Re: Organic Builder - new levels
Tim Hutton
Discussion about level 19:
On 11/30/05, Ralph Hartley <> wrote:
> Tim Hutton wrote:
> > On 11/27/05, Dave Mann <> wrote:
>
> >>The simplest solution I can think of would be to require that the cell
> >>be able reproduce many times over without the copies interfering with
> >>each other. Is there a problem with this that I haven't thought of apart
> >>from potential difficulties with evaluation?
> >
> > If the caustic agent weren't there, a 'trivial' solution would be to
> > break open the membrane, make a copy of the template (using your
> > crafty non-interfering template-replication rules), and then zip up
> > the two templates into new cells.
>
> That last step isn't so easy. How do you ensure that the chromosome, and
> nothing else, will be on the inside?
Yes, true, that is very tricky. So really the best solution to level 19 (which was where this discussion started) will be to do the full cell division. I'm more happy with this level now, a good solution is possible with only a minimal amount of dragging.
>
> If two cells are doing it at the same time, you can get a mess. The
> trick of only reacting with free atoms won't work when you have to close
> a loop.
>
> In fact, expanding a membrane and transport through one become much
> harder if you have more than one copy.
>
> This suggests some challenges that might be quite tricky:
>
> (1) Copy
>
> a1--a1--a1
> | |
> a1 b2 a1
> | |
> a1--a1--a1
>
> Only b atoms are permitted inside the copies, and exactly one must be
> present in each.
>
> The number of As can be allowed to vary, or not.
>
> (2) Make copies of the molecule in (1) enclosing *all* b atoms (one per
> copy, the number varies), with the additional requirement that the atoms
> of the copies must have the same *states* (1 and 2) as the original, not
> just the same types, nor may any other atoms be bound to the copies.
>
> (3) and (4) are the same as (1) and (2) respectively, but you must copy
>
> t1--u1--v1
> | |
> s1 b2 w1
> | |
> z1--y1--x1
>
> The letters (s-z) are different each time, and must be copied, but none
> are b.
>
> The difficulty of some of these may depend on the size of the loop.
>
> The order of these problems may not be quite right; (2) may be harder
> than (3).
>
> I haven't actually tried any of these yet.
These are great ideas for levels. I'll have a go at solving them and get back to you.
Tim
--
Tim Hutton - http://www.sq3.org.uk
Take the Organic Builder challenge - http://www.sq3.org.uk/Evolution/Squirm3/OrganicBuilder/
Dave Mann
> copy, the number varies), with the additional requirement that the atoms
> of the copies must have the same *states* (1 and 2) as the original, not
> just the same types, nor may any other atoms be bound to the copies.
Is it possible to do this with 100% probability? I can't see a way of
doing it without using vulnerable loop-closing reactions.
If there were a lot of extra A atoms, the original membrane could be
grown and copies could be created inside but solutions of this sort are
*cheats* to get around the interference problem.
As far as I can tell, the only way round this is for the original to
start off with the gate protected from other gates by mechanical means
- two rigid arms for example. The copy could then be grown internally
and all loop-closing reactions completed (including replicating the
protected gate) before the copy was released.
Is there a simpler solution than this?
Dave
Ralph Hartley
>>(2) Make copies of the molecule in (1) enclosing *all* b atoms (one per
>>copy, the number varies), with the additional requirement that the atoms
>>of the copies must have the same *states* (1 and 2) as the original, not
>>just the same types, nor may any other atoms be bound to the copies.
>
> Is it possible to do this with 100% probability?
I think so, but it isn't easy.
> I can't see a way of
> doing it without using vulnerable loop-closing reactions.
There isn't (at least that is what I intended), you have to use *safe*
loop closing reactions.
> If there were a lot of extra A atoms, the original membrane could be
> grown and copies could be created inside but solutions of this sort are
> *cheats* to get around the interference problem.
It would also violate the requirement (inherited from (1)) that each
membrane must enclose one b and nothing else. You could destroy the
outer membrane at the end *if* you knew that there were not a couple of
b atoms left, but you can never be 100% sure of that.
> As far as I can tell, the only way round this is for the original to
> start off with the gate protected from other gates by mechanical means
> - two rigid arms for example.
You can safely close loops with structures you can build from scratch.
See the attached rule-set.
It only works for 3 atom loops (triangles), but that is all you need.
To complete task (2), you would also need to move membrane atoms into
the interior. That also requires closing a triangle, and could be done
in a similar way.
The rest can be done inside.
You *could* even do all the reactions on the outside, but each would
have to be protected, so that would be a lot of work. Closing triangles
depends critically on the amount of space around the atoms, so it would
be hard to re-use code.
Ralph Hartley
Ralph Hartley
> (1) Copy
>
> a1--a1--a1
> | |
> a1 b2 a1
> | |
> a1--a1--a1
>
> Only b atoms are permitted inside the copies, and exactly one must be present in each.
>
> The number of As can be allowed to vary, or not.
>
> copy, the number varies), with the additional requirement that the atoms
> of the copies must have the same *states* (1 and 2) as the original, not
> just the same types, nor may any other atoms be bound to the copies.
> You can safely close loops with structures you can build from scratch.
> See the attached rule-set.
Here is a somewhat more debugged version, and the code I used to test it.
Ralph Hartley