Dependent Reflector Terminology and Analysis

[confocaloid]

Is "active region" meant to be a defined term, or just an informal convenience jargon?

It's not formal. There are edge cases, such as something that lasts 15 generations before becoming a beehive without really expanding.

Feel free to reword the page Active region.

Regarding the issue of signals, I believe the current definition of signal in LifeWiki is a correct definition of the concept "signal".
https://conwaylife.com/w/index.php?titl ... did=116950

There are other usages of the same word -- e.g.
* in the context of wires (Wire)
* in informal discussions, as a synonym of "something that moves through a circuit"
* ...
To handle multiple usages/meanings, one would need to create a disambiguation page Signal (disambiguation).
Lumping everything into a single page would not make a definition -- instead, the word "signal" would become entirely useless.

A signal is the movement of information through the Life universe. Signals can be carried by spaceships, fuses, drifters, or conduits. Spaceships can only transfer a signal at the speed of the spaceship, while fuses can transfer a signal at speeds up to the speed of light.

In practice, many signals are encoded as the presence or absence of a glider (or other spaceship) at a particular point at a particular time. Such signals can be combined by the collision of gliders to form logic operations such as AND, OR, and NOT gates. Signals can be duplicated using glider duplicators or other fanout devices, and can be used up by causing perturbations on other parts of the Life object.

Signals are used in pseudo-random glider generators, the unit Life cell and the Fermat prime calculator, among others.

(edited 2023-08-09 13:03 UTC)

[dvgrn]

I disagree. It is not "my definition" of signal. It's the LifeWiki definition of the concept signal

Your interpretation of the LifeWiki definition does not match my interpretation. I'd like to add a paragraph so that it does match my definition.

A key difference in our interpretations is this phrase in the second paragraph:

In practice, many signals are encoded as the presence or absence of a glider (or other spaceship) ...

The definition very deliberately says "many", not "all". Some signals -- like the signals in the p6 signal sources and sinks mentioned above -- can not be added or removed to convey bits of information in an information-theory sense.

In my opinion, the entire paragraph beginning "In practice..." refers only to this specific type of signal, since the next sentence starts "Such signals...". These particular signals are both dvgrn-signals and confocaloid-signals. In practice, they make up the great majority of the active objects traveling through signal circuitry.

The definition of "signal" is currently imprecise about what other types of signals there are, that don't fit into this "many signals" group. I'd say that the signals in the p6 signal sources and sinks would be included there. I'm fairly confident that a small adjustment to the "signal" article will allow a single definition to include these types of signals, while conforming to other current and past usage of the term, and avoiding a lot of changes to other definitions.

The rule "if it's moving through signal circuitry, then it can be safely called a signal" [...]

Then what is "signal circuitry", exactly? That notion is not defined.

As a good starting point, the patterns in the Signal-Circuitry folder in Golly are good examples of signal circuitry. Reflectors, splitters, logic gates, etc., are all signal circuitry. I've previously brought up the question of whether dependent reflectors should also be considered to be signal circuitry, as a matter of convenience and consistency -- they're called "reflectors", so it seems useful to treat them as if they're reflecting something. But that question opened a somewhat different can of worms, and then we went around and around about that for a long time.

@Everyone else:

I think I can speak for confocaloid on this point: confocaloid and I are still very much more interested in what other people have to say on this topic of the proper definition and usage of "signal". We've already heard what we each have to say, many times now. Without some additional input from others, I'm unwilling to have confocaloid continue to remove the term from LifeWiki articles due to what I see as a problematic application of a definition that isn't really usable in the context of signal circuitry.

[confocaloid]

@Everyone else: I think I can speak for confocaloid on this point: confocaloid and I are still very much more interested in what other people have to say on this topic of the proper definition and usage of "signal". We've already heard what we each have to say, many times now. Without some additional input from others, I'm unwilling to have confocaloid continue to remove the term from LifeWiki articles due to what I see as a problematic application of a definition that isn't really usable in the context of signal circuitry.

I'm unwilling to use "signal circuitry" without understanding what it is, exactly.

In practice, many signals are encoded as the presence or absence of a glider (or other spaceship) ...

The definition very deliberately says "many", not "all". Some signals -- like the signals in the p6 signal sources and sinks mentioned above -- can not be added or removed to convey bits of information in an information-theory sense.

Those "signals" are signals in a different sense. That's a different usage of the word "signal" (see Wire or light speed oscillators).

As a good starting point, the patterns in the Signal-Circuitry folder in Golly are good examples of signal circuitry.

Those examples are wonderful. However, they do not make a definition of "signal circuitry" that would be comparable to the current LifeWiki definition of the concept "signal".
"Signal circuitry" is not defined. "Signal" is:

EDIT: for the record, here's the current definition of "signal", copied here:

A signal is the movement of information through the Life universe. Signals can be carried by spaceships, fuses, drifters, or conduits. Spaceships can only transfer a signal at the speed of the spaceship, while fuses can transfer a signal at speeds up to the speed of light.

In practice, many signals are encoded as the presence or absence of a glider (or other spaceship) at a particular point at a particular time. Such signals can be combined by the collision of gliders to form logic operations such as AND, OR, and NOT gates. Signals can be duplicated using glider duplicators or other fanout devices, and can be used up by causing perturbations on other parts of the Life object.

Signals are used in pseudo-random glider generators, the unit Life cell and the Fermat prime calculator, among others.

There are different things people can mean when they say "signal". Trying to make a single definition to cover everything is misguided.
I'd strongly prefer to keep an existing definition that actually defines a concept signal. For other meanings, a disambiguation page is the way to go.

[dvgrn]

I'm unwilling to use "signal circuitry" without understanding what it is, exactly.

