Dependent Reflector Terminology and Analysis

[johamwit]

We could say that signals proper have the characteristic that in any circuit - including loops - using them, they can be either present or absent. That's what makes signals capable of "signaling" something, by its presence/absence, 1/0. In other words, signals are not compulsory for the guiding structure to maintain its integrity ( = not evolving into chaos ).

And any of the 2^N states of a glider loop will repeat itself given enough time, as golly's oscar script will happily prove. No dishes broken

Keeping in mind that, paraphrasing Alice, “When we use a word, it means just what we choose it to mean—neither more, nor less.”

[confocaloid]

I'm sorry. As far as I can understand, some of my posts caused attempts to give a detailed formal definition that would enable "tracking" moving active objects beyond the point where they could be tracked visually in the obvious intuitive way. I think it is possible to define dependent reflectors and explain how they work, without any references to the concept of signals, mentioning only streams of spaceships and not trying to interpret or count or track them.

So for the purpose of defining/explaining "dependent reflector", I think such formal definitions are unnecessary.

I think the question "when it is possible to meaningfully track a moving reaction?" is interesting in itself, but it is a separate question.

The discussion has also extended to the question of whether it's a good idea to edit out any mention of both "signals" and "loops" from the new "dependent reflector" article.
...
Trying to "edit away" the signals-and-loops way of describing these patterns, on the grounds that there "aren't really any signals there", still seems like a misguided idea. You shouldn't have to waste time inspecting some larger context to decide whether you can call a glider a "signal" or not. I've already given my view on why the loops are perfectly valid causal loops, so I won't repeat that here -- but if they are loops, then presumably something is traveling in those loops, and "signals" seems like a good general name for those traveling things.
...

It is unnecessary and confusing to discuss the concept of signals when defining and explaining dependent reflectors. A dependent reflector is a reflector that depends upon a regular and continuous input stream. The definition of Reflector already says that a reflector can reflect some specific type of spaceship.

Signals can exist in many different forms (as spaceships, as "holes" in streams of spaceships, as Herschels, as perturbations, as temporary stationary objects, ...) However, the terms "dependent reflector" and "reflector" are used only for streams of spaceships.

A spaceship does not have to be interpreted as a signal. Even if a spaceship doesn't carry any interesting or useful information, it is still possible to reflect it with a reflector. An uninterrupted stream can be reflected by a dependent reflector that accepts a stream of this type, regardless of the question of whether or not there are any actual signals transmitted by this stream.

In a discussion of signal streams, it is expected that there are different possibilities. But a dependent reflector in general requires the input stream to be uninterrupted, so there is only one "valid" input in this case.

For these reasons I suggest to reword the article "Dependent reflector" so that it doesn't mention signals at all. The discussion of signals is unnecessary in this case, and causes confusion.

I suggest to reword the following part so that it doesn't mention cycle length. (The cycle length is not well-defined with dependent reflectors. Trying to give a formal replacement definition in that article would probably mean failing to answer "what is a dependent reflector?" on the basic level.)

For 90° color-preserving dependent reflectors with periods 43 and above, a Snark can sometimes be used to replace a reflector, but the length of the resulting signal loop will very likely change and may have to be adjusted back to a cycle length that is a multiple of the dependent reflector's period.

I think the following part is unclear -- I do not understand which oscillators are excluded.

This list excludes engines that emit gliders but also require gliders to sustain themselves; oscillators like this are known for periods 53 and 69.

[hotdogPi]

This is the p53 I was referring to:

Code: Select all

x = 101, y = 114, rule = B3/S23
58b2o$57bobo$51b2o4bo$49bo2bo2b2ob4o$49b2obobobobo2bo$52bobobobo$52bob
ob2o$53bo2$66b2o$57b2o7bo$57b2o5bobo$64b2o2$51bo$49b2o$50b2o3$54b2o6b
2o$55bo6bobo10b2o$52b3o7bo12bo$52bo24bo$57b2o14b5o$58bo13bo$58bobo12b
3o$59b2o15bo$37bo35b4o$36bo31b2o3bo3b2o$36b3o29b2o4b3o2bo$76bob2o$60bo
15bo$60b2o13b2o$59bobo$79b2o$67b2o9bobo$67bo4b2o4bo$68b3o2bo2b2ob4o$
70b2obobobobo2bo$73bobobobo$73bobob2o$9bo13bobo48bo$9b3o11b2o$12bo11bo
15b2o15b2o28b2o$11b2o28bo15bo29bo$41bobo5b3o3bobo27bobo$42b2o4bo3bo2b
2o21bo6b2o$3b2o8b3o31bo5bo22b2ob2o$3bo8bo3bo30bo5bo22b2ob2o$2obo7bo5bo
29bo5bo22bo2b2o$o2b3o4bo3bo3bo29bo3bo24b2o$b2o3bo3bo2bobo33b3o$3b4o3bo
3bo$3bo7bo8bo54b2o$4b3o5bo3bob2obo54bo19b2o$7bo5b3o5bo14b3o34b3o20bo$
2b5o14b2o15bo24bo9bo24bo$2bo24bo9bo24bo15b2o14b5o$4bo20b3o34b3o14bo5b
3o5bo$3b2o19bo54bob2obo3bo5b3o$24b2o54bo8bo7bo$86bo3bo3b4o$49b3o33bobo
2bo3bo3b2o$22b2o24bo3bo29bo3bo3bo4b3o2bo$20b2o2bo22bo5bo29bo5bo7bob2o$
20b2ob2o22bo5bo30bo3bo8bo$20b2ob2o22bo5bo31b3o8b2o$14b2o6bo21b2o2bo3bo
4b2o$13bobo27bobo3b3o5bobo$13bo29bo15bo28b2o$12b2o28b2o15b2o15bo11bo$
76b2o11b3o$26bo48bobo13bo$22b2obobo$21bobobobo$18bo2bobobobob2o$18b4ob
2o2bo2b3o$22bo4b2o4bo$20bobo9b2o$20b2o$39bobo$24b2o13b2o$24bo15bo$21b
2obo$21bo2b3o4b2o29b3o$22b2o3bo3b2o31bo$24b4o35bo$24bo15b2o$25b3o12bob
o$28bo13bo$23b5o14b2o$23bo24bo$25bo12bo7b3o$24b2o10bobo6bo$37b2o6b2o3$
49b2o$50b2o$49bo2$35b2o$34bobo5b2o$34bo7b2o$33b2o2$47bo$43b2obobo$42bo
bobobo$39bo2bobobobob2o$39b4ob2o2bo2bo$43bo4b2o$41bobo$41b2o!

And the p69:

