liked pseudochemicals like "pain". I'd like to get rid of
> them
> somehow...

Agreed. *considers* What about having a conceptual 'mental model' of a
Norn's body within its brain somehow (or just skip that bit if you're not
differentiating between different locations where pain was applied to...
which actually, if we remember correctly, the original Creatures games never
die), and sending the pain signal directly to the brain along the
metaphorical/semi-implemented nervous system, ignoring chemicals and the
metaphorical/semi-implemented bloodstream altogether? Let the brain deal
with memory, instead of needing the chemical to stay in the bloodstream for
a set/degrading time... or just put it into a function of the body/skin,
that pain signals would be sent and continue to be sent, gradually
decreasing--in any case, the brain should deal with memory, associations and
the like, in whatever form(s).

....organs, now... hmm.

Random semi-ethical question: should a Norn suffering heart failure feel
(presumably copious amounts of) pain relating to it? And also, this
highlights a limitation of the 'undifferentiated pain' system--one would
imagine that, in a version at some point, a Norn should be able to recognise
the symptoms of its equivalent of a heart attack and go to get Sidhe-related
medical attention--or actually, there may be a way to get around it. Maybe
such a Norn, suffering extreme pain for no understood reason, with no
association to it, no clear reason, and nothing that could be done to stop
it... maybe such a Norn could be taught to seek such medical attention,
which would make a diagnosis and prescribe/apply the appropriate treatment,
without the Norn in question ever having to know anything more about the
pain other than that it existed.

....still. Three, four, or five versions onward... one would want to have a
Norn that could have a rudimentary understanding of its own body
functionality, be able to perceive cause and effect as relating to its
health, and be able to (rudimentarily in some cases, accurately in others)
diagnose 'problems' and treat them on their own, and later inform the
Hand--or write it down.

*[insert further future-related dreams here]*

>> However,
>> the idea(s) is/are... definitely interesting.
>>
>
> Thanks :-) Now, it needs improvement of course. Such as coming up with a
> way
> of separating complex chemicals. For example, how to model the ATP and ADP
> stuff in this way, glucose and glycogen, etc.

[see bd's posts]

--
User of 'Thingamajig!'
Refractor Dragon -=(UDIC)=-
Anonymous
February 18, 2005 12:32:11 PM

On 2005-02-15, nornagon <nornagon@gmail.com> wrote:
> emmel InSaNiTised:
>
>>> It will be interesting to see what comes out of it. Fuzzie's goal is to
>>> make a clone, mine is to make something new and hopefully better AL-wise.
>>> We had a very interesting discussion about how to make a better
>>> biochemistry the other day in IRC.
>>
>> And someday we'll even have the computing power to actually run it...
>> ;-)
>
> Hehe. Actually it was based on binary strings for chemicals that define what
> it can and can't bind with, and how well.

<g> Maybe we should lease computing power from some alien empire then...
all of their computing power. Then again the ping would be enourmous.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 18, 2005 12:37:05 PM

>
> emmel wrote:
>
>>> I don't like it much, myself, so I'm rolling my own. Nothing very
>>> impressive is done yet, but I'm getting there :-)
>>
>> Keep us updated.
> I will :-)
>
> See also the GE forums. Posted there as well because I wasn't getting much
> of a reaction here in the beginning.

What do you expect with that few people around? It's a bit busier now,
but. Doesn't mean nobody's listening, though.

>>> It will be interesting to see what comes out of it. Fuzzie's goal is to
>>> make a clone, mine is to make something new and hopefully better AL-wise.
>>> We had a very interesting discussion about how to make a better
>>> biochemistry the other day in IRC.
>>
>> And someday we'll even have the computing power to actually run it...
>> ;-)
>
> Well, bd came up with a very nice idea: Simulate chemicals with bit fields.
> Perhaps we could simulate the interactions between chemicals without
> accurately modelling anything.

Too low level :-?

> For the example, I tried to describe carbon monoxide poisoning in this
> model. The goal is to make it close to the way it happens for humans:
>
> CO binds with hemoglobin, but it binds so strongly it makes the red cell
> useless for transporting oxygen. Death happens wen enough red cells are
> taken out of circulation, and the organism suffocates.
>
> First, we define the chemicals in terms of binary strings:
> Oxygen:Â Â Â Â Â 110100
> CO:Â Â Â Â Â Â Â Â Â 110110
> hemoglobin: 001000
>
> A reaction combines chemicals with a XOR. The reaction is defined in terms
> of bits, not chemical numbers. The reaction only works if the bits don't
> clash (ChemA AND ChemB == 0)
>
> Reaction:
> 110110 + 001000 = 111110
>
> The reaction can also happen if there isn't a complete match - this way more
> than one chemical can fit. But the product will be different - a XOR of the
> reactants. So we get:
>
> CO + hemoglobin: 110110 + 001000 = 111110
> O2 + hemoglobin: 110100 + 001000 = 111100
>
> Now, the next stage in the respiration process is perhaps something like
> this:
>
> 111100 + 000011 = 111111 (O2 hemoglobin + blah = energy)
> 111110 + 000011 = clash (doesn't work)
>
> In this stage, it turns out that the product of CO + hemoglobin doesn't work
> for the respiration process, which results in hemoglobin being taken out of
> circulation. And this way we get CO poisoning as a side effect, instead of
> because the genome defined it explicitly.