Nobody is forcing you to use "signal circuitry" or any other term.

Not surprisingly, I've used the term "signal circuitry" a whole lot, over many years. No one has ever had any trouble understanding what I meant (or if they did, I didn't hear about it, I guess).

It's probably just like many other Life-related concepts, in that the meaning is generally very clear from context, but it's just plain not possible to create an iron-clad definition that covers every possible weird case. Just like trying to define "methuselah" unambiguously, you can easily create infinite pointless arguments by making the attempt, but what you can't do is end up with a definition that everybody agrees is the best possible definition.

Signal circuitry is -- according to previous usage, which is more often than not me saying "signal circuitry" -- circuitry that's capable of repeatably carrying dvgrn-signals from one point to another, or processing them in various ways.

Conduits and splitters, as well as switches and other types of logic gates, all fall under the umbrella of "signal circuitry". I don't think there's any particular confusion on this point -- "signal circuitry" hasn't even needed a LifeWiki definition until now, apparently.

I would say that based on traditional usage going back at least to the 1990s, p6 signal sources and sinks -- and other circuitry that depends on receiving a constant stream of dvgrn-signals -- could still be counted as "signal circuitry".

It's not particularly important to me whether dependent circuitry is part of the formal definition of "signal circuitry" or not, since it's generally better to use the word "dependent" when that feature is important to the discussion. But I do definitely want to be able to say things like, "Such-and-such dependent reflector mechanism has to be supported by a continuous stream of period-N input signals, and produces a continuous period-N stream of output signals". (By "signal" here, of course I'm meaning dvgrn-signals, not confocaloid-signals.)

I'm sorry, but I just don't think it's a good idea to try to fit the information-theory meaning of "signal" into this particular context. "Signal" hasn't been consistently used with that meaning in the past, here on the forums, and it's very unlikely to start getting consistently used that way now.

Appeals to the commonly understood real-world definitions of "signal" aren't really relevant, because "signal" has a whole lot of common real-world meanings too; e.g., carrier signals are more like the signals in p6 signal sources and sinks than like the bit-carrying confocaloid-signals that you're focusing on.

[confocaloid]

Conduits and splitters and reflectors are circuitry. Adding the word "signal" does nothing but create confusion. (I'm not the only one who expressed this concern in this thread -- please read earlier replies from other people.)

p6 signal sources and sinks are not circuitry. They're (parts of) oscillators. Oscillators have rotors, stators, etc.

Trying to fit everything into "signal circuitry" at all costs is misguided and counter-productive. There are many many things that just plain are not circuitry.


"Signal circuitry" is not defined -- just because it's a name of a folder in a pattern collection, doesn't mean it makes sense.
"Signal" is defined:

EDIT: for the record, here's the current definition of "signal", copied here:

A signal is the movement of information through the Life universe. Signals can be carried by spaceships, fuses, drifters, or conduits. Spaceships can only transfer a signal at the speed of the spaceship, while fuses can transfer a signal at speeds up to the speed of light.

In practice, many signals are encoded as the presence or absence of a glider (or other spaceship) at a particular point at a particular time. Such signals can be combined by the collision of gliders to form logic operations such as AND, OR, and NOT gates. Signals can be duplicated using glider duplicators or other fanout devices, and can be used up by causing perturbations on other parts of the Life object.

Signals are used in pseudo-random glider generators, the unit Life cell and the Fermat prime calculator, among others.

There are different things people can mean when they say "signal". Trying to make a single definition to cover everything is misguided.
I'd strongly prefer to keep an existing definition that actually defines a concept signal. For other meanings, a disambiguation page is the way to go.

[dvgrn]

Conduits and splitters and reflectors are circuitry. Adding the word "signal" does nothing but create confusion. (I'm not the only one who expressed this concern in this thread -- please read earlier replies from other people.)

I've read them -- several times now. The opinions I'm seeing expressed are far more ambiguous than the hard line you've been taking. There are also people who have said "there's no harm in calling them signals". There is not a consensus on this point.

So far, there is one very strong advocate of your interpretation of "signal" -- you. I'm mirroring that by putting up an equally strong defense of my preferred definition. I think that in this particular case, the definition of "signal" can be usefully extended and clarified to include the signals in p6 signal sources and sinks, in a single "dvgrn-signal" definition, without damaging people's understanding of current usage or requiring extensive changes to other definitions. I don't think that it's possible in practice to limit the existing term "signal" to mean only information-theory "confocaloid-signals", so I'd like to discourage attempts to make changes according to that theory.

The rest of what you've said in the most recent message seems to be copied and pasted without any changes from previous posts. I'm not finding that a helpful discussion tactic, since I've already read all of it, and responded to it to the best of my ability.

[confocaloid]

I think that in this particular case, the definition of "signal" can be usefully extended and clarified to include the signals in p6 signal sources and sinks, in a single "dvgrn-signal" definition, without damaging people's understanding of current usage or requiring extensive changes to other definitions. [...]

If it's going to be a "dvgrn-signal" definition, it should be clearly separated from the existing definition of signal.

The p6 signal sources and sinks don't contain any signals, under the existing LifeWiki definition. That's correct -- those oscillators are not examples of circuitry.

The rest of what you've said in the most recent message seems to be copied and pasted without any changes from previous posts. I'm not finding that a helpful discussion tactic, since I've already read all of it, and responded to it to the best of my ability.

This is a defense tactic.
You said it several times that you would like to hear opinions from other people -- yet you're continuing to defend your own interpretation, at the expense of everyone else. I'm trying to defend myself against this.
(It's not "the other way round" because you're a long-term member and you certainly ought to know/understand what's the issue and what are related pitfalls/confusion/etc.)