Code: Select all

x = 99, y = 128, rule = B3/S23
15bo67bo$15b3o63b3o$18bo7b3o41b3o7bo$17b2o11bo37bo11b2o$27bo2bobo33bob
o2bo$22bo53bo$21bobo4b4obo31bob4o4bobo$4b2o3bo11bo2bo4b3o2bo29bo2b3o4b
o2bo11bo3b2o$b2o2bo2bobo11b2o8bo33bo8b2o11bobo2bo2b2o$2bobo27bo2bo27bo
2bo27bobo$2bob2o2bo26bo27bo26bo2b2obo$2o6b2o24bo29bo24b2o6b2o$bobob2ob
2o25b3o23b3o25b2ob2obobo$bobo4b2o27bo23bo27b2o4bobo$2bobo3bo7b2o63b2o
7bo3bobo$4bobo10b2o61b2o10bobo$3bobo10bo65bo10bobo$3b2o89b2o$9bo21bo
14b2o3b2o14bo21bo$8bobo21b2o12bo5bo12b2o21bobo$7bo2bo20b2o11bobo5bobo
11b2o20bo2bo$8b2o34b2o7b2o34b2o3$4b2o34b2o15b2o34b2o$3bobo33bo2bo13bo
2bo33bobo$3bo35bobo15bobo35bo$2b2o36bo17bo36b2o$45b2o5b2o$21b2o21bobo
5bobo21b2o$20b2o23bo7bo23b2o$22bo17b3ob2obo3bob2ob3o17bo$44bobobobobob
o$29bo9bo5bo2bobo2bo5bo9bo$27b2o14bo2bobobobo2bo14b2o$28b2o10b2o2b2obo
bobob2o2b2o10b2o$40bo4bobo3bobo4bo$9b2o29bo4bobo3bobo4bo29b2o$10bo29bo
5bo5bo5bo29bo$10bobo73bobo$11b2o73b2o3$15b2o65b2o$14bo2bo63bo2bo$15bob
o63bobo$16bo21b2o19b2o21bo$38bobo17bobo$11bo11bo14bo21bo14bo11bo$11bob
o10b2o47b2o10bobo$11bo3bo7b2o49b2o7bo3bo$15b2o65b2o$12b2ob2o25bo13bo
25b2ob2o$15b2o24b2o13b2o24b2o$13bo3b2o21bo2bo11bo2bo21b2o3bo$15bo24bob
o13bobo24bo$16bo3bo20b2o13b2o20bo3bo$18b2obo55bob2o$21bo55bo$21b2o53b
2o9$21b2o53b2o$21bo55bo$18b2obo55bob2o$16bo3bo20b2o13b2o20bo3bo$15bo
24bobo13bobo24bo$13bo3b2o21bo2bo11bo2bo21b2o3bo$15b2o24b2o13b2o24b2o$
12b2ob2o25bo13bo25b2ob2o$15b2o65b2o$11bo3bo7b2o49b2o7bo3bo$11bobo10b2o
47b2o10bobo$11bo11bo14bo21bo14bo11bo$38bobo17bobo$16bo21b2o19b2o21bo$
15bobo63bobo$14bo2bo63bo2bo$15b2o65b2o3$11b2o73b2o$10bobo73bobo$10bo
29bo5bo5bo5bo29bo$9b2o29bo4bobo3bobo4bo29b2o$40bo4bobo3bobo4bo$28b2o
10b2o2b2obobobob2o2b2o10b2o$27b2o14bo2bobobobo2bo14b2o$29bo9bo5bo2bobo
2bo5bo9bo$44bobobobobobo$22bo17b3ob2obo3bob2ob3o17bo$20b2o23bo7bo23b2o
$21b2o21bobo5bobo21b2o$45b2o5b2o$2b2o36bo17bo36b2o$3bo35bobo15bobo35bo
$3bobo33bo2bo13bo2bo33bobo$4b2o34b2o15b2o34b2o3$8b2o34b2o7b2o34b2o$7bo
2bo20b2o11bobo5bobo11b2o20bo2bo$8bobo21b2o12bo5bo12b2o21bobo$9bo21bo
14b2o3b2o14bo21bo$3b2o89b2o$3bobo10bo65bo10bobo$4bobo10b2o61b2o10bobo$
2bobo3bo7b2o63b2o7bo3bobo$bobo4b2o27bo23bo27b2o4bobo$bobob2ob2o25b3o
23b3o25b2ob2obobo$2o6b2o24bo29bo24b2o6b2o$2bob2o2bo26bo27bo26bo2b2obo$
2bobo27bo2bo27bo2bo27bobo$b2o2bo2bobo11b2o8bo33bo8b2o11bobo2bo2b2o$4b
2o3bo11bo2bo4b3o2bo29bo2b3o4bo2bo11bo3b2o$21bobo4b4obo31bob4o4bobo$22b
o53bo$27bo2bobo33bobo2bo$17b2o11bo37bo11b2o$18bo7b3o41b3o7bo$15b3o63b
3o$15bo67bo!

[dvgrn]

We could say that signals proper have the characteristic that in any circuit - including loops - using them, they can be either present or absent. That's what makes signals capable of "signaling" something, by its presence/absence, 1/0. In other words, signals are not compulsory for the guiding structure to maintain its integrity ( = not evolving into chaos ).
...
Keeping in mind that, paraphrasing Alice, “When we use a word, it means just what we choose it to mean—neither more, nor less.” :D

The problem is that while we could say that, people have now spent over half a century of Conway's Life discussion very carefully not saying that.

A few posts up I tried to outline the key idea that binary 0=absent, 1=present signals are not not the only kinds of signals that can move information through the Life universe. A "signal" is not the same thing as a "bit". The signals in dependent reflector loops are an example of a type of signal that moves information, but can't be added or removed in this way.

The attempt to rewrite the dependent reflector article to remove all reference to "signal" seemed to be motivated by this underlying confusion. I.e., there seems to be this implicit belief that "signal" does mean the same thing as "bit", or that it should from now on only be used to mean the same thing as "bit".

That doesn't seem likely to happen -- too many people have been using "signal" to mean "the movement of information through the Life universe" for too many decades. This subtle distinction about the meaning of "signal" may well be confusing to some people, but we need to address that directly in the definition of "signal", not just bury it by conflating "signal" with "bit" and pretending that all Life signals are simple binary presence-or-absence choices. That would be a huge and sudden change in the usage of the term "signal", losing an important nuance, and ultimately causing a much more dangerous level of confusion.

Clearly I need some good examples
I've mentioned the QuickSilver Demonoid's use of a continuous stream of signals to convey the information "don't self-destruct yet!" through a dependent reflector.

Here's maybe a simpler example of the inadequacy of the "signals must be able to be either present or absent" claim. Have a look at the weekender|doo-dah converter.

The choice of weekender or doo-dah is obviously a binary choice. In the pattern in the link, that binary choice is being communicated by the small glider loop on the left side, to the larger structure on the right.

If signals were always presence-or-absence choices, then there would be some object that could be added or removed to change a weekender to a doo-dah, or vice versa. But there isn't any such object in this case. The information is conveyed by a "carrier signal", so to speak -- there's a glider that always has to be present, but can be in one of two different locations.

-- Yes, that's still a binary choice between two locations ... but it doesn't allow for simple presence-or-absence tests of the type that people are trying to use here to detect the existence of "actual signals". Just try removing any one glider and see what happens... and yet it's perfectly reasonable to use "signal" in this context, because clearly the signal is getting through to the converter.

[Scorbie]

Ever since from the first post what I tried to explain is why dependent reflector loop-or-whatever looks like a loop.

I think it is possible to define dependent reflectors and explain how they work, without any references to the concept of signals, mentioning only streams of spaceships and not trying to interpret or count or track them.

What I did is precisely that. That definition described dependent reflectors by explaining how they work, without any references to the concept of signals, mentioning only streams of spaceships and not trying to interpret or count or track(*) them.
(*) I used the word "track" in the explanation but it's generalizable beyond reflectors that allow following the glider by eye and tracking them.

So for the purpose of defining/explaining "dependent reflector", I think such formal definitions are unnecessary.

I disagree. I had a hard time understanding your ideas behind your posts just from reading it. (Probably because I feel pretty tired today, or maybe because one needs to know what you are thinking beforehand in order to understand what you wrote.)
Because I tried to define dependent reflectors as rigorously as I can, I could understand what you were thinking between the lines etc.

I'll try to explain the intutive idea.

A dependent reflector can be thought of as a device that redirects an infinite glider generator generating a glider at A to an infinite glider generator generating a glider at B after delay generations.