It *has* certain advantages. Actually it sounds pretty much like the
things defined in the Norn genomes.

> The above is of course incomplete - we need to come up with doing more
> things, like separating big chemicals into pieces - but this is the basic
> idea of what I'd like biochemistry to be like.
>
> It's of course a lot more complex than the CL model. This wouldn't be done
> by hand, of course. We'd have some program that would generate random
> binary strings following a list of requirements for the initial genome.
>
> So, what do you think? I'm very interested in discussing this part.

In fact I miss to see why that's be so much more complex than the CL
model. Enlighten me, will you?

> [darn knode with its similar icons... clicked the wrong one and replied only
> by email the first time. Took me a while to realize why it wasn't showing
> up on AGC. *moves reply by email button*]

Use slrn.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 18, 2005 12:41:08 PM

>
>
>>
>> For the record: haemoglobin/hemoglobin isn't actually a chemical.
> Well, I'd say conceptually it's about the same thing. Of course it's *huge*
> compared to O2, but both are groups of atoms, so I think it's quite fair to
> include them in the same category.

Just a question: what else is it? I mean, chemicals *means* molecules,
doesn't it?

> Now, I never liked pseudochemicals like "pain". I'd like to get rid of them
> somehow...

Hard to do, but then again perhaps not. It depends if you go with 'pain
messenger' as a chemical or want to implement a neural network for that.

>> However,
>> the idea(s) is/are... definitely interesting.
>>
>
> Thanks :-) Now, it needs improvement of course. Such as coming up with a way
> of separating complex chemicals. For example, how to model the ATP and ADP
> stuff in this way, glucose and glycogen, etc.

Well, actually ATP and ADP are not *that* complex if you focus on what
it does rather than what it is made up off.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 18, 2005 12:47:39 PM

On 2005-02-17, The Triad <wanderer@beeb.web> wrote:
> news:ZFOQd.438825\$A7.623607@telenews.teleline.es...
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>>
>>>
>>> For the record: haemoglobin/hemoglobin isn't actually a chemical.
>> Well, I'd say conceptually it's about the same thing. Of course it's
>> *huge*
>> compared to O2, but both are groups of atoms, so I think it's quite fair
>> to
>> include them in the same category.
>
> Hmm. Conceptually... they work in the same sorts of ways, yes. (Note to
> selves: Find a good definition for 'chemical'.)

See above.

>> Now, I never liked pseudochemicals like "pain". I'd like to get rid of
>> them
>> somehow...
>
> Agreed. *considers* What about having a conceptual 'mental model' of a
> Norn's body within its brain somehow (or just skip that bit if you're not
> differentiating between different locations where pain was applied to...
> which actually, if we remember correctly, the original Creatures games never
> die), and sending the pain signal directly to the brain along the
> metaphorical/semi-implemented nervous system, ignoring chemicals and the
> metaphorical/semi-implemented bloodstream altogether? Let the brain deal
> with memory, instead of needing the chemical to stay in the bloodstream for
> a set/degrading time... or just put it into a function of the body/skin,
> that pain signals would be sent and continue to be sent, gradually
> decreasing--in any case, the brain should deal with memory, associations and
> the like, in whatever form(s).

So we are going to model a neuronal network?

> ...organs, now... hmm.
>
> Random semi-ethical question: should a Norn suffering heart failure feel
> (presumably copious amounts of) pain relating to it? And also, this

Most probably.

> highlights a limitation of the 'undifferentiated pain' system--one would
> imagine that, in a version at some point, a Norn should be able to recognise
> the symptoms of its equivalent of a heart attack and go to get Sidhe-related
> medical attention--or actually, there may be a way to get around it. Maybe
> such a Norn, suffering extreme pain for no understood reason, with no
> association to it, no clear reason, and nothing that could be done to stop
> it... maybe such a Norn could be taught to seek such medical attention,
> which would make a diagnosis and prescribe/apply the appropriate treatment,
> without the Norn in question ever having to know anything more about the
> pain other than that it existed.

Interesting.