Please let other people to give their opinions, without entering into basically the same type of argument each time.

[dvgrn]

This is a defense tactic...
Please let other people to give their opinions, without entering into basically the same type of argument each time.

I believe that all of what you've just written applies equally well to your recent posts, as to mine. I don't believe I've been using any tactics that you haven't been using.

If you find these tactics unpleasant when they're aimed at you, please consider how I'm feeling. I've spent a lot of time in the last decade working on LifeWiki and Life Lexicon definitions, to the best of my ability. Now they're gradually being damaged (in my view) by someone who seems to be refusing to stop making edits that remove the word "signal", even after four polite requests to do so.

So -- here's the next question for you:

Suppose, hyoothetically, that everyone else on the forums is completely and permanently turned off and/or scared away from this thread, and that no one wants to express an opinion in favor or in opposition to either of our positions? (As I've said before, I wouldn't blame them -- we've made this a very unpleasant and unreadable thread, unlike almost everything else on the forums.)

I've been trying to ask for more independent opinions and/or moderation here for quite a while now. hotdogPi brought up a different point, about "active region", but so far things aren't working out well for getting a broad community response about "signal".

Supposing that that is true, what should the next step be?

@Everyone else: I'm definitely not trying to discourage other responses here -- and I won't respond to them in the way that I've been having to respond to confocaloid, because nobody else is repeatedly removing the word "signal" from LifeWiki articles while this discussion is still ongoing. Please, bring on the independent opinions!

[confocaloid]

[...]
Supposing that that is true, what should the next step be?
[...]

My answer is "wait longer".

Instead of demanding/expecting that everyone interested respond in a short time (as you did on 2023-07-03 on the LifeWiki talk page Talk:Dependent reflector) -- give more time, until things become clear.

If it's not clear after a week / a month / several months whether there's a consensus one way or another, give more time.

Saying something like "I'll probably give this a day to collect any other thoughts, then go ahead..." is counter-productive -- and that your reply was at the very beginning of this discussion.
---
That same reply by Dvgrn also contains phrase:
"The first name was invented by (username omitted) back in 2014, and the second was just a copycat use of the same term by (username omitted). Once those naming errors are repaired..."

In my view, this is a failure to acknowledge that the two members mentioned there are enthusiasts like yourself, and have made significant contributions.
In my view, this is also a failure to acknowledge that naming is hard (so you shouldn't blame people for not doing it in the best possible way).

That happened at the very beginning of this discussion (in particular, before any of my replies). Instead of trying to help people, you're attacking them.
With greater power comes greater responsibility.

[dvgrn]

[...]
Supposing that that is true, what should the next step be?
[...]

My answer is "wait longer".
...
If it's not clear after a week / a month / several months whether there's a consensus one way or another, give more time.

That sounds fine to me. In the mean time, my request still stands.

Please stop removing "signal" from LifeWiki articles, or making any other edits that rely on your implicit assumption that your "confocaloid-signal" definition of "signal" is the correct one. Once a consensus emerges, as you say, then it will be safe to go back to changing the current status quo -- assuming that there's broad agreement with your position, of course.

That same reply by Dvgrn also contains phrase:
"The first name was invented by (username omitted) back in 2014, and the second was just a copycat use of the same term by (username omitted). Once those naming errors are repaired..."

In my view, this is a failure to acknowledge that the two members mentioned there are enthusiasts like yourself, and have made significant contributions.
In my view, this is also a failure to acknowledge that naming is hard (so you shouldn't blame people for not doing it in the best possible way).

That happened at the very beginning of this discussion (in particular, before any of my replies). Instead of trying to help people, you're attacking them.
With greater power comes greater responsibility.

@confocaloid, this all seems to me like an irrelevant and somewhat wild accusation. No blame is assigned in that quote -- it looked to me like just a minor terminology error, the kind of thing that you and I have both routinely fixed in passing on the LifeWiki, in hundreds of cases. This particular error could have been very easily corrected as long as no one objected. You didn't include the part of that quote where I specifically said "(unless I hear objections)". Someone did object, and so I have not made those suggested changes.

This is all exactly the way this is supposed to work, as far as I can tell.

I'm very interested to hear directly from anyone that I've supposedly attacked, and am very happy to apologize for any feelings that I've inadvertently hurt. I certainly had no intention of offending or insulting anyone, but as Sokwe has pointed out, it's always very easy to make mistakes in this area.

I don't think it's a good idea for you to try to defend other people in cases like this. I was discussing this exact issue with one of the people who did the naming, at the time that I made that post. The other person involved hasn't been active on the forums for a couple of years.

As far as I can tell, nobody was particularly offended or insulted -- so nobody really needs defending in this case. Again, I expected at the time that this would be a non-controversial renaming issue, just like dozens or hundreds of other minor mistakes and inconsistencies that you and I have dealt with over the years.

[confocaloid]

You should not require "a day to collect any other thoughts". Give more time by default.

You should not blame/single out people for imperfect naming. Naming is hard.

Apart from that, the same reply shows a general "superiority" attitude. Other people may know/understand something that you do not know / do not understand.

(Same holds for many other replies by dvgrn in this discussion, on Talk:Dependent reflector and in Oscillator Discussion Thread and in this thread. I'm not going to list all cases explicitly, as that's useless unless dvgrn agrees with me on this.)

Just to give my two cents here, "dependent reflector loop" seems to me like a much better name for this article than "dependent glider shuttle". Mainly this is because the patterns being referenced, [[p26 glider shuttle]] and [[p31 glider shuttle]], are not actually shuttles and badly need to be renamed. Shuttles go back and forth, not round and round.