Code: Select all

x = 43, y = 88, rule = LifeHistory
3D5.3D7.25D$D7.D.D6.D12.A11.D$D.D5.D.D5.D12.A.A10.D$D.D5.D.D4.D13.A.
3A8.D$3D5.3D3.D11.2A.A4.A7.D$13.D12.A.A.ABABA7.D$12.D15.A.A.2A4.2A2.D
$11.D2.2A11.ABA4B4.A3.D$10.D4.A9.3BA7BA.A3.D$10.D4.A.A6.12B2A4.D$10.D
5.2A5.13B6.D$10.D9.16B2.2D2.D$10.D8.17B2.D.D.D$10.D7.2AB3A12BD.3D.D$
10.D6.A2BA12B.DBDBD.D.D$10.D3.2A.3BAB2A8B3.2DB2D2.D$10.D2.A3B2AB2A4B
2A2B6.4B2.D$10.D2.A2B.4A5B2A2B7.4BD$10.D3.A3.2BA7B10.2BDB$10.D8.3A8B
9.D3B$10.D6.7B4.2A8.D.3B$10.D5.3AB.2B.B3.A8.D3.2B$11.D3.3BA3.B2AB3.3A
4.D5.B$12.D.3BA5.2A6.A3.D$13.D3B17.D$12.2BDB17.D$11.4B18D$10.4B$3D6.
4B$D.D5.4B$3D4.4B$D.D3.4B$D.D2.4B$4.4B$3.4B$2.E3B$.B2EB$BEBE13$3D.3D.
3D$D3.D3.D.D19.A$D.D.3D.3D18.A.A$D.D3.D3.D18.A.3A$3D.3D.3D15.2A.A4.A$
26.A.A.ABABA$28.A.A.2A4.2A$14.2A11.ABA4B4.A$15.A9.3BA7BA.A$15.A.AB5.
12B2A$16.2AB.B2.13B$18.18B$18.18B$18.2AB3A12BE$17.A2BA12B.EBEB$14.2A
4BAB2A8B3.2E2B$13.A3B2AB2A4B2A2B6.4B$13.A3B4A5B2A2B7.4B$14.A3.2BA7B
10.4B$19.3A8B9.4B$17.7B4.2A10.3B$16.3AB.2B.B3.A12.2B$15.3BA3.B2AB3.3A
10.B$14.3BA5.2A6.A$13.4B$12.4B$11.4B$10.4B$9.4B$8.4B$7.4B$6.4B$5.4B$
4.4B$3.4B$2.C3B$.B2CB$BCBC!

For instance, with this definition, the pattern inside the red polygon is a p59 dependent reflector. It reflects a NEwards glider stream with the firstmost glider at A to a SEwards glider stream with the firstmost glider at B after 59 generations.
You can "bind" the glider at A at gen 0 to the glider at B at gen 59.
The bound glider pairs may change if we enclose a different region to serve as a dependent reflector, but it's okay to me as long as the explanation is consistent. If you think of the reflectors as a black box then the two different enclosings are different reflectors.

Edit: I was not satisfied because I cannot define the bindings at gens 0..58. If it bothers you then just let the reflector already have an output glider at gen 0 (let delay = 0) and bind A and B.

[Scorbie]

My 2 cents about the signal discussion.

I think the reason why someone could be against the word "signal" is because in some fields the word "signal" is precisely used for something that can convey information and thus have nonzero entropy and have more than one possible states. For example, with this context, an infinite period p glider stream starting at glider g, as a whole, for example, have only one possible pattern, thus cannot be considered as a signal. I'll underline the term used in this context.

If you look at the Dependent reflector entry you see that it receives a "continuous stream of input signal"s. It seems colloquially that the word "signal" in this definition means something like "object that can be used as a signal (outside of the dependent reflector use case)". (Edit: brain dump: Personally this feels analogous to sending all 1s through a wire.)

If one's not satisfied with the use of this term then we could use a term like "signal object" for disambiguation, for example. (This is just an example. I don't have any preferences to how this is resolved. Just let me know how it's settled. And off course there are signals that can be used as signals that work on non-vacuum backgrounds, so I'm not sure if object is the best word to call them.)

Edit:

Signals can exist in many different forms (as spaceships, as "holes" in streams of spaceships, as Herschels, as perturbations, as temporary stationary objects, ...) However, the terms "dependent reflector" and "reflector" are used only for streams of spaceships.

If someone finds a 2c/3 signal turner would it be called a reflector? I assume people would think of it as a reflector, but I'll have to see.

[dvgrn]

My 2 cents about the signal discussion.

I think the reason why someone could be against the word "signal" is because in some fields the word "signal" is precisely used for something that can convey information and thus have nonzero entropy and have more than one possible states. For example, with this context, an infinite period p glider stream starting at glider g, as a whole, for example, have only one possible pattern, thus cannot be considered as a signal.

Sure -- nobody's going to draw a box around a half-infinite glider stream and call the entire thing a "signal".

Individual gliders in such a stream are clearly signals, though, in the sense defined by signal: they carry information from one part of the Life universe to another. Even if that information is just basically, "Glider coming through here!", that's still perfectly valid information.

If you look at the Dependent reflector entry you see that it receives a "continuous stream of input signals"...

It's probably best to bear in mind that the dependent reflector article doesn't make a particularly authoritative reference. It was all written very recently, and quite a bit of it was written by me, with no attempt to avoid the use of the words "signal" or "loop". We'll have to look back a lot farther to validate the standard usages of "loop" and "signal" that are under discussion here.

My general sense is that "signal" has never been synonymous with "bit", or with "spaceship" -- it has always had the broader meaning of "traveling information", nothing more specific than that. Rightly or wrongly, I saw the recent edit to the dependent-reflector-loop article as an attempt to rather heavy-handedly implement a particular suggested change in this standard terminology, ignoring several decades of previous common usage that conflicts with this new interpretation.

Maybe the standard terminology _is_ confusing, but it seems to me that the source of the confusion is somewhere other than the dependent-reflector article.

Other Appeals to Old Standard Usage
... I suppose that my opinions about other naming proposals that have come up, are somewhat related to this viewpoint: e.g., "shuttles" have always been some single item oscillating back and forth -- consistent for forty years, until in just a few scattered cases it started getting misapplied, about a decade ago. "Relays" have pretty much always been a single spaceship bouncing back and forth, but not necessarily in an oscillator -- there are such things as "growing relays", and "fluff relays" were a proposed hypothetical type of spaceship where a glider went back and forth between two halves, moving each half in the same direction.

The case for "loop" applied to dependent reflectors is less clear, I suppose. It seems totally fine to me to apply the concept of "loop" to a closed conduit where discrete pieces of information -- signals -- are flowing around and back to the starting point, but the information can't be conveniently divided up into simple binary bits. But I can't prove offhand that that's what "everyone else" would have thought, a decade or two ago when dependent reflectors were somewhat newer and stranger.

[Scorbie]

Just to let you all know I'm trying to suggest possible solutions for resolving the conflict.

Individual gliders in such a stream are clearly signals, though, in the sense defined by signal: they carry information from one part of the Life universe to another. Even if that information is just basically, "Glider coming through here!", that's still perfectly valid information.

I want to point out that as far as I can infer, confocaloid is using a specific term for "information", specifically what's used in "information theory". This definition has a unit in bits and requires the signal to have more than one possible state or otherwise it contains zero information. Which explains why confocaloid wants to disambiguate gliders that can be missing (more than 1 possible configurations) to gliders that cannot be missing (exactly 1 possible configuration)

It's probably best to bear in mind that the dependent reflector article doesn't make a particularly authoritative reference. It was all written very recently, and quite a bit of it was written by me, with no attempt to avoid the use of the words "signal" or "loop". We'll have to look back a lot farther to validate the standard usages of "loop" and "signal" that are under discussion here.

My general sense is that "signal" has never been synonymous with "bit", or with "spaceship" -- it has always had the broader meaning of "traveling information", nothing more specific than that. Rightly or wrongly, I saw the recent edit to the dependent-reflector-loop article as an attempt to rather heavy-handedly implement a particular suggested change in this standard terminology, ignoring several decades of previous common usage that conflicts with this new interpretation.