> ...still. Three, four, or five versions onward... one would want to have a
> Norn that could have a rudimentary understanding of its own body
> functionality, be able to perceive cause and effect as relating to its
> health, and be able to (rudimentarily in some cases, accurately in others)
> diagnose 'problems' and treat them on their own, and later inform the
> Hand--or write it down.
>
> *[insert further future-related dreams here]*

Consciousness.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 18, 2005 12:54:48 PM

<snip>

> Too much insulin will result in hipoglucemia, which results in ATP not
> getting produced, which results in death quite fast. That's easy.
>
> Lack of properly working glucagon probably results in bad resistance against
> low levels of glucose. Creature would probably do fine, except when very
> active. Would faint and probably recover consciousness all by itself.
>
> Now, too much glucose is a problem. From looking at wikipedia, it's hard to
> reproduce in a straightforward way, since it doesn't seem so immediately
> dangerous. According to it, the result is pH changes, problems due to
> osmosis which results in dehydration, etc.
>
> For these cases, we probably could make the reaction rate dependent on the
> amount of the chemical present. When there's a huge amount of glucose
> available, it's everywhere and all reactions that involve it happen much
> faster than expected. That probably end resulting in something bad. Perhaps
> overheating due to a huge increase in energy production?

Just a thought: What about (maybe just in the beginning) to model
chemistry not closely to what goes on in *our* bodies, but instead to
create a completely new system. There's no reason for sticking with
borders defined by real life chemistry and reactions. Of course that's
take away that great learning-about-what-happens-in-a-body effect of
Creatures, but then again simplifying things is very human. Anyone here
wanting to try and explain a, say, five year old why a plane flies?
*Really* flies?
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 18, 2005 11:26:08 PM

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

emmel wrote:
| On 2005-02-17, The Triad <wanderer@beeb.web> wrote:
|

[much snip]

|>*[insert further future-related dreams here]*
|
|
| Consciousness.

A specification would be nice...
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCFpWvek8FNvG1IuYRA1PVAJ4uvrCu/vMBEIZR+dK1PmPVYMgxYACfaZ/4
QbvqQykuGY/ktDlVZGhtuoU=
=J8WT
-----END PGP SIGNATURE-----
Anonymous
February 18, 2005 11:28:33 PM

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

| emmel wrote:
|
|
|>>What do you expect with that few people around? It's a bit
|>>busier now, but. Doesn't mean nobody's listening, though.
|
| I didn't mean it as criticism. And the conversation here is more
| technical too, which is nice :-)
|
|
|>>>So, what do you think? I'm very interested in discussing this
|>>>part.
|>>
|>>In fact I miss to see why that's be so much more complex than
|>>the CL model. Enlighten me, will you?
|
| Ok. First, we're trying to simulate chemistry closer to the real
| world. This means that those binary strings are our molecules,
| and they really need to make sense in the global view of things.
|
| For example, we can't just have:
|
| 1 Starch -> 3 glucose
|
| Or anything like it. Starch would be defined as a long string of
| 1's and 0's that can be split in several units of glucose by an
| enzyme coded by the norn's genome. And a change in the enzyme
| could cause it to split starch in something different, a new
| chemical.
|
|
| The complexity of this is that we have to come up with the
| formula for every compound, and in such a way that they have the
| right interactions. The initial creation will probably be done
| by a program.

How about we throw some random genomes in and run it on fast mode for a
while?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCFpZAek8FNvG1IuYRA5ibAKCGKxCG2UiPZZPuwgCyJ6nehj5BQgCeLMVW
SVTb/vPtUrC9OjOXRAom4Y8=
=Opya
-----END PGP SIGNATURE-----
Anonymous
February 19, 2005 3:23:18 AM

emmel InSaNiTised:

> On 2005-02-15, nornagon <nornagon@gmail.com> wrote:
>> emmel InSaNiTised:
>>
>>>> It will be interesting to see what comes out of it. Fuzzie's goal is to
>>>> make a clone, mine is to make something new and hopefully better
>>>> AL-wise. We had a very interesting discussion about how to make a
>>>> better biochemistry the other day in IRC.
>>>
>>> And someday we'll even have the computing power to actually run it...
>>> ;-)
>>
>> Hehe. Actually it was based on binary strings for chemicals that define
>> what it can and can't bind with, and how well.
>
> <g> Maybe we should lease computing power from some alien empire then...
> all of their computing power. Then again the ping would be enourmous.

....not if you do crazy quantum spin things!

--
- nornagon
http://www.nornrock.com
mailto: nornagon@gmail.com
DS Species range: 10001-10100
Anonymous
February 19, 2005 3:23:19 AM

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

nornagon wrote:
| emmel InSaNiTised:
|
|
|>On 2005-02-15, nornagon <nornagon@gmail.com> wrote:
|>
|>>emmel InSaNiTised:
|>>
|>>
|>>>>It will be interesting to see what comes out of it. Fuzzie's goal is to
|>>>>make a clone, mine is to make something new and hopefully better
|>>>>AL-wise. We had a very interesting discussion about how to make a
|>>>>better biochemistry the other day in IRC.
|>>>
|>>>And someday we'll even have the computing power to actually run it...
|>>>;-)
|>>
|>>Hehe. Actually it was based on binary strings for chemicals that define
|>>what it can and can't bind with, and how well.
|>
|><g> Maybe we should lease computing power from some alien empire then...
|>all of their computing power. Then again the ping would be enourmous.
|
|
| ...not if you do crazy quantum spin things!
|

