Technological Challenges

simsim314 · Post by **simsim314** » April 24th, 2014, 4:09 am

All "big" life patterns are built from small "technologies". Gemini for example is based on stable reflector, G->H converter, and Herschel duplicators and conduits. Gemini also uses some special H -> G converter that can pack gliders nicely together. It's also uses 2-3-4 Glider packs for arm moving and reflection operations, and syntheses of Spartan parts with two 90 degree glider collisions. All those technologies were "small discoveries" that accumulated into something like Gemini.

So to improve designs in Life we need some small "technological solutions". So I suggest we post here what we need (preferably for what purpose). And hopefully with time also solve at least partially some of the challenges.

Here is my "request list":

1. A cross signal adjuster (needed for multiple channels replicators): given two signals (say two gliders at any phase and on some known lanes), that would "collide" or intervene each with others "functionality", the adjuster will get those two signal as an input, and will ALWAYS give an output of the same signal on some other output lanes, regardless of the input phase and the location on the lanes. The adjuster can have constant delay (which is always applied).

2. A better G->H converter. The current spartan G->H converter is of about 500 ticks, this slows down all current replicators designs.

3. A simple gate mechanism (or a flip-flop - needed for computational designs). Given a stream of gliders (on some input lane), and another input glider (a key) that could "open and close a gate". If the gate state is "open" the input glider will "close" the gate and the stream will be blocked. If the gate was closed the input glider will "open the gate" and the stream will pass freely. This open-close gate mechanism should be reusable of course and as small as possible.

4. A fast and small stable duplicator. Like the Snark but instead one output glider this one will have two (or more). As with Snark small, stable, and fast are qualities that are needed.

Post by **dvgrn** » April 25th, 2014, 2:12 pm

simsim314 wrote:... to improve designs in Life we need some small "technological solutions". So I suggest we post here what we need (preferably for what purpose). And hopefully with time also solve at least partially some of the challenges.

1. A cross signal adjuster (needed for multiple channels replicators): ...

This is a somewhat trickier project than the old p30/p46/p60 signal crossers (if I'm understanding it correctly). If there are no constraints at all on the timing of Glider 1 and Glider 2, and the signals have to cross each other and end up at known spacetime locations relative to their inputs... well, I think this would take some creative use of multiple regulators, which would have to be redesigned to be constructible.

Even minimal constraints would make this problem much easier. If you can guarantee that Glider 1 and Glider 2 will both show up within some number of ticks, for example, and that you don't need the outputs to appear until after both input gliders have been processed, then I think you can get away with just the reaction at the heart of the regulator, plus some signal splitters and miscellaneous wiring. If you can guarantee that Glider 1 will appear before Glider 2, that also makes the problem relatively trivial.

A similar problem that I've had on the back burner for a while is an exact-timing merge circuit (ETM). If you want to combine two signal gliders that may be either present or absent, G1@(X1, Y1, T1) and G2@(X2, Y2, T2), how do you get G1 OR G2 = G3@(X3, Y3, T3)? I've been slowly working out various non-Spartan solutions for this -- they do exist, and they're gradually getting more compact. But really it seems as if there should be a small transparent glider-emitting Herschel converter that can do this -- inside a 20x30 box, let's say -- by producing a spark that suppresses an incoming glider, and then replacing it a little later with an exactly-timed equivalent glider. No luck there so far.

simsim314 wrote:2. A better G->H converter. The current spartan G->H converter is of about 500 ticks, this slows down all current replicators designs.

Hear, hear! The Silver reflector/converter was a wonderful discovery in its time, but there must be something smaller that can be found with modern computational power. Distributed search, anyone?

simsim314 wrote:3. A simple gate mechanism (or a flip-flop - needed for computational designs).

Here's a first effort in that direction. The inputs are Herschels instead of gliders, but we can just add on the improved G->H converter from #2 above, and we'll be all set...!