Maybe the standard terminology _is_ confusing, but it seems to me that the source of the confusion is somewhere other than the dependent-reflector article.

Okay if this is longstanding historically used standard terminology I'm personally towards keeping backwards compatibility for the terms.

According to wikipedia, A signal may also be defined as any observable change in a quantity over space or time (a time series), even if it does not carry information. I'm assuming information here is the term in information theory. Therefore I think it's acceptable that the terms don't strictly align with the terms in information theory.

Therefore a possible solution would be keeping the term but explaining the differences of the terms "signal" and "information" from information theory.

[confocaloid]

Individual gliders in such a stream are clearly signals, though, in the sense defined by signal: they carry information from one part of the Life universe to another. Even if that information is just basically, "Glider coming through here!", that's still perfectly valid information.

I disagree. When a glider is not actually used to carry an information, it does not make sense to describe it as a signal.

I think it is extremely confusing to describe all gliders in the pattern under discussion as "signals", on the grounds that their presence in a specific location at a specific time technically counts as "information". When someone is interested in describing some pattern of interest, I think it makes sense to restrict use of the word "signal" to those cases where there is some actual interesting information.

When I'm reading the page Glider in LifeWiki, I'm reading an article about a spaceship -- not an article about a signal.
As a pattern, the glider is a spaceship, not a signal.

Gliders can be used to transmit information in the Life universe. Gliders are very commonly used in this way.
However, gliders are not always used in this way (and gliders are not always used in some way).

In a discussion of a pattern where a tandem glider is used to convey some information, it can be convenient/useful to describe that tandem glider as a signal. There are two gliders, but there is only one signal.

The R-pentomino releases 6 gliders during its evolution. Neither of these gliders is a signal (because this sentence belongs to a discussion of the R-pentomino pattern, and here neither of six gliders is actually used to convey information).

When a newly created glider is eaten by an eater without any intermediate interactions, of course that particular glider is not used to transmit any information. An evolving Herschel releases the first natural glider (FNG). It is common that the FNG is suppressed. When the FNG is suppressed, it does not make sense to describe that FNG as a signal. A suppressed first natural glider is not a signal.

An infinite stream of gliders from a glider gun is not a signal. Neither is any of gliders in that stream, unless the surrounding context explains how those gliders are actually used to convey some interesting information.

---

As far as I understand, the word "signal" means "something that is used to convey information". (Not just "can be used", but is actually used.) If something is not used to convey information then it is not a signal. There should be a simpler term (active object, reaction, perturbation, spaceship, glider, recipe, ...) that accurately describes that something.

[confocaloid]

This is likely offtopic (and certainly not CA-related), but this is an example of use of the word "signal".

The term nonmanual signal is a gesture or action such as change of your facial expression; the tilt, shake, or nod of your head; and/or the hunching of one or both shoulders -- that is used convey information.

If such movements are not used to convey information then they are not "signals" -- instead we would simply call them twitches.

[Scorbie]

As far as I understand, the word "signal" means "something that is used to convey information". (Not just "can be used", but is actually used.) If something is not used to convey information then it is not a signal. There should be a simpler term (active object, reaction, perturbation, spaceship, glider, recipe, ...) that accurately describes that something.

Considering there are uses of the term "signal" for things that do not carry information in the strict sense[1], and considering the word "signal" is prevalently used with its current meaning already, what do you think about keeping the term overloaded as it is and explaining how it's different from the definitions from information theory? And possibly define a new term for your definition?

[1] https://en.wikipedia.org/wiki/Carrier_signal

Edit: Oh, there could be uses of dependent reflectors that do reflect signals that carry information in the strict sense (and thus have nonzero entropy) as a one-time-turner of a glider stream of unary-encoded data. I didn't know people did this. I think that justfies more of the use of the current definition as the current definition is consistent with the information theory terms.

[confocaloid]

I think the current common usage is that the word "signal" is normally used when referring to something that actually carries some interesting information. In other cases, other words are used instead, such as "glider", "spaceship", "moving object", "perturbation".

E.g. a simple presence of a glider somewhere in the Life universe is not normally described as a "signal". Occurrences of gliders from random soups are not signals, and normally are not described as signals. And so on.

I think the current version of Dependent reflector uses the word "signal" in an uncommon way that does not reflect common usage. So it needs to be rewritten in some way. I suggest to change the page "Dependent reflector" so that it doesn't mention signals at all - the discussion of signals is unnecessary in this case, and causes confusion.

Edit: Oh, there could be uses of dependent reflectors that do reflect signals that carry information in the strict sense (and thus have nonzero entropy) as a one-time-turner of a glider stream of unary-encoded data. I didn't know this. I think that justfies more of the use of the current definition as the current definition is consistent with the information theory terms.

In some cases, a dependent reflector effectively "becomes" a non-dependent reflector (as long as the pattern functions as intended).

I think the definition of a dependent reflector should not refer to the idea of signals at all. Both dependent and non-dependent reflectors reflect streams of spaceships. Whether or not these reflected spaceships (or subsets of reflected spaceships) can be usefully viewed/described/discussed as things carrying some interesting signals, the definitions of a reflector/dependent reflector should remain the same.

If someone finds a 2c/3 signal turner would it be called a reflector? I assume people would think of it as a reflector, but I'll have to see.

I think it could be called an elbow. I don't know how it will be actually named when/if it is discovered.

[Scorbie]

I think the current common usage is that the word "signal" is normally used when referring to something that actually carries some interesting information.

What is your definition of "interesting imformation?" The same data 33 can be represented as a binary stream 100001 or a unary stream of 33 consecutive gliders. I don't see why one would be more interesting than the other.

[dvgrn]

I think it is extremely confusing to describe all gliders in the pattern under discussion as "signals", on the grounds that their presence in a specific location at a specific time technically counts as "information". When someone is interested in describing some pattern of interest, I think it makes sense to restrict use of the word "signal" to those cases where there is some actual interesting information.

This paragraph might summarize a key point of disagreement. Here are three observations in response:

1) Nobody is in fact going to describe every glider as a "signal" in every context. Quite often the more convenient word is "glider". It seems okay to me for people to say the word "glider" when they mean "glider", and the word "signal" when they mean "signal". It actually doesn't seem okay to me to attempt to restrict the use of "signal" in the way you describe.

2) At the moment it looks to me as if there's some information flowing in dependent reflectors that is interesting to me, but not interesting to you -- which may be what has caused this very long and confusing discussion in which we mostly just talk past each other rather than understanding each other's point of view.

I would prefer in the future to avoid the confusion that has been on display in this discussion -- in good part by not attempting to restrict the use of the words "signal" and "loop" in the way that you're suggesting.

3) If you think it makes sense to restrict the use of the words "signal" and "loop" in some way, then by all means be careful not to use those words in ways that you personally object to.

However, speaking for myself, I'm finding that I very strongly object to being told that I'm using the word "signal" wrong -- and that this wrongness requires that the dependent-reflector article be rewritten to arbitrarily remove various phrases, which I believe are standard and generally well-understood uses of standard terminology.

By this time I've done quite a bit of review of past and current usage, and I'm not convinced yet that I'm doing anything out of line with the common CA meaning of the terms "signal" and "loop".

I don't claim that I have yet proven that point to the satisfaction of every possible reader -- but then I haven't heard any support from anyone else for this idea of restricting the meaning of "signal" and "loop", so I'm not convinced that the burden of proof lies with me here.

It almost seems as if you believe that the word "loop" should only be used synonymously with "memory loop", and that -- by logical extension -- "dependent reflectors" should really be called something other than "reflector" because they don't reflect "signals" in your narrow definition. EDIT: Sorry, this has been mentioned in passing before -- I mean, they don't reflect spaceships, because that would require input spaceships and output spaceships, and you seem to be saying that that's not how dependent reflectors really work.