The ping still takes ages. It just makes a whirring sound from the
spinning particles as it approaches.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCFpj6ek8FNvG1IuYRAwiMAJ90HrCri+KcKg7mZ6GXxwXWRDTxtgCaAw0m
2bmFI3SsduYazzv0SbUvO2A=
=EtQG
-----END PGP SIGNATURE-----
Anonymous
February 19, 2005 10:46:13 PM

bd InSaNiTised:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: RIPEMD160
>
> | emmel wrote:
> |
> |
> |>>What do you expect with that few people around? It's a bit
> |>>busier now, but. Doesn't mean nobody's listening, though.
> |
> | I didn't mean it as criticism. And the conversation here is more
> | technical too, which is nice :-)
> |
> |
> |>>>So, what do you think? I'm very interested in discussing this
> |>>>part.
> |>>
> |>>In fact I miss to see why that's be so much more complex than
> |>>the CL model. Enlighten me, will you?
> |
> | Ok. First, we're trying to simulate chemistry closer to the real
> | world. This means that those binary strings are our molecules,
> | and they really need to make sense in the global view of things.
> |
> | For example, we can't just have:
> |
> | 1 Starch -> 3 glucose
> |
> | Or anything like it. Starch would be defined as a long string of
> | 1's and 0's that can be split in several units of glucose by an
> | enzyme coded by the norn's genome. And a change in the enzyme
> | could cause it to split starch in something different, a new
> | chemical.
> |
> |
> | The complexity of this is that we have to come up with the
> | formula for every compound, and in such a way that they have the
> | right interactions. The initial creation will probably be done
> | by a program.
>
> How about we throw some random genomes in and run it on fast mode for a
> while?

Heh. Could work. With like four crays working on it

--
- nornagon
http://www.nornrock.com
mailto: nornagon@gmail.com
DS Species range: 10001-10100
Anonymous
February 19, 2005 10:46:14 PM

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

nornagon wrote:
| bd InSaNiTised:
|
|
|>-----BEGIN PGP SIGNED MESSAGE-----
|>Hash: RIPEMD160
|>
|>| emmel wrote:
|>|
|>|
|>|>>What do you expect with that few people around? It's a bit
|>|>>busier now, but. Doesn't mean nobody's listening, though.
|>|
|>| I didn't mean it as criticism. And the conversation here is more
|>| technical too, which is nice :-)
|>|
|>|
|>|>>>So, what do you think? I'm very interested in discussing this
|>|>>>part.
|>|>>
|>|>>In fact I miss to see why that's be so much more complex than
|>|>>the CL model. Enlighten me, will you?
|>|
|>| Ok. First, we're trying to simulate chemistry closer to the real
|>| world. This means that those binary strings are our molecules,
|>| and they really need to make sense in the global view of things.
|>|
|>| For example, we can't just have:
|>|
|>| 1 Starch -> 3 glucose
|>|
|>| Or anything like it. Starch would be defined as a long string of
|>| 1's and 0's that can be split in several units of glucose by an
|>| enzyme coded by the norn's genome. And a change in the enzyme
|>| could cause it to split starch in something different, a new
|>| chemical.
|>|
|>|
|>| The complexity of this is that we have to come up with the
|>| formula for every compound, and in such a way that they have the
|>| right interactions. The initial creation will probably be done
|>| by a program.
|>
|>How about we throw some random genomes in and run it on fast mode for a
|>while?
|
|
| Heh. Could work. With like four crays working on it
|

Or just a bunch of PCs. Norn@Home?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCF5dDek8FNvG1IuYRA0HZAKCDw5KBa8zx0SVpyIuGo0pjny98MACfbFIA
uW7jfeX2637QvLkWrlVj3SM=
=wh8A
-----END PGP SIGNATURE-----
Anonymous
February 21, 2005 12:00:04 PM

On 2005-02-19, bd <bdonlan@bd.beginyourfear.com> wrote:
>
> emmel wrote:
>| On 2005-02-17, The Triad <wanderer@beeb.web> wrote:
>|
>
> [much snip]
>
>|>*[insert further future-related dreams here]*
>|
>|
>| Consciousness.
>
> A specification would be nice...

That's the problem, isn't it? We all know what it is - somehow. But we
don't neccessarily know the same.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 21, 2005 12:01:34 PM

On 2005-02-18, nornagon <nornagon@gmail.com> wrote:
> emmel InSaNiTised:
>
>> On 2005-02-15, nornagon <nornagon@gmail.com> wrote:
>>> emmel InSaNiTised:
>>>
>>>>> It will be interesting to see what comes out of it. Fuzzie's goal is to
>>>>> make a clone, mine is to make something new and hopefully better
>>>>> AL-wise. We had a very interesting discussion about how to make a
>>>>> better biochemistry the other day in IRC.
>>>>
>>>> And someday we'll even have the computing power to actually run it...
>>>> ;-)
>>>
>>> Hehe. Actually it was based on binary strings for chemicals that define
>>> what it can and can't bind with, and how well.
>>
>> <g> Maybe we should lease computing power from some alien empire then...
>> all of their computing power. Then again the ping would be enourmous.
>
> ...not if you do crazy quantum spin things!