The first name was invented by (username omitted) back in 2014, and the second was just a copycat use of the same term by (username omitted). Once those naming errors are repaired, it doesn't seem like there's going to be any reason to rename this article. These things really '''are''' loops made out of [[dependent reflector]]s, and that's a term that's been in use for a long time and is well recognized.

I'll probably give this a day to collect any other thoughts, then go ahead and rename those two non-glider-shuttles and patch up this article appropriately (unless I hear objections). [[User:Dvgrn|Dvgrn]] ([[User talk:Dvgrn|talk]]) 19:36, 3 July 2023 (UTC)

In my view, that's highly arrogant. That reply came before any of my replies.

[yoleo]

To summarize:

confocaloid-signal: informational content. Immutable and unchanging: signals don't flip or turn, because they are not spatial. Avoids confusion with "active objects."

dvgrn-signal: the information or the encoding, depending on context. May be used to refer to "objects that can carry information" in circuitry-adjacent patterns, like gliders in a memory loop or wire-based oscillators. We've been using the term like this for a while now.

Looking through confocaloid's list of objections to use of "signal," they're basically all places where the word is used to mean "the input/output of a conduit" (like signal circuitry). [Now, is the word "signal" adding something meaningful to each of those articles? Perhaps not. But that's a separate discussion.] That seems like reasonably typical usage of the word to me. Just from a practical perspective: when building things in CGoL, we often need to talk about conduit input/output. "Signal" is a convenient and suitable word for this. This sort of double usage is commonplace: text, speech, mail, etc. can refer to the medium (how it is sent) or to the message (the informational content). I don't see a real risk for confusion. Even Life Lexicon uses the term "signal converter" under the entry for "splitter," when strictly speaking it's only the medium that's changing, not the informational content.

[dvgrn]

Even Life Lexicon uses the term "signal converter" under the entry for "splitter," when strictly speaking it's only the medium that's changing, not the informational content.

Heh, unfortunately the current Life Lexicon probably shouldn't be considered to be an unbiased reference for the usage of "signal". There's no entry for "splitter" before 2018, which was when I put in the big Lexicon update. The appearance of the word "signal" in any new definition was most likely my doing.

If my memory is correct, I even thought of that update as being in large part a "signal circuitry update". Mike Playle's and Guam's and Kazyan's and others' work on new circuitry mechanisms all needed to be documented -- the Snark, the CC semi-Snark, the CP semi-Snark, tremi-Snark, and quadri-Snark, the syringe, the H-to-MWSS, and a pile of other small Lexicon-compatible conduits. And by that time I had started the collection thread that I called "Stable signal converters".

I used the word "signal" in many of these Lexicon definitions as a short common label. I agree with confocaloid that many of the definitions could get along all right if "signal" were removed. But I did add every instance of the term on purpose, not mechanically as confocaloid has suggested -- on the theory that it would be familiar terminology that would let readers know that these definitions were all in the same category. I was trying to make the use of the term consistent across the hundreds of uses of "signal" in the Lexicon.

One of the questions under discussion at this point is whether those additions of the term "signal" were an okay idea -- or whether they were a bad enough idea that they should be removed from the next version of the Life Lexicon, and from the LifeWiki entries based on the 2018 definitions, and so on.

Other new perspectives on this would be much appreciated!

[confocaloid]

confocaloid-signal: informational content. Immutable and unchanging: signals don't flip or turn, because they are not spatial. Avoids confusion with "active objects."

Agree.
To be clear: I believe that in these contexts, 'active object' is a strictly better wording than 'signal'. It's longer, but clearer.
Note that 'active object' was previously used for a long time in the context of inputs/outputs of conduits -- e.g. in the definition of 'track' or in the definition of 'conduit' in previous releases of Life Lexicon. More examples of 'active object' used in this sense can be found in other places.
(In other words, this is an existing usage of 'active object' -- not something invented by me.)

[Now, is the word "signal" adding something meaningful to each of those articles? Perhaps not. But that's a separate discussion.]

Indeed part of my objection is that 'signal' is often redundant (and therefore can be confusing to a reader).

If my memory is correct, I even thought of that update as being in large part a "signal circuitry update".

But I did add every instance of the term on purpose, not mechanically as confocaloid has suggested -- on the theory that it would be familiar terminology that would let readers know that these definitions were all in the same category. I was trying to make the use of the term consistent across the hundreds of uses of "signal" in the Lexicon.

I don't believe 'signal circuitry' is a well-defined category. (To be clear: having a folder named 'Signal-Circuitry' in a pattern collection doesn't suffice.)
Without a definition, there is no consistency -- one cannot reliably determine whether or not a term/concept/pattern belongs to the category.

Other new perspectives on this would be much appreciated!

Agree.

What 'signal circuitry' means? Is it a defined term? Is it a defined category?

I really think it should be written down explicitly, just what 'signal circuitry' generally means.
If you consider it a term, then it should be defined.
If you consider it a category, then there should be a good understanding of the distinction (what belongs vs. what doesn't belong).

I feel like 'signal circuitry' would be more restrictive than 'circuitry' (whatever that would mean). To me, 'signal circuitry' sounds like an advertisement of something that does some nontrivial/interesting computation. It is left unexplained what is 'nontrivial' or 'interesting', but it would be a restriction.

The p43 Snark loop is an oscillator. To me, it feels wrong to describe it as 'signal circuitry'.
Similarly, neither the p256 machine gun nor the p59 glider shuttle are examples of 'signal circuitry', in my opinion. The former is a gun that emits periodic glider stream; the latter is an oscillator.