All of this is a large change in usage to attain this new and different type of consistency that you seem to be aiming for. It will require extraordinary proof that such a large change is really necessary.

As far as I understand, the word "signal" means "something that is used to convey information". (Not just "can be used", but is actually used.) If something is not used to convey information then it is not a signal. There should be a simpler term (active object, reaction, perturbation, spaceship, glider, recipe, ...) that accurately describes that something.

Can you please re-consider and respond to Scorbie's suggestion above?

what do you think about keeping the term overloaded as it is and explaining how it's different from the definitions from information theory? And possibly define a new term for your definition?

It's seems quite possible that the existing use of "signal" is better left the way it is, and a new term can be invented with the exact connotation that you prefer. If that term catches on over time and is used by a lot of people, then that will be evidence that you've discovered a useful distinction in types of information flow, and after a year or two it could be added to the LifeWiki.

[GUYTU6J]

By-stander's view:
When mentioning "signal" or "information", there are two concepts that may be involved: 1)microscopic ON/OFF state of one cell, or 2)mesoscopic presence/absence of an object. (Mesoscopic, not macroscopic, which is reserved for larger constructions. As I have stated in Tutorials/General_technology, "signal circuitry [is] the very connection between microscopic evolution and macroscopic construction.") The first appears in the discussion of speed of light and zone of influence, and the second is in Herschel conduit and Glider-accepting devices.

This is not a new idea, because it has been conveyed to different extent in previous posts like:

Golly is cell level editor. In GOL many times we have "logical components" (like reflectos, duplicators, blockic seeds etc.). So we need more tools to allow "Smart Editing". By Smart Editing I mean tools that allow move reflectors, duplicators, place reflectors from gliders, and even manage Hershels from inside golly using input-output interfaces style.
...

There are two ways to think about ConwayLife: cell-based and pattern-based. The cell-based approach focuses on the individual cells and whether they're on or off. The individual cells do not move. They merely turn on or off. The pattern-based approach focuses on entire patterns. Unlike the cell-based approach, the pattern-based approach allows objects, such as gliders, to move.
...

Occasionally the two concepts are mixed, see for a subtle example:

Let's make things confusing, shall we?

Code: Select all

x = 73, y = 26, rule = LifeHistory
11.2A54.2A$12.A55.A$10.A55.A$10.5A51.5A$14.A55.A$10.2A54.2A$9.A.A53.A
.A$9.2A54.2A$13.2A54.2A$13.A.A53.A.A$.2D12.A39.2D14.A$2D13.2A37.2D15.
2A$.2D52.2D$2.D53.D8$.3C53.3C7.2A$2.C8.2A45.C8.A.A$2.3C6.A46.3C8.A$
12.3A54.2A$14.A!

Let's make things confusing, shall we?

Code: Select all

[HR44B, but with block-eating eater shifted into Fx77 position,
which magically moves the output by two cells and changes it to HR48B]

Nice find! It's not really too confusing in practice, just a new HR48B -- but it's a wonderful minimalist example of how that subtle change in eater placement makes a large difference several ticks later and many cells away. The information about which eater is present propagates through the active reaction faster than c/2 diagonally (!).
...

In my opinion, it is reasonable to say that a dependent glider reflector transmits microscopic information but not mesoscopic information. One can toggle the state of one cell at the input end and observe variation(s) of the state of cell(s) at the output end, both being describable using "ON/OFF". However, after changing a glider from presence to absence, one cannot describe the output glider by still using "presence/absence" — since the dependent reflector is broken and does not work any more, the output glider should actually be "null".

EDIT: sorry, but I kept losing connection just now.

[hotdogPi]

The new p34 found today was explicitly called a "dependent reflector loop" by its finder, who hasn't posted in this thread.

[confocaloid]

The new p34 found today was explicitly called a "dependent reflector loop" by its finder, who hasn't posted in this thread.

Ever since from the first post what I tried to explain is why dependent reflector loop-or-whatever looks like a loop.

To me, 'dependent reflector loop' accurately describes this particular type of oscillator. The word 'signal' doesn't feature in 'dependent reflector loop', so even though I agree with confocaloid's point that these* dependent reflectors cannot reusably transmit information, I don't see why that should disqualify these oscillators from being called 'loops'.

*the overclocked Silver reflector, on the other hand, can transmit information, because there's freedom in the relative timings of gliders.

Sorry for not replying earlier. These points are convincing. The phrase "dependent reflector loop" is now in common use outside of this thread, and it is used to refer to this type of oscillator. I have no objections to documenting this term on LifeWiki as a commonly used term.

[confocaloid]

I think the current common usage is that the word "signal" is normally used when referring to something that actually carries some interesting information.

What is your definition of "interesting imformation?" The same data 33 can be represented as a binary stream 100001 or a unary stream of 33 consecutive gliders. I don't see why one would be more interesting than the other.

I think this depends on the specific situation. When there is a specific engineered pattern, and someone would like to describe or discuss this pattern, there may be some "interesting" information flow that is best described in terms of signals, and some "boring" information flow that is best described in terms of (e.g.) spaceships.

E.g. suppose the pattern is a large Herschel loop and contains circulating signals that are duplicated and sent (e.g. as gliders) for some purpose to some other place. It can be convenient/useful to describe these circulating active reactions as signals -- this is the "interesting information" in this case. At the same time, something like suppressed first natural gliders from those Herschels would be "boring information" -- if those suppressed gliders are mentioned in discussion, probably they would be described as gliders (rather than as signals).

In oscillators that are now called dependent reflector loops, there are glider streams moving between dependent reflectors. I think these glider streams are best described as glider streams that keep the dependent reflector reactions from self-destructing and allow them to work correctly at their period. So these glider streams are a way to support the interesting reactions that are demonstrated in the wiki page. I believe the interesting thing in this case is the dependent reflector itself -- the glider streams are "boring", and I suggest to describe them as (uninterrupted) glider streams.

The same happens in the p53 and p69 oscillators that are posted by hotdogPi -- there is a central active reaction that emits glider streams, and those glider streams are redirected to allow the reaction to work correctly:

This is the p53 I was referring to:

Code: Select all

x = 101, y = 114, rule = B3/S23
58b2o$57bobo$51b2o4bo$49bo2bo2b2ob4o$49b2obobobobo2bo$52bobobobo$52bob
ob2o$53bo2$66b2o$57b2o7bo$57b2o5bobo$64b2o2$51bo$49b2o$50b2o3$54b2o6b
2o$55bo6bobo10b2o$52b3o7bo12bo$52bo24bo$57b2o14b5o$58bo13bo$58bobo12b
3o$59b2o15bo$37bo35b4o$36bo31b2o3bo3b2o$36b3o29b2o4b3o2bo$76bob2o$60bo
15bo$60b2o13b2o$59bobo$79b2o$67b2o9bobo$67bo4b2o4bo$68b3o2bo2b2ob4o$
70b2obobobobo2bo$73bobobobo$73bobob2o$9bo13bobo48bo$9b3o11b2o$12bo11bo
15b2o15b2o28b2o$11b2o28bo15bo29bo$41bobo5b3o3bobo27bobo$42b2o4bo3bo2b
2o21bo6b2o$3b2o8b3o31bo5bo22b2ob2o$3bo8bo3bo30bo5bo22b2ob2o$2obo7bo5bo
29bo5bo22bo2b2o$o2b3o4bo3bo3bo29bo3bo24b2o$b2o3bo3bo2bobo33b3o$3b4o3bo
3bo$3bo7bo8bo54b2o$4b3o5bo3bob2obo54bo19b2o$7bo5b3o5bo14b3o34b3o20bo$
2b5o14b2o15bo24bo9bo24bo$2bo24bo9bo24bo15b2o14b5o$4bo20b3o34b3o14bo5b
3o5bo$3b2o19bo54bob2obo3bo5b3o$24b2o54bo8bo7bo$86bo3bo3b4o$49b3o33bobo
2bo3bo3b2o$22b2o24bo3bo29bo3bo3bo4b3o2bo$20b2o2bo22bo5bo29bo5bo7bob2o$
20b2ob2o22bo5bo30bo3bo8bo$20b2ob2o22bo5bo31b3o8b2o$14b2o6bo21b2o2bo3bo
4b2o$13bobo27bobo3b3o5bobo$13bo29bo15bo28b2o$12b2o28b2o15b2o15bo11bo$
76b2o11b3o$26bo48bobo13bo$22b2obobo$21bobobobo$18bo2bobobobob2o$18b4ob
2o2bo2b3o$22bo4b2o4bo$20bobo9b2o$20b2o$39bobo$24b2o13b2o$24bo15bo$21b
2obo$21bo2b3o4b2o29b3o$22b2o3bo3b2o31bo$24b4o35bo$24bo15b2o$25b3o12bob
o$28bo13bo$23b5o14b2o$23bo24bo$25bo12bo7b3o$24b2o10bobo6bo$37b2o6b2o3$
49b2o$50b2o$49bo2$35b2o$34bobo5b2o$34bo7b2o$33b2o2$47bo$43b2obobo$42bo
bobobo$39bo2bobobobob2o$39b4ob2o2bo2bo$43bo4b2o$41bobo$41b2o!