No, sorry, but that doesn't work. You still have to transfer the date at
the speed of light - at most.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 21, 2005 12:05:25 PM

>
> emmel wrote:
>
>>
>>> ...still. Three, four, or five versions onward... one would
>>> want to have a Norn that could have a rudimentary
>>> understanding of its own body functionality, be able to
>>> perceive cause and effect as relating to its health, and be
>>> able to (rudimentarily in some cases, accurately in others)
>>> diagnose 'problems' and treat them on their own, and later
>>> inform the Hand--or write it down.
>>>
>>> *[insert further future-related dreams here]*
>>
>> Consciousness.
>
> Mmm. Does it imply consciousness, though? Surely we're not the
> only animal who can associate bad food to a stomach ache?

Well, associating it with concepts like getting a doctor *has*. Abstract
thinking implies consciousness IMHO.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 21, 2005 12:11:13 PM

>
> emmel wrote:
>
>> Just a thought: What about (maybe just in the beginning) to
>> model chemistry not closely to what goes on in *our* bodies,
>> but instead to create a completely new system. There's no
>> reason for sticking with borders defined by real life chemistry
>> and reactions. Of course that's take away that great
>> learning-about-what-happens-in-a-body effect of Creatures, but
>> then again simplifying things is very human. Anyone here
>> wanting to try and explain a, say, five year old why a plane
>> flies? *Really* flies?
>
> Well, that just sounds harder to me. I mean, evolution came up
> with all those mechanisms that work quite well and make sense.
> It'd be awfully hard to come up with something like that.
>
> Besides, if the goal is to replicate life, it seems to make sense
> to try to immitate existing life first.
>
> Of course if you think you can roll your own, please do so :-)
> I'd like to see several completely different kinds of creatures
> in my world.

Life has developed this way because of the environment. There is no
reason metabolism could be completely different without oxygen for
example - and still work. The question is if we try to model life as
closely as possible or creating something close enough to our
environment to get an AI to understand us. I mean that's the goal, isn't
it?
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
Anonymous
February 21, 2005 5:16:31 PM

"emmel" <the_emmel*whatever*@gmx.net> wrote in message
news:slrnd1bebp.h8s.the_emmel*whatever*@btcips73x3.cip.uni-bayreuth.de...
> On 2005-02-17, The Triad <wanderer@beeb.web> wrote:
>> news:ZFOQd.438825\$A7.623607@telenews.teleline.es...
>>> -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA1
>>>
>>>
>>>>
>>>> For the record: haemoglobin/hemoglobin isn't actually a chemical.
>>> Well, I'd say conceptually it's about the same thing. Of course it's
>>> *huge*
>>> compared to O2, but both are groups of atoms, so I think it's quite fair
>>> to
>>> include them in the same category.
>>
>> Hmm. Conceptually... they work in the same sorts of ways, yes. (Note
>> to
>> selves: Find a good definition for 'chemical'.)
>
> See above.

(Er?)

>>> Now, I never liked pseudochemicals like "pain". I'd like to get rid of
>>> them
>>> somehow...
>>
>> Agreed. *considers* What about having a conceptual 'mental model' of a
>> Norn's body within its brain somehow (or just skip that bit if you're not
>> differentiating between different locations where pain was applied to...
>> which actually, if we remember correctly, the original Creatures games
>> never
>> die), and sending the pain signal directly to the brain along the
>> metaphorical/semi-implemented nervous system, ignoring chemicals and the
>> metaphorical/semi-implemented bloodstream altogether? Let the brain deal
>> with memory, instead of needing the chemical to stay in the bloodstream
>> for
>> a set/degrading time... or just put it into a function of the body/skin,
>> that pain signals would be sent and continue to be sent, gradually
>> decreasing--in any case, the brain should deal with memory, associations
>> and
>> the like, in whatever form(s).
>
> So we are going to model a neuronal network?

Crudely. If you'll recall, the original Creatures games did it to some
extent, though hopefully the imagined version could be at least a step or so
further along in complexity.

>> ...organs, now... hmm.
>>
>> Random semi-ethical question: should a Norn suffering heart failure feel
>> (presumably copious amounts of) pain relating to it? And also, this
>
> Most probably.

*nods*

>> highlights a limitation of the 'undifferentiated pain' system--one would
>> imagine that, in a version at some point, a Norn should be able to
>> recognise
>> the symptoms of its equivalent of a heart attack and go to get
>> Sidhe-related
>> medical attention--or actually, there may be a way to get around it.
>> Maybe
>> such a Norn, suffering extreme pain for no understood reason, with no
>> association to it, no clear reason, and nothing that could be done to
>> stop
>> it... maybe such a Norn could be taught to seek such medical attention,
>> which would make a diagnosis and prescribe/apply the appropriate
>> treatment,
>> without the Norn in question ever having to know anything more about the
>> pain other than that it existed.
>
> Interesting.