Life Lexicon

The current Lexicon entry 'CC semi-snark' says: (I added the links to other entries)

CC semi-Snark A small 90-degree colour-changing glider reflector requiring two input gliders on the same lane for each output glider. It was discovered by Sergei Petrov on 1 July 2013, using a custom-written search utility. It functions as a very compact period doubler in some signal circuitry, for example the linear propagator. The semi-Snark can period-double a regular glider stream of period 51 or more, or an intermittent stream with two gliders every 67 ticks or more, since the block reset glider can be sent just 16 ticks before its partner.

Note that 'signal circuitry' is linked as two separate terms (a separate link 'signal', and a separate link 'circuitry').
If 'signal circuitry' is considered a single term, then the two words should instead be linked to one definition (as a single link).

The current Life Lexicon entry 'circuit' says: (I omitted the links)

circuit Any combination of conduits or converters that moves or processes an active signal. This includes components with multiple states such as period multipliers or switches, which can be used to build guns, logic gates, universal constructors, and other computation or construction circuitry.

I believe it should say 'active object' instead of 'active signal', for reasons already explained above, and because 'active signal' is neither a defined term nor a commonly understood phrase, as far as I can tell.

[dvgrn]

I don't believe 'signal circuitry' is a well-defined category. (To be clear: having a folder named 'Signal-Circuitry' in a pattern collection doesn't suffice.)
Without a definition, there is no consistency -- one cannot reliably determine whether or not a term/concept/pattern belongs to the category.
...
What 'signal circuitry' means? Is it a defined term? Is it a defined category?

I'll have a try at answering this question, since it's been asked a couple of times.

Signal circuitry is circuitry that can be re-used by dvgrn-signals traveling through it. This is distinct from one-time circuitry, which is generally cleanly destroyed after a single use.

So if I say "signal circuitry", I'm conveying the information that the circuitry in question is intended to accept dvgrn-signals repeatedly -- maybe strictly periodic, maybe intermittent, maybe asynchronous, but it won't be damaged after a single use. Conversely, if I say "one-time circuitry", I'm communicating the fact that the input (usually called a "trigger", but it's certainly also a type of dvgrn-signal) will only show up once.

The LifeWiki definition of circuit, as I read it, is compatible with my use of both "signal circuitry" and "one-time circuitry", if one-time converters are allowed as a type of converter. "One-time converter" is definitely not commonly used terminology -- the short form "OTT" is more generally used -- but it would be perfectly understandable in the context of a discussion about one-time circuitry.

A circuit is any combination of conduits or converters that moves or processes an active signal.

Again, to be clear, this all seems correct to me because I believe that the intended definition of the word "signal" in that context is "dvgrn-signal", not "confocaloid-signal". I don't want the definition of "signal circuitry" to hinge in any way on the question of whether its moving active objects are representing, or are capable of representing, unencumbered information-theory bits. What's important is the movement of the active objects themselves.

... I see that the definition of "circuit" was added in the 2018 update to the Life Lexicon. That means I probably wrote it myself. So this isn't anything like an unbiased data point, either! On the other hand, I think I should be considered to be a fairly reliable authority on what I intended the definition to mean, when I wrote it: by "signal" I meant "dvgrn-signal", since up to that point nobody had tried using "signal" in that context to mean "confocaloid-signal".

-- Or if they did, I hadn't noticed, which is always possible! Most actual usages of "signal" could equally well mean either "dvgrn-signal" or "confocaloid-signal". Where the distinction is important, I think the context will reliably make it clear which meaning was intended. When the distinction is *not* important, it seems very awkward to try to limit the term "signal" to the strict information-theory sense. That's not the only sense in which the term has been used in the context of Life.

The p43 Snark loop is an oscillator. To me, it feels wrong to describe it as 'signal circuitry'.

The way I've always used the term "signal circuitry" is completely agnostic to the question of whether bits of information can be encoded in the signals or not. A p43 Snark loop, according to this view, is a clear example of signal circuitry. It's a very small and simple example, but there's no question that it counts as signal circuitry -- as well as being an oscillator, of course.

[confocaloid]

... On the other hand, I think I should be considered to be a fairly reliable authority on what I intended the definition to mean, when I wrote it: by "signal" I meant "dvgrn-signal", since up to that point nobody had tried using "signal" in that context to mean "confocaloid-signal".

-- Or if they did, I hadn't noticed, which is always possible! Most actual usages of "signal" could equally well mean either "dvgrn-signal" or "confocaloid-signal". Where the distinction is important, I think the context will reliably make it clear which meaning was intended. When the distinction is *not* important, it seems very awkward to try to limit the term "signal" to the strict information-theory sense. That's not the only sense in which the term has been used in the context of Life.

I believe the common use/meaning of the word 'signal' is the meaning 'information carried by a moving object'. Information implies some choice and preservation-as-it-moves. The word 'signal' commonly means 'information' -- not 'moving pattern'.

The current LifeWiki definition of 'signal' says:

A signal is the movement of information through the Life universe. Signals can be carried by spaceships, fuses, drifters, or conduits. Spaceships can only transfer a signal at the speed of the spaceship, while fuses can transfer a signal at speeds up to the speed of light.

In practice, many signals are encoded as the presence or absence of a glider (or other spaceship) at a particular point at a particular time. Such signals can be combined by the collision of gliders to form logic operations such as AND, OR, and NOT gates. Signals can be duplicated using glider duplicators or other fanout devices, and can be used up by causing perturbations on other parts of the Life object.

I believe it's clear from the definition, that the intended meaning is "information carried through a pattern by a moving thing" -- i.e., not the moving thing itself (even though that occurs as a common jargon in conversations), and certainly not a moving thing that doesn't carry any information.
From what is said in the definition of 'signal', it follows that gliders in "dependent reflector loop" oscillators are not signals.
From what is said in the definition of 'signal', it follows that there has to be some possible variation in how the information-carrying things are used -- otherwise the information could not move.