And the p69:

Code: Select all

x = 99, y = 128, rule = B3/S23
15bo67bo$15b3o63b3o$18bo7b3o41b3o7bo$17b2o11bo37bo11b2o$27bo2bobo33bob
o2bo$22bo53bo$21bobo4b4obo31bob4o4bobo$4b2o3bo11bo2bo4b3o2bo29bo2b3o4b
o2bo11bo3b2o$b2o2bo2bobo11b2o8bo33bo8b2o11bobo2bo2b2o$2bobo27bo2bo27bo
2bo27bobo$2bob2o2bo26bo27bo26bo2b2obo$2o6b2o24bo29bo24b2o6b2o$bobob2ob
2o25b3o23b3o25b2ob2obobo$bobo4b2o27bo23bo27b2o4bobo$2bobo3bo7b2o63b2o
7bo3bobo$4bobo10b2o61b2o10bobo$3bobo10bo65bo10bobo$3b2o89b2o$9bo21bo
14b2o3b2o14bo21bo$8bobo21b2o12bo5bo12b2o21bobo$7bo2bo20b2o11bobo5bobo
11b2o20bo2bo$8b2o34b2o7b2o34b2o3$4b2o34b2o15b2o34b2o$3bobo33bo2bo13bo
2bo33bobo$3bo35bobo15bobo35bo$2b2o36bo17bo36b2o$45b2o5b2o$21b2o21bobo
5bobo21b2o$20b2o23bo7bo23b2o$22bo17b3ob2obo3bob2ob3o17bo$44bobobobobob
o$29bo9bo5bo2bobo2bo5bo9bo$27b2o14bo2bobobobo2bo14b2o$28b2o10b2o2b2obo
bobob2o2b2o10b2o$40bo4bobo3bobo4bo$9b2o29bo4bobo3bobo4bo29b2o$10bo29bo
5bo5bo5bo29bo$10bobo73bobo$11b2o73b2o3$15b2o65b2o$14bo2bo63bo2bo$15bob
o63bobo$16bo21b2o19b2o21bo$38bobo17bobo$11bo11bo14bo21bo14bo11bo$11bob
o10b2o47b2o10bobo$11bo3bo7b2o49b2o7bo3bo$15b2o65b2o$12b2ob2o25bo13bo
25b2ob2o$15b2o24b2o13b2o24b2o$13bo3b2o21bo2bo11bo2bo21b2o3bo$15bo24bob
o13bobo24bo$16bo3bo20b2o13b2o20bo3bo$18b2obo55bob2o$21bo55bo$21b2o53b
2o9$21b2o53b2o$21bo55bo$18b2obo55bob2o$16bo3bo20b2o13b2o20bo3bo$15bo
24bobo13bobo24bo$13bo3b2o21bo2bo11bo2bo21b2o3bo$15b2o24b2o13b2o24b2o$
12b2ob2o25bo13bo25b2ob2o$15b2o65b2o$11bo3bo7b2o49b2o7bo3bo$11bobo10b2o
47b2o10bobo$11bo11bo14bo21bo14bo11bo$38bobo17bobo$16bo21b2o19b2o21bo$
15bobo63bobo$14bo2bo63bo2bo$15b2o65b2o3$11b2o73b2o$10bobo73bobo$10bo
29bo5bo5bo5bo29bo$9b2o29bo4bobo3bobo4bo29b2o$40bo4bobo3bobo4bo$28b2o
10b2o2b2obobobob2o2b2o10b2o$27b2o14bo2bobobobo2bo14b2o$29bo9bo5bo2bobo
2bo5bo9bo$44bobobobobobo$22bo17b3ob2obo3bob2ob3o17bo$20b2o23bo7bo23b2o
$21b2o21bobo5bobo21b2o$45b2o5b2o$2b2o36bo17bo36b2o$3bo35bobo15bobo35bo
$3bobo33bo2bo13bo2bo33bobo$4b2o34b2o15b2o34b2o3$8b2o34b2o7b2o34b2o$7bo
2bo20b2o11bobo5bobo11b2o20bo2bo$8bobo21b2o12bo5bo12b2o21bobo$9bo21bo
14b2o3b2o14bo21bo$3b2o89b2o$3bobo10bo65bo10bobo$4bobo10b2o61b2o10bobo$
2bobo3bo7b2o63b2o7bo3bobo$bobo4b2o27bo23bo27b2o4bobo$bobob2ob2o25b3o
23b3o25b2ob2obobo$2o6b2o24bo29bo24b2o6b2o$2bob2o2bo26bo27bo26bo2b2obo$
2bobo27bo2bo27bo2bo27bobo$b2o2bo2bobo11b2o8bo33bo8b2o11bobo2bo2b2o$4b
2o3bo11bo2bo4b3o2bo29bo2b3o4bo2bo11bo3b2o$21bobo4b4obo31bob4o4bobo$22b
o53bo$27bo2bobo33bobo2bo$17b2o11bo37bo11b2o$18bo7b3o41b3o7bo$15b3o63b
3o$15bo67bo!

With dependent reflectors, in general there are only glider/spaceship streams, and there are no actual signals of interest. It was already mentioned in this thread that there are some ways to send signals through a dependent reflector -- but then the interesting thing becomes something else (and I think the dependent reflector is not anymore dependent reflector, but is best described as some other device).

[Scorbie]

I think this depends on the specific situation. When there is a specific engineered pattern, and someone would like to describe or discuss this pattern, there may be some "interesting" information flow that is best described in terms of signals, and some "boring" information flow that is best described in terms of (e.g.) spaceships.

E.g. suppose the pattern is a large Herschel loop and contains circulating signals that are duplicated and sent (e.g. as gliders) for some purpose to some other place. It can be convenient/useful to describe these circulating active reactions as signals -- this is the "interesting information" in this case. At the same time, something like suppressed first natural gliders from those Herschels would be "boring information" -- if those suppressed gliders are mentioned in discussion, probably they would be described as gliders (rather than as signals).

The problem with this approach to wiki editing is that your "interesting" information depends on context, while patterns themselves in isolation don't have context. Your approach can be applied to Snarks, or any other signal circuitry, because it may or may not transmit "interesting" information depending on context.

In oscillators that are now called dependent reflector loops, there are glider streams moving between dependent reflectors. I think these glider streams are best described as glider streams that keep the dependent reflector reactions from self-destructing and allow them to work correctly at their period. So these glider streams are a way to support the interesting reactions that are demonstrated in the wiki page. I believe the interesting thing in this case is the dependent reflector itself -- the glider streams are "boring", and I suggest to describe them as (uninterrupted) glider streams.

