hibbsa
2012-12-17 04:15:57 UTC
I get it that the theory is necessarily bounded by not addressing how a
transformation can be accomplished (the constructor), but whether such a
transformation is possible (i.e. because this distinction is tied in
with its positioning at a fundamental level).
But something you might be envisaging that wouldn't violate that
boundary, would be representing transformations that are conditionally
possible, dependent on whether other transformations take place first,
local to some space.
One reason you might want to do this would be because it seems possible
that in the limit...the resultant network would have to share common
features with the multiverse. Wouldn't it? Or is that wrong?
The multiverse is also constrained by the laws of physics in that from a
given fungible locality the different ways things converge in every
physically possible way.
Isn't that the same as saying that at some locality within in this one
universe, given some set of transformations have taken place, there is a
corresponding set of networks of knock-on transformations, also
constrained by the laws of physics.
At some network property level, how does the multiverse of alternate
worlds look different to the Constructorverse of alternate possibilities
within this one world?
Anyway.....if any of that is relevant then the work linked here could be
of interest to you. What they discover is that so long as they
physically occur in time and space, totally different networks are
asymptotically identical for certain dynamical measures. They show that
this equivalence can be used to simplify analysis and computation of a
given network. They say that some analysis they had to do that
previously took years in a supercomputer was reduced to days, or a day,
using this equivalence.
http://www.nature.com/srep/2012/121113/srep00793/full/srep00793.html
<http://www.nature.com/srep/2012/121113/srep00793/full/srep00793.html>
[Non-text portions of this message have been removed]
transformation can be accomplished (the constructor), but whether such a
transformation is possible (i.e. because this distinction is tied in
with its positioning at a fundamental level).
But something you might be envisaging that wouldn't violate that
boundary, would be representing transformations that are conditionally
possible, dependent on whether other transformations take place first,
local to some space.
One reason you might want to do this would be because it seems possible
that in the limit...the resultant network would have to share common
features with the multiverse. Wouldn't it? Or is that wrong?
The multiverse is also constrained by the laws of physics in that from a
given fungible locality the different ways things converge in every
physically possible way.
Isn't that the same as saying that at some locality within in this one
universe, given some set of transformations have taken place, there is a
corresponding set of networks of knock-on transformations, also
constrained by the laws of physics.
At some network property level, how does the multiverse of alternate
worlds look different to the Constructorverse of alternate possibilities
within this one world?
Anyway.....if any of that is relevant then the work linked here could be
of interest to you. What they discover is that so long as they
physically occur in time and space, totally different networks are
asymptotically identical for certain dynamical measures. They show that
this equivalence can be used to simplify analysis and computation of a
given network. They say that some analysis they had to do that
previously took years in a supercomputer was reduced to days, or a day,
using this equivalence.
http://www.nature.com/srep/2012/121113/srep00793/full/srep00793.html
<http://www.nature.com/srep/2012/121113/srep00793/full/srep00793.html>
[Non-text portions of this message have been removed]