When one is using a spaceship to carry a signal, one can delay/advance the spaceship, or one can choose between sending or not sending it, or one can send a signal encoded by a specific number of spaceships, etc
When one is using a conduit like Fx77 or R64 to carry a signal, one can delay/advance the input Herschel (delaying/advancing the output Herschel by the same number of ticks), or choose between sending/not sending a Herschel, or send a specific number of Herschels, etc.

That's why I believe that neither p43 Snark loop nor p26 glider shuttle are patterns containing any signals, in the LifeWiki sense of signal. They're oscillators, not signal-processing devices. There's no movement of information here; the description 'p43 Snark loop' already implies that gliders come strictly at p43 after each other, and no glider is absent; the whole arrangement repeats after 43 ticks, which is much earlier than one would expect if any kind of information travel actually happened.

TL,DR: I believe the LifeWiki definition of signal defines it in the context where moving things are used to carry information, and it (correctly) defines "signal" to mean "information carried by a moving object/pattern/perturbation".
Information can be measured in bits; there has to be some variation. The signal is preserved as it moves through the pattern, while the moving thing can be changed/transformed freely.

[Pavgran]

In the p43 Snark loop, the local behaviour (around a single Snark) is indistinguishable from the local behaviour of some signal-processing circuitry involving Snark. So, we might as well call the gliders in a p43 Snark loop 'signals', as they are locally signals.
With dependent reflectors, things are slightly more complicated. I would view a single reflector as a signal-processing device (with the gliders being the signal in the same local sense as with the Snark example), albeit with very low information entropy. It can essentially either continue operating, or fail to recieve an input signal and destruct. Essentially, the only information it can process is 'when to destruct', and in closed oscillator loop, the answer might be 'never', but that will be, in essence, 'locally never'. An information in the form of almost any glider from outside will destroy the loop, in one way or another.

I didn't want to post to that thread before, but now I formulated a clear enough (I hope) point that I want to share.
Also, I support more of dvgrn's points than confocaloid's.

[confocaloid]

In the p43 Snark loop, the local behaviour (around a single Snark) is indistinguishable from the local behaviour of some signal-processing circuitry involving Snark. So, we might as well call the gliders in a p43 Snark loop 'signals', as they are locally signals.
With dependent reflectors, things are slightly more complicated. I would view a single reflector as a signal-processing device (with the gliders being the signal in the same local sense as with the Snark example), albeit with very low information entropy. It can essentially either continue operating, or fail to recieve an input signal and destruct. Essentially, the only information it can process is 'when to destruct', and in closed oscillator loop, the answer might be 'never', but that will be, in essence, 'locally never'. An information in the form of almost any glider from outside will destroy the loop, in one way or another.

I didn't want to post to that thread before, but now I formulated a clear enough (I hope) point that I want to share.
Also, I support more of dvgrn's points than confocaloid's.

By definition, a dependent reflector is a periodic (period-N) reflector that requires a constant incoming period-N stream to operate.

It is left unspecified what will happen if the incoming stream is irregular. Here is an example:

Code: Select all

x = 117, y = 76, rule = B3/S23
32b2o58b2o$33b2o58b2o$32bo59bo$10b2o58b2o$10bo59bo$12bo59bo$8b5o55b5o$
8bo59bo$11b4o56b4o$11bo2bo56bo2bo6$2b2o58b2o$2bo59bo$3b3o52b2o3b3o$5bo
17b2o33bo6bo17b2o$2o21bo35b3o21bo$bo13b3o3bobo37bo13b3o3bobo$bobo10bo
3bo2b2o38bobo10bo3bo2b2o$2b2o9bo5bo42b2o9bo5bo$13bo5bo53bo5bo$13bo5bo
53bo5bo$14bo3bo55bo3bo$15b3o57b3o2$3b2o58b2o$4bo59bo$b3o14b2o41b3o14b
2o$bo16bo42bo16bo$19b3o57b3o$16bo4bo54bo4bo$15b2o58b2o$15bobo57bobo18$
35b2o58b2o$35bobo57bobo$35bo59bo18$55b2o58b2o$54b2o58b2o$56bo59bo!

The example on the left and the example on the right are (stator variants of) the same p79 dependent reflector. As long as they are viewed as dependent reflectors and used correctly, there is no functional difference -- the timing and positions of gliders are the same in both cases. However, they self-destruct in two different ways when the input stream is interrupted.

Similarly, it is possible to engineer two dependent reflectors (even with the same bounding box/envelope, so that either can be substituted for another) that function in the same way, but fail in two different ways (so that one dependent reflector continues to produce output gliders for some time, even after the other dependent reflector already stopped producing output gliders).
I would say such dependent reflectors are equivalent -- to be able to use a dependent reflector correctly, one only needs to know the relative positions/timing of input and output glider streams.

I think my issue with viewing dependent reflectors (and independent reflectors too) as "signal-related things" is that the concepts of signal, information, entropy are unneeded to explain what is a dependent reflector / what is a reflector. I believe the relevant concepts are "physical" (i.e. gliders and other spaceships, streams of spaceships, timing, phase, and so on) rather than "logical".
(edited 21:34 UTC)

[Jormungant]

Well, I mentioned my position before, It is by no mean a popularity contest, but I lean on the side of confocaloid. I am 100% fine with the definition of signal, but I see no gain in using the term "signal circuitry" over "circuitry", If someone sees the term "circuitry" and think it is too ambiguous, given that the assumed context is "game of life", I assume no reasonable being would think it is "penguin circuitry" or "lymphatic circuitry". So, as strange as is may, I would only use "signal circuitry" as a more constrained class of circuitry, I see the point of Pavgran that the lack of input could be seen as a signal, since destroying cleanly circuits allow catterloopillar-like patterns, thought I have yet to see one that involves dependent reflectors (some work in progress I am not aware of?) Still, if you re-read the last sentence I wrote, and take as literally as what is the I-do-not-know-why-it-is-mandatory definition of "signal circuit" implied by the "signal" definition, the lack of input is a signal, so is an "empty pattern" a signal then?

