Page 1 of 1

So far, nothing too strange...

PostPosted: Mon May 07, 2007 11:25 pm
by GreatLimmick
"Second stage anneal"? I remember the word "annealing" from my Materials Science course. It has something to do with forging carbon steel. If I could remember more, maybe I could figure out specifically what that has to do with what's going on... but the fact that it involves combining two substances into an alloy doesn't bode terribly well, in my opinion.

Or am I completely off base here?

PostPosted: Tue May 08, 2007 12:12 am
by Cyril_Dran
I have no idea, my brain won't get past "ooooer, big update" and "colortastic!"

PostPosted: Tue May 08, 2007 3:20 am
by The_Fox
Well, considering the word that ASCII used before was "gestalt" and how he talked about them still being a heterogeneous mental salad in the first stage, I figure the point behind the experiment is to turn two minds into one. The problem being pulling them back apart, which should be difficult to do without trauma.

- Jarylan

PostPosted: Tue May 08, 2007 5:12 am
by Kels
I'm just curious how many wheels within wheels there actually are. Cal said the traps ASCII had inserted into the code weren't for him, but to what extent did he anticipate Cal and simply use the simpler traps as distraction, concealing something far more dangerous?

Of course, Cal and his lovely wife aren't above a few tricks of their own, so this should be interesting...

PostPosted: Tue May 08, 2007 6:16 am
by Allan_ecker
I'm trying to reign in my urge to spew latin at people and expect them to catch up with a dictionary, encycolopedia, or convenient silicon process engineer as the case may be, but the term "anneal" is staying in because one doesn't necessarily need to understand its meaning out of context to eventually grasp its meaning in context.

In my line of work, we think of "annealing" as something you do to diamond-configured silicon (silicon microchips are diamonds-- just with different atoms!) to get the dopants to settle into "bonding sites" where they actively participate the crystal's semiconductor properties as opposed to "interstitial sites" where they just increase resistance.

What you do is you heat the silicon up, then allow it to cool very, very slowly, with all the pieces settling in... right where they fit best...

Kind of alarming thought, applying this to people...

PostPosted: Tue May 08, 2007 8:32 am
by Buck
It's also an algorithmic technique for searching solutions in a large problem space. Just saying. ;)

Cal must have gone through ASCII's code amazingly fast. Although I guess Amanda took her time, too.

PostPosted: Tue May 08, 2007 8:54 am
by Kels
Ah, Allan's description makes it make more sense to me now. Up until this point, I'd assumed all ASCII had been wanting to do was make a group-mind, which of course would be near-impossible to separate, but it seems that he plans on going a step beyond that.

As to Amanda taking her time, dude, she just wore out an ANDROID! :twisted:

PostPosted: Tue May 08, 2007 1:30 pm
by Nyamaza
Well, as I see it there's actually two possibilities.

Amanda wore ASCII out, such that he cannot actively function for a while.
Amanda has left ASCII bound/teased/distracted in such a way that he will nto be willing/able to free himself for a while.

So there ARE other possibilities to the Skunk Minx managing to completely wear out a Hundecouph Power Supply.

PostPosted: Tue May 08, 2007 1:32 pm
by Kels
Hey, you have your <strike>fantasies</strike> interpretation, I'll have mine! :twisted:

PostPosted: Tue May 08, 2007 5:22 pm
by Patchwork cat
I seem to remember annealing metal can be to do with hammering or rolling, not good with silicon slices or minds, methinks...
chi

PostPosted: Tue May 08, 2007 6:21 pm
by GreatLimmick
kels wrote:As to Amanda taking her time, dude, she just wore out an ANDROID! :twisted:

He could be worn out emotionally, or his senses may be overwhelmed, rather than simply having exhausted his power supply or structurally damaged.

patchwork cat wrote:I seem to remember annealing metal can be to do with hammering or rolling, not good with silicon slices or minds, methinks...
chi

No, now that Allen described silicon annealing, I remember how it works with steel. The iron/carbon mixture can have a lot of phases depending on weight-percent carbon and temperature. The procedure for annealing steel is basically the same as the procedure for annealing silicon, except that the purpose is to make the finished product hard and durable rather than semiconductive. I'd have to go get the textbook to see the exact properties you're looking for when you're annealing steel, though.

