[Noisebridge-discuss] Philosophy and Computer Language Question

Sat Dec 31 07:29:30 UTC 2011

You made some good points. I'd argue that anything other than simulating sub-atomic particles is simply based on empirical observations and simplifications. It isn't what is really happening, but is rather an attempt to guess at a model.

How accurate is a model? I saw somebody on a forum trying to build a perpetual motion machine in a physics simulator. He didn't seem to understand that the physics simulator was built with the rules of physics in mind, which don't allow that. A perpetual motion machine in the simulator would most likely just demonstrate the limitations of the simulator.

I think most simulation systems are going to be boring from a philosophical perspective.

Here's a quick rundown as an example of how many different levels of simulation accuracy there can be, using a microprocessor:

Highest level: non-real time instruction simulation. The simulator has a list of registers, looks at an instruction, "add register 5 and register 6, put the result in register 7"
This depends on the instructions being accurately described - CPUs often have bugs so a simulator like this will not always have the bugs properly emulated. The simplest version of this will not have any concept of time - it'll just work as fast as possible (a few times to a few hundred times slower than the real processor...)
Some versions of this will have time calculations. They'll tell you how many clock cycles it took to do each calculation.

The next level deeper is the logic level. A processor is a collection of logic gates of various kinds. AND, OR,  NAND, buffers, flip flops, etc. They have digital inputs and outputs. Some systems will simulate at this level. If memory serves, a logic level simulation of a 486 took a week to boot. Very slow. Some of these simulations are time accurate and some are not,

Eg: the clock on a CPU keeps everything in sync. The clock pulse happens, and then all the logic gates change, and, in theory, they have finished changing by the time the next pulse happens. Let's say each gate takes 5 nS from input to output. If the clock pulse is 100 nS long, you'd better not have 25 gates in a row, because the signal won't come out the other end for 125 nS - at which point the next clock pulse will have already started! This will likely cause the hardware to sometimes fail. Some simulators will track the time of each gate changing, while others will just sort of assume that it will work - at each pulse they will follow through all the gates and change the, and assume it will finish on time. It takes longer to simulate accurate time.

Next layer deeper is transistors. Logic gates are made up of transistors, with analog values throughout. You could simulate a processor entirely by transistors, although that is not very commonly done. 

The next layer would be to simulate the individual particles. That would be absolutely insanely complex to do, and I'm not even sure if our knowledge of the particle behaviors would allow us to meaningfully simulate transistors out of them. 

Thanks,

gopi at iPad

On Dec 30, 2011, at 14:13, Ryan Rawson <ryanobjc at gmail.com> wrote:

> See the problem with this 'process philosophy' is it sounds like a
> MODEL of the real world to me, rather than "how it works".  Without an
> evidential backing, how can this process philosophy be anything other
> than wishful thinking?
> 
> As for the computer science side of things, ultimately when you get
> down to it, everything is bits and bytes and CPU instructions. So we
> can talk in terms of floating point precisions, ranges of integers,
> memory sizes, cpu speed, etc.
> 
> And on a more 'out there' moment, it would require a computer with as
> much computing power that is implied by the universe itself to
> simulate the universe. That is we can consider the universe as a
> real-time computing platform that is simulating itself, and all the
> computations that would be required to make it happen would also have
> to be replicated. Eg: all the quantum interactions and other processes
> that build up and cause reality.  Futhermore, we cant simulate reality
> because we dont know how it fully works yet.  So whatever we build as
> a simulation is merely a model of reality, and thus will always
> diverge from it in eventually (presumably?) important ways.
> 
> On Fri, Dec 30, 2011 at 1:08 PM, Caleb Grayson <calebgrayson at gmail.com> wrote:
>> Well.. I'm not sure. I'm more of a philosopher than a computer scientist.
>> 
>> In Whitehead's Process Philosophy he said everything in reality is a
>> function or process that takes in the entire universe at every moment and
>> spits out Actual Occasions that become apart of the Creative Advance, the
>> history of functional  results in time and space.
>> 
>> There is a question as to what time an space are. It is m suspicion that
>> functions and their solutions are not in time and space, but time and space
>> are  in functions and their solutions. Why would times and spaces for which
>> nothing is happening be generated by an efficient system?
>> CS, if I understand correctly, time and space have to be predefined by
>> establishing their numerical domains first.
>> I'm hoping CS in its attempt to simulate reality can give inside into it.
>> 
>> Of course CS being a rational/material system has no place to calculate for
>> spirit/soul outside of its system which Whitehead does allow for.
>> 
>> On Dec 30, 2011, at 12:34 PM, Will Sargent <will.sargent at gmail.com> wrote:
>> 
>>> 
>>> my particular interest is how simulations handle events in time and space.
>>>  how does a simulation taken multiple input from multiple functions or users
>>> and align them in time•space.  what would those functions look like?
>> 
>> 
>> Are you talking about multi-agent systems?
>> 
>> Will.
>> 
>> 
>> _______________________________________________
>> Noisebridge-discuss mailing list
>> Noisebridge-discuss at lists.noisebridge.net
>> https://www.noisebridge.net/mailman/listinfo/noisebridge-discuss
>> 
> _______________________________________________
> Noisebridge-discuss mailing list
> Noisebridge-discuss at lists.noisebridge.net
> https://www.noisebridge.net/mailman/listinfo/noisebridge-discuss