I do not really care either way, but this is definitely confusing, I guess I shared a weird pattern for some thought, I did put a p128 oscillator first in this thread, but I later realized (thanks to dvgrn's) that it could indeed host 2 signal independently, and is hence a true signal circuit, so I did not provide a good counter example. I falsely remembered that both signal stabilizes the other, and needs to be both in or out. Anyways, it is not as pretty, and feels like I am digging dirt to engineer some bizarre thing, but I got the following:

Code: Select all

x = 143, y = 128, rule = LifeHistory
8$83.A$83.3A$74.A11.A$74.3A8.2A5.4B$77.A7.11B$60.2A14.2A3.B5.11B$59.B
2AB13.8B2.11B2A$60.2B16.19B2A$55.B3.2B17.20B$54.2AB.4B15.20B$54.2A8B
11.21B$55.B.B2A6B2.2B2.25B$58.2A15BD14B4.4B$58.17BDBD4B.6B6.4B$54.21B
3D4B2.B.5B5.4B$54.23BD4B7.2A6.4B$53.2A26B8.A8.4B$53.2A14B.4B16.3A6.4B
$54.B.11B2.4B19.A7.4B$56.10B2.4B29.4B$57.14B31.4B$56.14B33.4B$57.13B
34.4B$57.13B35.4B$59.2B.4B3DB36.4B$62.5BD2B37.4B$61.4B3D2B38.4B$61.9B
39.4B$59.2CB2.7B39.4B$58.C.CB2.8B39.4B$58.C6.8B39.4B13.A$57.2C6.8B40.
4B10.3A$64.6B2.B2A39.4B8.A$64.7B.BA.A23.2A14.4B7.2A$65.6B4.A4.2A.A.2A
11.B2AB14.4B3.5B$65.6B4.2A2.A2.2A.A2.A.2A.A4.3B16.4B2.3B$65.6B7.A.A4.
A.2A.A.2A5.B.B16.9B7.2A$64.8B7.A.5A2.B8.5B17.8B8.A$63.8B10.A4.AB2A6.
6B18.10B3.B.A.2A$32.2A29.9B8.A2.BA.A.2A4.8B18.4BD2B2A2B.B3A2.A$32.A.A
28.9B7.A.2BA.A2.13B19.2BDBD2B2A3BAB2.2A$34.A4.2A21.10B7.2A.2BA5.13B
17.3B2D7B4A$30.4A.2A2.A2.A19.3B2A5B11.3B3.15B14.2AB.7B3.2B.A$30.A2.A.
A.A.A.2A18.4B2A5B12.4B.15B13.A.AB.7B2.B3A$32.BABABA.A21.11B11.4B.17B
12.A5.4B4.A$33.B2ABA.A16.2A.A7BD4B4.29B11.2A5.4B5.5A$34.2B.BA17.A.2A
2.4B3DB5.4BC11B2A2BD10B16.4B10.A$33.3B27.2B2D2BD3B2.4B3C9B2A2BD11B2.
2A10.4B9.A$24.2A6.4B28.10B.4BCBC13BD6B3.B2A2.A9.4B10.2A$25.A6.B2A3B
26.17BC19B4.A.B2A9.4B$25.A.AB3.B2A3B25.22B2.2B2.B3.6B5.A11.4B$26.2AB.
10B23.19B14.6B5.3A7.4B$28.13B18.2B.20B12.9B7.A5.4B$28.5B3D6B16.2C24B
10.2A4.4B5.2A4.4B$28.7BD7B15.2C25B10.A5.4B4.9B$30.8B2.4B10.A4.27B6.3A
7.4B5.6B$30.6B5.4B7.3A6.27B4.A10.4B2.8B$29.9B4.4B5.A10.25B2C15.7BD7B$
28.4B4.2A5.4B4.2A10.24B2C16.6B3D5B$27.4B5.A7.9B12.20B.2B18.13B$26.4B
7.3A5.6B14.19B23.10B.B2A$25.4B11.A5.6B3.B2.2B2.22B25.3B2AB3.BA.A$24.
4B9.2AB.A4.19BC17B26.3B2AB6.A$11.2A10.4B9.A2.2AB3.6BD13BCBC4B.10B28.
4B6.2A$12.A9.4B10.2A2.11BD2B2A9B3C4B2.3BD2B2D2B27.3B$10.A10.4B16.10BD
2B2A11BC4B5.B3D4B2.2A.A17.AB.2B$10.5A5.4B5.2A11.29B4.4BD7BA.2A16.A.AB
2AB$15.A4.4B5.A12.17B.4B11.11B21.A.ABABAB$12.3AB2.7B.BA.A13.15B.4B12.
5B2A4B18.2A.A.A.A.A2.A$11.A.2B3.7B.B2A14.15B3.3B11.5B2A3B19.A2.A2.2A.
4A$11.4A7B2D3B17.13B5.A2B.2A7.10B21.2A4.A$9.2A2.BA3B2A2BDBD2B19.13B2.
A.A2B.A7.9B28.A.A$8.A2.3AB.2B2A2BD4B18.8B4.2A.A.AB2.A8.9B29.2A$8.2A.A
.B3.10B18.6B6.2ABA4.A10.8B$11.A8.8B17.5B8.B2.5A.A7.8B$11.2A7.9B16.B.B
5.2A.A.2A.A4.A.A7.6B$21.3B2.4B16.3B4.A.2A.A2.A.2A2.A2.2A4.6B$19.5B3.
4B14.B2AB11.2A.A.2A4.A4.6B$19.2A7.4B14.2A23.A.AB.7B$20.A8.4B39.2AB2.
6B$17.3A10.4B40.8B6.2C$17.A13.4B39.8B6.C$32.4B39.8B2.BC.C$33.4B39.7B
2.B2C$34.4B39.9B$35.4B38.2B3D4B$36.4B37.2BD5B$37.4B36.B3D4B.2B$38.4B
35.13B$39.4B34.13B$40.4B33.14B$41.4B31.14B$42.4B29.4B2.10B$43.4B7.A
19.4B2.11B.B$44.4B6.3A16.4B.14B2A$45.4B8.A8.26B2A$46.4B6.2A7.4BD23B$
47.4B5.5B.B2.4B3D21B$48.4B6.6B.4BDBD17B$49.4B4.14BD15B2A$50.25B2.2B2.
6B2AB.B$51.21B11.8B2A$50.20B15.4B.B2A$49.20B17.2B3.B$48.2A19B16.2B$
48.2A11B2.8B13.B2AB$49.11B5.B3.2A14.2A$51.11B7.A$51.4B5.2A8.3A$60.A
11.A$61.3A$63.A!