Work-hardening can be helpful, to a degree, with any metal (although doing it too much will make the metal too brittle for the increased hardness to be of any use), but annealing really only works on alloys (giving them some of the properties of composite materials). If you want to use both, you have to do the annealing first, because heating a material enough to anneal it will typically eliminate the internal stress that makes work-hardening worthwhile (by transforming the material into a different phase).

PostPosted: Tue May 08, 2007 9:51 pm
by Schol-R-LEA
On the other side of the hardware/software fence (which, like usual, I am currently straddling - my Comp Arch professor keeps trying to drag me into his VLSI class, and the only reason I didn't was because I was overloaded already) is 'simulated annealing', a heuristic technique for eliminating parts of a large search space if memory serves (I don't know off hand what the connection to metallurgical annealing is).

PostPosted: Tue May 08, 2007 10:22 pm
by Allan_ecker
Correct-o-mundo, Simulated Annealing is a technique for preventing the settling into of Local Optima when searching for a solution to an optimization problem with multiple variables.

This is totally a reference to silicon annealing.

Essentially you compute the optimization factor for each of a hundred or so points in the design space, have them move about randomly, then slowly reduce the amount by which they may move. Ultimately they will tend to rattle down to a few minimum cost points, with one being the lowest: your global minimum.

There's a more advanced technique called "particle swarm" which was actually developed at my old lab by a guy named Jinho Park if I recall correctly...

PostPosted: Wed May 09, 2007 4:42 am
by Kels
Hmm...I'm getting this odd feeling that while ASCII's experiment is to create this "annealed" groupmind, that might not be his final objective. Could it be that he wants to create it in order to do something?

PostPosted: Wed May 09, 2007 7:14 am
by The_Fox
kels wrote:Hmm...I'm getting this odd feeling that while ASCII's experiment is to create this "annealed" groupmind, that might not be his final objective. Could it be that he wants to create it in order to do something?


While I don't rule out the possibility, one must recall that many scientists pursue science for science's sake, with little in the way of imagining concrete outputs. Some of the greatest breakthroughs we have had have been scientists who were simply pursuing something that didn't look useful. And might not have been at the time, but ended up useful later.

And given that this is mad science, and often treated more like a hobby, I have to wonder. It may simply have been a project motivated by curiousity over the possible results.

- Jarylan, with a long post that said very little really.

PostPosted: Wed May 09, 2007 2:50 pm
by GreatLimmick
So, if simulated annealing is a method of working for an optimal solution... What's the problem that ASCII's network is trying to solve?

PostPosted: Thu May 10, 2007 1:18 am
by Buck
They all have in common that they have a "temperature"* that "cools down", eventually settling on a "solution".

I would assume that it will have some mind-altering effect on Cal and Amanda, either temporary or permanent.


* in case of simulated annealing, the temperature is a probability, or is used to compute a probability.

PostPosted: Thu May 10, 2007 7:32 am
by Tetramorpheus
I'm just an English Major, so forgive my lack of knowledge, but...
Is it likely that the little 'magnetic forces' they're feeling are what are supposed to position them correctly in this annealing process?

PostPosted: Fri May 11, 2007 5:13 am
by Patchwork cat
patchwork cat wrote:I seem to remember annealing metal can be to do with hammering or rolling, not good with silicon slices or minds, methinks...
chi

No, now that Allen described silicon annealing, I remember how it works with steel. The iron/carbon mixture can have a lot of phases depending on weight-percent carbon and temperature.


whoa! First, i said 'metal' and i have known this refer to brass- if erroneously (anyway the softer alloys would respond more easily to working)- further it was a loosely phrased joke, so i don't believe it needed quoting..

PostPosted: Fri May 11, 2007 3:05 pm
by GreatLimmick
Sorry, didn't realize it was a joke.

As far as I know, brass could very well be subject to annealing. It's an alloy, after all. I've never seen a phase diagram for brass, though. It depends largely on how mutually soluble they are. Silver and copper, for example, are almost completely soluble at almost all temperatures (IIRC), so they're pretty much impossible to anneal. Iron and carbon, on the other hand, aren't really mutually soluble at all.

As far as I recall, the solubility of metals depends mostly on their crystal structure and atomic radius; the closer, the more soluble.