Thank you.

>> ...still. Three, four, or five versions onward... one would want to
>> have a
>> Norn that could have a rudimentary understanding of its own body
>> functionality, be able to perceive cause and effect as relating to its
>> health, and be able to (rudimentarily in some cases, accurately in
>> others)
>> diagnose 'problems' and treat them on their own, and later inform the
>> Hand--or write it down.
>>
>> *[insert further future-related dreams here]*
>
> Consciousness.

Ah. Yes. The question is--how do you tell when you've attained it?
/Seeming/ conciousness, at least... an individual whom one could speak to
as an equal, and could speak back to one in the same fashion...

Somehow, I/we think that Darwinian evolution is important. The more that
things are hardwired, the less the likelihood that there will ever evolve
anything greater than the original creators forged.

*considers* On the subject of neural networks... one idea, one concept, is
to first make certain that the system of Darwinian evolution, particularly
as applied to the brain, is in place. Then, start hatching Norns with
psyches various degrees along the spectrum. There, a near-completely
hardwired Norn, there a Norn with general instincts, here more complex
instincts, there more simple instincts, a few with completely random Kohonen
networks, others only semi-random, a careful order...

The 'organic' machine which would create the eggs would, with the aid of
other systems (from an in-game perspective), keep tabs on the hatchees--and
produce ever more subtle combinations, experimenting, combining, selecting
as it chooses the Norn with the best possible chance--within its data, and
the constrains of the simulated physics--of long-term survival. Once the
population rose over a certain threshold, it would probably more-or-less
stop, because there would be no need for it; however, it would still watch
the genetics and the lifetimes of the Norns, and every so often, output a
single egg (ever read Diaspora, by Greg Egan? Such a Norn would be a little
like an Orphan...) to add to the gene pool, to add to those existing.

....the Kohonen networks and the like... that data we would provide the
'machines' with from the start, as creators. However, the precise
/selection/ process would only be coded in based on basic perception, the
length of a life for example. If a Hand was overly protective, and so his
Norns didn't reach their full potential... well, so be it. That's what
would actually happen! That's realistic! Survival of the fittest, or the
most intelligent...

Some day. Some day, it would be wonderful, even if we couldn't understand
the Norn that had been born of our creation, to see such a Norn that was
truly concious, truly intelligent, and could truly think as well or better
as one of us.

And after that... what next? Why should an evolving intelligence stop at
present human level? Speed time up by thousands, millions of days to a
second, leave a Wolfing Run to go on for longer than any have been able to
try or any Norns have been able to take full advantage of, and see what is
found when time returns to normal...!

*smiles*

....some day...

--
User of 'Thingamajig!'
Refractor Dragon -=(UDIC)=-
Anonymous
February 21, 2005 5:20:37 PM

"emmel" <the_emmel*whatever*@gmx.net> wrote in message
news:slrnd1j90h.5rf.the_emmel*whatever*@btcips73x6.cip.uni-bayreuth.de...
>>
>> emmel wrote:
>>
>>>
>>>> ...still. Three, four, or five versions onward... one would
>>>> want to have a Norn that could have a rudimentary
>>>> understanding of its own body functionality, be able to
>>>> perceive cause and effect as relating to its health, and be
>>>> able to (rudimentarily in some cases, accurately in others)
>>>> diagnose 'problems' and treat them on their own, and later
>>>> inform the Hand--or write it down.
>>>>
>>>> *[insert further future-related dreams here]*
>>>
>>> Consciousness.
>>
>> Mmm. Does it imply consciousness, though? Surely we're not the
>> only animal who can associate bad food to a stomach ache?
>
> Well, associating it with concepts like getting a doctor *has*. Abstract
> thinking implies consciousness IMHO.

Hmm. For the record, the example in question doesn't necessarily require
conciousness--said Norn suddenly feels great amounts of pain, and nothing to
associate it with, nothing to stop doing or get away from. Since it cannot
stop it itself, it goes to what /will/ stop it, and pushes what it has been
taught to push (or the equivalent) in such situations--it drinks, or eats,
or whatever that which has been delivered--and the pain disappears. The
association between the ending of unstoppable pain and that particular
course of action is strengthened, even if the Norn cannot understand,
abstractly, exactly what just happened [as we would understand it].

....associations... /patterns/, really. It all comes down to patterns. One
of the brain's primary purposes... to ascertain patterns, to make
associations, to classify, to categorise...

Hmm.

--
User of 'Thingamajig!'
Refractor Dragon -=(UDIC)=-
Anonymous
February 21, 2005 5:21:14 PM

"emmel" <the_emmel*whatever*@gmx.net> wrote in message
news:slrnd1j8mf.5rf.the_emmel*whatever*@btcips73x6.cip.uni-bayreuth.de...
> On 2005-02-19, bd <bdonlan@bd.beginyourfear.com> wrote:
>>
>> emmel wrote:
>>| On 2005-02-17, The Triad <wanderer@beeb.web> wrote:
>>|
>>
>> [much snip]
>>
>>|>*[insert further future-related dreams here]*
>>|
>>|
>>| Consciousness.
>>
>> A specification would be nice...
>
> That's the problem, isn't it? We all know what it is - somehow. But we
> don't neccessarily know the same.