I believe that's a narrow view of the dependent reflector that only considers the context of dependent reflector loops. You don't need to "keep the reflector from destructing" for it to be meaningful. I believe it's a valid take to consider them as one-time turners of consecutive gliders (this can encode any natural number N so someone can make a circuit that contains "interesting" information with this). This is in line with dvgrn's interpretation of "signal", and I think is a valid formalization of dvgrn's idea. Also depending on context the "dependent reflector" or "dependent reflector loop" may contain "interesting" data. One might use the it as a cheap way to encode infinity for example.

My point is this approach can lead to subjective and inconsistent edits, and I'm not sure if this approach is worth it especially considering it's breaking backwards compatibility.

It was already mentioned in this thread that there are some ways to send signals through a dependent reflector -- but then the interesting thing becomes something else (and I think the dependent reflector is not anymore dependent reflector, but is best described as some other device).

So that's why you want to keep them alive. To me an object is meaningful even if it can be used once. A one-time-pad is a one-time-pad even if it doesn't work perpetually. Maybe for someone who might misunderstand that this pattern does (transmit signals) AND (live perpetually) an explanation would be nice.

[confocaloid]

I suggest the following definition for the Dependent reflector article:

A dependent reflector is a periodic reflector that depends upon a regular and uninterrupted input stream of spaceships -- i.e. it will fail if any of spaceships is missing.

The definition I'm proposing should remain valid in different contexts, regardless of whether or not it is convenient/useful to describe the input/output spaceships/subsets of spaceships as something that carries a meaningful signal from one place to another.
If those spaceship streams are nothing more than a way to support some active reaction, then they're still streams of spaceships.
If those spaceship streams are actually ways to deliver some interesting information through an engineered pattern, then again, they are still streams of spaceships.

On the other hand, I think it is incorrect to define a dependent reflector in terms of signal streams. I think so, because in general those spaceships are not required to carry any particular signals, and also because signals may exist in many other forms besides spaceships. So I think the current definition in Dependent reflector does not accurately answer the question "What is a dependent reflector?" (especially to a newcomer who may be struggling to understand basic concepts).

To be clear: my suggestion would not change the common usage of existing terms. I'm only suggesting (what I think to be) a valid definition of the term "dependent reflector".

Further, in the definition of "dependent reflector loop" on the same LifeWiki page, there is a sentence "Any missing signal in the loop will cause the oscillator to fail." This sentence is redundant (because of the definition of "dependent reflector"), and it unnecessarily requires that the spaceships be considered "signals" (it may or may not be convenient/useful to view them as signals, depending on context -- but they're always spaceships). I suggest to remove that sentence due to redundancy and potential for confusion.

I think this depends on the specific situation. (...)

The problem with this approach to wiki editing is that your "interesting" information depends on context, while patterns themselves in isolation don't have context. Your approach can be applied to Snarks, or any other signal circuitry, because it may or may not transmit "interesting" information depending on context.

I think when a wiki page discusses some specific pattern (e.g. a Snark) in itself, that pattern (along with some ways to use and understand it for what it is on the basic level) becomes the context. So the Snark is a stable glider reflector. On the basic level, it reflects gliders. It does not "know" whether those gliders are carrying some important signals, it's just reflecting them (i.e. doing what it's supposed to do).

The R-pentomino is a methuselah, a long-lived five-cell configuration. Just because that configuration appears in signal circuitry, doesn't mean the signal-circuitry-related discussion should appear on the top of the wiki page. On the basic level, gliders emitted by the R-pentomino evolving in surrounding empty space are just gliders and nothing more -- even though someone might e.g. catch them and build a large high-period oscillator that repeatedly constructs a R-pentomino that goes all the way through its 1103-tick evolution, and then clean up the ash after it, and repeat again.

I don't think my suggestions break backwards compatibility in any way. Maybe there's a misunderstanding.

[Scorbie]

Okay, did I understand correctly that you think the following sentence is correct:

All dependent reflectors can depend on a regular and uninterrupted input stream of signals; i.e. it will fail if any of the signals are missing.

But you want the term "dependent reflector" to be more general than this?

If the above is true, then the idea you have in mind is consistent as far as I can see.
I don't have any objections to any system as long as they're consistent and is practical (e.g. preserves backwards compatibility), but I'm not everybody so I can't say how others would think about it. Specifically, I expect there to be conficting views on:

1. Whether to see a term of a pattern as a pattern itself or a pattern with a context
2. Whether all gliders transfer information

There may be plenty of other things worth of discussion that I have missed.

[dvgrn]

I suggest the following definition for the Dependent reflector article:

A dependent reflector is a periodic reflector that depends upon a regular and uninterrupted input stream of spaceships -- i.e. it will fail if any of spaceships is missing.

I fired up the LifeWiki article intending to try editing this definition in verbatim (plus or minus a "the" where needed).

I don't really object to the definition in that form, but it made me think about possible exceptions. When I tried to generalize the definition so that it wouldn't run in to trouble if and when new types of dependent reflector are discovered, I found that I really wanted to continue to be able to use the word "signal". Here's what I ended up with:

A '''dependent reflector''' is a [[periodic]] [[reflector]] that depends upon a regular and continuous input stream -- i.e., it will fail if any input [[signal]] is missing. Known dependent reflectors are supported by streams of [[spaceship]]s, usually [[glider]]s. [[Elementary]] dependent mechanisms that turn other [[signal]] types are not currently known; they might be more likely be referred to as '[[dependent]] [[elbow]]s' rather than 'dependent reflectors', since the definition of 'reflector' currently applies specifically to spaceships.

That's maybe too complicated to be a first-paragraph definition, though, so I didn't make the change.

It's interesting that "dependent elbow" would be kind of a new term. If someone happened to find a 2c/3 signal loop with a corner that only worked at p19 when supported by incoming 2c/3 signals, they might well end up calling it a "p19 2c/3 dependent reflector loop", and they wouldn't be terribly wrong. It would be only a slight extension of the term "reflector", not anything that would invalidate current usage.

Anyway, I didn't make any changes to the article yet. Will think about it some more when I get another chance.

How About Swimmers? (for example)
Meanwhile, here's a thought experiment I'm curious about: it's not beyond the realm of possibility that an elementary 90-degree turning mechanism for a switch-engine channel might be found someday. Seems like that might well get called a "reflector", even though swimmers aren't spaceships.

We can build engineered swimmer elbows today if we want to. Dependent swimmer elbows aren't likely to be terribly useful or interesting -- but they will probably be a little smaller on average than independent swimmer elbows, because building the new swimmer requires synchronization between several signal streams generated by the input swimmer. If you can shorten those streams independently mod N, you'll often end up being able to pack things tighter than if you require absolute synchronization -- but that will make the whole thing into a dependent mechanism.

[confocaloid]

(edit: removed part -- there is at least one error in my reply, and anyway it's unlikely to be important for the discussion of the terminology/content of the wiki page.)

I'll try to explain the intutive idea.

A dependent reflector can be thought of as a device that redirects an infinite glider generator generating a glider at A to an infinite glider generator generating a glider at B after delay generations.