We perceive it--and yet we don't understand it. We don't understand its
cause, its roots.

--
User of 'Thingamajig!'
Refractor Dragon -=(UDIC)=-
Anonymous
February 21, 2005 5:28:59 PM

"emmel" <the_emmel*whatever*@gmx.net> wrote in message
news:slrnd1bdvi.h8s.the_emmel*whatever*@btcips73x3.cip.uni-bayreuth.de...
>>
>>
>>>
>>> For the record: haemoglobin/hemoglobin isn't actually a chemical.
>> Well, I'd say conceptually it's about the same thing. Of course it's
>> *huge*
>> compared to O2, but both are groups of atoms, so I think it's quite fair
>> to
>> include them in the same category.
>
> Just a question: what else is it? I mean, chemicals *means* molecules,
> doesn't it?

Mm. Not certain. I could have a piece of wood in front of me, and you
might acknowledge that there were chemicals in there, but you wouldn't call
wood a chemical--haemoglobin is (as far as I'm aware) composed of chemicals,
and of elements, but it itself is not necessarily either a chemical or an
element.

Or possibly we're overestimating its complexity, misremembering its
structure.

>> Now, I never liked pseudochemicals like "pain". I'd like to get rid of
>> them
>> somehow...
>
> Hard to do, but then again perhaps not. It depends if you go with 'pain
> messenger' as a chemical or want to implement a neural network for that.

*nods* ...or pseudo-neural network. Instead of--oh, skah.

Just realised... 'Hunger' and 'Tiredness' and the like were 'chemicals' as
well, even though by right they should have existed in the brain, nothing to
do with the bloodstream at all... augh.

How much does the (Norn) brain do? How much /should/ it do?

....centralisation... we need to make the brain more self-contained, need to
make it... heh.

Ironic, in a way. In order to try to attain an ideal, we actually have to
make it /harder/ for us to mess around with their minds. We can recall
injecting Norns with Sleepiness, or Tiredness, and having them flopping down
to the ground... now, a chemical compound which /caused/ them to feel tired
or sleepy by affecting their brain, yes! But not the pure essense of
emotion! Skah it, you could even inject them with Anger!

....cause and effect... we have to make it indirect, otherwise the very
means that make the meddling possible will hold the brain back from the
heights it could achieve if only it was self-contained, a separate system...

....or maybe we're babbling. But it might, somehow, help. Hopefully.

>>> However,
>>> the idea(s) is/are... definitely interesting.
>>>
>>
>> Thanks :-) Now, it needs improvement of course. Such as coming up with a
>> way
>> of separating complex chemicals. For example, how to model the ATP and
>> stuff in this way, glucose and glycogen, etc.
>
> Well, actually ATP and ADP are not *that* complex if you focus on what
> it does rather than what it is made up off.

*nods; smiles*

....yes. Good.

....effects... we're starting from the middle, working upwards. Or maybe
they started at the top, and we want to work from the middle instead, and
maybe one day, a few versions down the line, we'll be able to start close
enough to the bottom that no one will be able to tell the difference without
inventing an Albian electron microscope first! *grins*

Some day...! And so much that can be done in the present, to lead to that
day!

--
User of 'Thingamajig!'
Refractor Dragon -=(UDIC)=-
Anonymous
February 22, 2005 1:46:29 AM

emmel InSaNiTised:

> On 2005-02-18, nornagon <nornagon@gmail.com> wrote:
>> emmel InSaNiTised:
>>
>>> On 2005-02-15, nornagon <nornagon@gmail.com> wrote:
>>>> emmel InSaNiTised:
>>>>
>>>>>> It will be interesting to see what comes out of it. Fuzzie's goal is to
>>>>>> make a clone, mine is to make something new and hopefully better
>>>>>> AL-wise. We had a very interesting discussion about how to make a
>>>>>> better biochemistry the other day in IRC.
>>>>>
>>>>> And someday we'll even have the computing power to actually run it...
>>>>> ;-)
>>>>
>>>> Hehe. Actually it was based on binary strings for chemicals that define
>>>> what it can and can't bind with, and how well.
>>>
>>> <g> Maybe we should lease computing power from some alien empire then...
>>> all of their computing power. Then again the ping would be enourmous.
>>
>> ...not if you do crazy quantum spin things!
>
> No, sorry, but that doesn't work. You still have to transfer the date at
> the speed of light - at most.

Nope. Wrong.

There are pairs of particles, I'm not sure what they're called. But when one of
them has their spin changed, the other one spins the opposite way. Instantly.
Yup, quantum physics is a headf**k.

--
- nornagon
http://www.nornrock.com
mailto: nornagon@gmail.com
DS Species range: 10001-10100
Anonymous
February 22, 2005 1:46:30 AM

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

nornagon wrote:
| emmel InSaNiTised:
|
|
|>On 2005-02-18, nornagon <nornagon@gmail.com> wrote:
|>
|>>emmel InSaNiTised:
|>>
|>>
|>>>On 2005-02-15, nornagon <nornagon@gmail.com> wrote:
|>>>
|>>>>emmel InSaNiTised:
|>>>>
|>>>>
|>>>>>>It will be interesting to see what comes out of it. Fuzzie's goal
is to
|>>>>>>make a clone, mine is to make something new and hopefully better
|>>>>>>AL-wise. We had a very interesting discussion about how to make a
|>>>>>>better biochemistry the other day in IRC.
|>>>>>
|>>>>>And someday we'll even have the computing power to actually run it...
|>>>>>;-)
|>>>>
|>>>>Hehe. Actually it was based on binary strings for chemicals that define
|>>>>what it can and can't bind with, and how well.
|>>>
|>>><g> Maybe we should lease computing power from some alien empire then...
|>>>all of their computing power. Then again the ping would be enourmous.
|>>
|>>...not if you do crazy quantum spin things!
|>
|>No, sorry, but that doesn't work. You still have to transfer the date at
|>the speed of light - at most.
|
|
| Nope. Wrong.
|
| There are pairs of particles, I'm not sure what they're called. But
when one of
| them has their spin changed, the other one spins the opposite way.
Instantly.
| Yup, quantum physics is a headf**k.
|

Actually, no. What happens is two particles are generated with unknown
spins. When the spin of one is observed, it picks a spin at random, and
the other instantly changes to the opposite spin - but since you can't
control the spin you can't send data. This is somehow subtly different
from choosing the spins at time of particle creation.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iQEVAwUBQhpX7fzMPsvJqdmPAQNetggAkPK4lRwHWShq7D294QFibqM+zKqLUqeG
UtdIc2i9p2v8Trc9xAHaAJOaFXVDW3auK9b84BCmM7sZNJ9K8NK6SoASjkI4JlZ6
GmM82dL4vLA9n4cU+mIoBfe8FukckQndr/I3ta26GR/4jzvRInnZl2QAW5dW66ay
3tKU7FGKeo436iMK+lY58GOs3XJGOZX3gvxAhxTYpCmwT4sqOhvdBUnMURrMBqv6
Oj4hxkbObi9FU0/EZpWpQzGqs4LsXC484CSlxVe0KexpvzTyj9x0O2vnWAKSy+2/
TCgUr8WY1PNMJDLJXxj95jsdd04z3e81TBSlh3E/6YIzVCpd8m78mw==
=+tkE
-----END PGP SIGNATURE-----
Anonymous
February 22, 2005 2:57:40 PM

On 2005-02-21, nornagon <nornagon@gmail.com> wrote:
> emmel InSaNiTised:
>
>> On 2005-02-18, nornagon <nornagon@gmail.com> wrote:
>>> emmel InSaNiTised:
>>>
>>>> On 2005-02-15, nornagon <nornagon@gmail.com> wrote:
>>>>> emmel InSaNiTised:
>>>>>
>>>>>>> It will be interesting to see what comes out of it. Fuzzie's goal is to
>>>>>>> make a clone, mine is to make something new and hopefully better
>>>>>>> AL-wise. We had a very interesting discussion about how to make a
>>>>>>> better biochemistry the other day in IRC.
>>>>>>
>>>>>> And someday we'll even have the computing power to actually run it...
>>>>>> ;-)
>>>>>
>>>>> Hehe. Actually it was based on binary strings for chemicals that define
>>>>> what it can and can't bind with, and how well.
>>>>
>>>> <g> Maybe we should lease computing power from some alien empire then...
>>>> all of their computing power. Then again the ping would be enourmous.
>>>
>>> ...not if you do crazy quantum spin things!
>>
>> No, sorry, but that doesn't work. You still have to transfer the date at
>> the speed of light - at most.
>
> Nope. Wrong.
>
> There are pairs of particles, I'm not sure what they're called. But when one of
> them has their spin changed, the other one spins the opposite way. Instantly.
> Yup, quantum physics is a headf**k.

I *know*. The problem is that you can't figure out what it means without
the data you get, when putting information in the other quant. And
*that* has to be transfered by other means - by the speed of light. You
can do quantum teleportaion, but you'll still need an additional data
stream, thus preserving the law of maximum data transport.
--
emmel <the_emmel*you-know-what-that's-for*@gmx.net>
(Don't forget to remove the ** bit)

Official AGC feedback maniac

"God is playing creatures - and we're the norns."

"A hundred dead are a tragedy - a hundred thousand are statistics."

"I guess you can call yourself lucky." -
"I could, but Linda suits me a little better...
Things called lucky tend to get hit by trucks."

Hi, I'm a .sig virus. Just copy me to your .signature. And don't worry.
• 1 / 17
• 2
• 3
• 4
• 5
• More pages
• Next