Code: Select all

x = 43, y = 88, rule = LifeHistory
3D5.3D7.25D$D7.D.D6.D12.A11.D$D.D5.D.D5.D12.A.A10.D$D.D5.D.D4.D13.A.
3A8.D$3D5.3D3.D11.2A.A4.A7.D$13.D12.A.A.ABABA7.D$12.D15.A.A.2A4.2A2.D
$11.D2.2A11.ABA4B4.A3.D$10.D4.A9.3BA7BA.A3.D$10.D4.A.A6.12B2A4.D$10.D
5.2A5.13B6.D$10.D9.16B2.2D2.D$10.D8.17B2.D.D.D$10.D7.2AB3A12BD.3D.D$
10.D6.A2BA12B.DBDBD.D.D$10.D3.2A.3BAB2A8B3.2DB2D2.D$10.D2.A3B2AB2A4B
2A2B6.4B2.D$10.D2.A2B.4A5B2A2B7.4BD$10.D3.A3.2BA7B10.2BDB$10.D8.3A8B
9.D3B$10.D6.7B4.2A8.D.3B$10.D5.3AB.2B.B3.A8.D3.2B$11.D3.3BA3.B2AB3.3A
4.D5.B$12.D.3BA5.2A6.A3.D$13.D3B17.D$12.2BDB17.D$11.4B18D$10.4B$3D6.
4B$D.D5.4B$3D4.4B$D.D3.4B$D.D2.4B$4.4B$3.4B$2.E3B$.B2EB$BEBE13$3D.3D.
3D$D3.D3.D.D19.A$D.D.3D.3D18.A.A$D.D3.D3.D18.A.3A$3D.3D.3D15.2A.A4.A$
26.A.A.ABABA$28.A.A.2A4.2A$14.2A11.ABA4B4.A$15.A9.3BA7BA.A$15.A.AB5.
12B2A$16.2AB.B2.13B$18.18B$18.18B$18.2AB3A12BE$17.A2BA12B.EBEB$14.2A
4BAB2A8B3.2E2B$13.A3B2AB2A4B2A2B6.4B$13.A3B4A5B2A2B7.4B$14.A3.2BA7B
10.4B$19.3A8B9.4B$17.7B4.2A10.3B$16.3AB.2B.B3.A12.2B$15.3BA3.B2AB3.3A
10.B$14.3BA5.2A6.A$13.4B$12.4B$11.4B$10.4B$9.4B$8.4B$7.4B$6.4B$5.4B$
4.4B$3.4B$2.C3B$.B2CB$BCBC!

For instance, with this definition, the pattern inside the red polygon is a p59 dependent reflector. It reflects a NEwards glider stream with the firstmost glider at A to a SEwards glider stream with the firstmost glider at B after 59 generations.
You can "bind" the glider at A at gen 0 to the glider at B at gen 59.
The bound glider pairs may change if we enclose a different region to serve as a dependent reflector, but it's okay to me as long as the explanation is consistent. If you think of the reflectors as a black box then the two different enclosings are different reflectors.

Edit: I was not satisfied because I cannot define the bindings at gens 0..58. If it bothers you then just let the reflector already have an output glider at gen 0 (let delay = 0) and bind A and B.

I think this "infinite-glider-generator-redirecting device" explanation is much more intuitive than anything that requires the spaceships to be interpreted as "signals".

For me, a period-N dependent reflector is (essentially) a period-N gun -- except that it requires an incoming period-N stream of spaceships of the same type. There are no meaningful signals in this picture.

[dvgrn]

I think this "infinite-glider-generator-redirecting device" explanation is much more intuitive than anything that requires the spaceships to be interpreted as "signals".

For me, a period-N dependent reflector is (essentially) a period-N gun -- except that it requires an incoming period-N stream of spaceships of the same type. There are no meaningful signals in this picture.

It seems clear by this point that that's your opinion. Have I also been able to make it clear that not everyone shares your opinions about what is "less intuitive" or "more confusing"? Seems like we're still going to have to be careful not to have LifeWiki edit wars where I "fix" some wording so that it looks right to me, and then you revert it, or vice versa.

Meaning Seems to Be in the Eye of the Beholder
With my usual way of communicating about signal circuitry, there are two sets of (what I would call) "meaningful signals" in a period-N dependent reflector. There's a period-N input stream, made up of incoming signals -- gliders, *WSSes, or whatever. Without those continuous input signals, the reflector will fail. Again, this is my own personal standard phrasing.

The dependent reflector also produces a continuous stream of output signals. It's easy to set up a 1-to-1 correspondence between inputs and outputs, and usually it's very clear which input signal goes with which output signal: the first output that goes missing when you remove an input, is most likely the corresponding output signal for that input signal.

-- Yes, there are rare, mostly deliberately engineered exceptions, and it's okay that they exist -- they don't affect the terminology question here. For dependent reflectors, the specific choice of 1-to-1 correspondence isn't particularly important to any likely potential uses, and it doesn't imply that bits of binary information are moving through the reflector. But none of that means that it's bad or confusing to call the input and output streams "period-N signal streams".

Common Example Not Mentioned Yet: Pseudo-Period Guns
I'm probably mostly thinking about medium-complexity constructions like pseudo-period glider guns, which often use multiple synchronized period-N signal streams to build a period N/k gun. Like the dependent-reflector case, if any single signal is removed anywhere, the entire structure will generally fail.

The above is a case where I say "signal" because "glider or *WSS or Herschel or B or R or pi or whatever" is highly awkward. The term "spaceship" is also too clunky to use, because it makes people think of *WSSes when gliders are more common, and it doesn't apply to Herschels and the like. Only "signal" is general enough to be usable.

The smallest 90-degree signal turn that fits in some particular weird-shaped corner in a pseudo-period-N gun may well turn out to be a dependent reflector instead of a Snark, bumper or bouncer. But "signal" is still the correct term to use for all of the periodic ... signals ... that are flowing through all of the signal circuitry that makes up the pseudo-period gun. Nobody's going to stop and worry about whether they really shouldn't call a glider in some particular location a "signal" because it just came out of a dependent reflector, or is about to go into one. It can still be very correctly and conveniently referred to as a period-N signal.

[confocaloid]

I don't really object to the definition in that form, but it made me think about possible exceptions. When I tried to generalize the definition so that it wouldn't run in to trouble if and when new types of dependent reflector are discovered, I found that I really wanted to continue to be able to use the word "signal". Here's what I ended up with:

A '''dependent reflector''' is a [[periodic]] [[reflector]] that depends upon a regular and continuous input stream -- i.e., it will fail if any input [[signal]] is missing. Known dependent reflectors are supported by streams of [[spaceship]]s, usually [[glider]]s. [[Elementary]] dependent mechanisms that turn other [[signal]] types are not currently known; they might be more likely be referred to as '[[dependent]] [[elbow]]s' rather than 'dependent reflectors', since the definition of 'reflector' currently applies specifically to spaceships.

That's maybe too complicated to be a first-paragraph definition, though, so I didn't make the change.

Common Example Not Mentioned Yet: Pseudo-Period Guns
I'm probably mostly thinking about medium-complexity constructions like pseudo-period glider guns, which often use multiple synchronized period-N signal streams to build a period N/k gun. Like the dependent-reflector case, if any single signal is removed anywhere, the entire structure will generally fail.

The above is a case where I say "signal" because "glider or *WSS or Herschel or B or R or pi or whatever" is highly awkward. The term "spaceship" is also too clunky to use, because it makes people think of *WSSes when gliders are more common, and it doesn't apply to Herschels and the like. Only "signal" is general enough to be usable.

I assume that the word "signal" is used here as a shorthand for "glider or *WSS or Herschel or B or R or pi or whatever".
I think the current wording is not helpful.
With the above assumption, "signal reflector" reads as "reflector that can reflect gliders, *WSSes, Herschels, ... or whatever". But this is too general to be helpful -- the term "reflector" is used/defined only for things that can reflect spaceships specifically.
Other uses of "signal" on that page also create the impression that the word "reflector" is used for things that can reflect something that isn't a spaceship. I think this does not reflect current usage.

I think "dependent reflector" is currently used only for things that reflect (uninterrupted period-N) spaceship streams.
I think the wiki page should accurately reflect the current use, without trying to predict future terminology.
Attempting to predict possible future uses will lead to a too-general-to-be-useful definition, likely to create more and more and more confusion.

Edit 2: attempted to clarify the wording.