So now, you cannot remove only 1 signal (unless you plan to destroy the whole thing with a "blank signal" I guess), so I just wonder if people are comfortable calling this a "signal circuit". I personally would just call it a circuit or an oscillator.

[confocaloid]

I wanted to share a concrete illustrative example.

...
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. ...

... (Of course if you don't view it as a black box it obviously contains bit-carrying signals internally, but it's not inconceivable there could be a reflector with these properties that isn't just a self-sabotaging independent reflector, and that wouldn't have bit-carrying signals anywhere. This is why I like viewing things as black boxes.)

...
With dependent reflectors, things are slightly more complicated. I would view a single reflector as a signal-processing device (with the gliders being the signal in the same local sense as with the Snark example), albeit with very low information entropy. It can essentially either continue operating, or fail to recieve an input signal and destruct. Essentially, the only information it can process is 'when to destruct', and in closed oscillator loop, the answer might be 'never', but that will be, in essence, 'locally never'. An information in the form of almost any glider from outside will destroy the loop, in one way or another.
...

Here's a large engineered period-79 dependent reflector.

As long as this device is used for what it is advertised to do, the only things one needs to know are "physical":
- Where and when input gliders must appear? (timing mod 79)
- Where and when output gliders will appear? (timing mod 79)
- Does the device fit into the available space?

Code: Select all

x = 747, y = 828, rule = B3/S23
214b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o
2b2o2b2o2b2o2b2o$214b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b
2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o3$214b2o78b2o$214b2o78b2o$229b2o$
229b2o$214b2o78b2o$214b2o78b2o3$214b2o20bo57b2o$214b2o10b2o8b3o8b2o45b
2o$226b2o4b2o5bo7bo$232bobo3b2o5bobo$214b2o18bo10b2o47b2o$214b2o18b2o
58b2o3$214b2o7b2o69b2o$214b2o7bo70b2o$220b2obo$220bo2bob2o$214b2o6b2ob
o68b2o$214b2o9bo13b3o52b2o$225b2o11bo3bo$237bo5bo7b2o$214b2o21bo5bo6b
2o42b2o$214b2o21bo5bo8bo41b2o$238bo3bo4b2o24b2o$239b3o5bobo24bo$214b2o
33bo22bo21b2o$214b2o21b2o10b2o21b5o17b2o$238bo37bo$235b3o32b4o$214b2o
19bo34bo2bo8b2o10b2o$214b2o66b2o10b2o2$289b2o$214b2o73b2o3b2o$214b2o
78b2o$281b2o$282bo$217bobo59b3o12b2o$217b2o41b2o17bo14b2o$218bo42bo21b
2o$261bobo3b3o13bo$262b2o2bo3bo10bobo10b2o$265bo5bo9b2o11b2o$265bo5bo$
265bo5bo$266bo3bo23b2o$267b3o24b2o2$280b2o$214b2o64bo13b2o$214b2o49b2o
14b3o10b2o$266bo16bo$263b3o$214b2o47bo4bo25b2o$214b2o52b2o24b2o$267bob
o$198bo$197bo16b2o78b2o$197b3o14b2o78b2o3$214b2o78b2o$214b2o78b2o3$
214b2o78b2o$214b2o78b2o3$214b2o78b2o$214b2o78b2o3$214b2o32b2o44b2o$
214b2o31bobo44b2o$249bo$179bo$177b2o35b2o78b2o$178b2o34b2o78b2o3$214b
2o2b2o2b2o2b2o2b2o2b2o14b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b
2o$214b2o2b2o2b2o2b2o2b2o2b2o14b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o2b2o
2b2o2b2o11$228b2o$229b2o$228bo$158bo$158bobo$158b2o15$208b3o$210bo$
209bo$138bobo$138b2o$139bo14$189bo$189b2o$188bobo$119bo$118bo$118b3o
12$76bo$76b3o$79bo$78b2o89b2o$168bobo$170bo$100bo$98b2o$99b2o15$149b2o
$150b2o$149bo$79bo47b2o$79bobo45bo$79b2o48bo$125b5o$125bo$128b4o$118b
2o8bo2bo$118b2o5$119b2o$119bo$120b3o$122bo17b2o$117b2o21bo$118bo13b3o
3bobo$118bobo10bo3bo2b2o$119b2o9bo5bo$59bobo68bo5bo$59b2o69bo5bo$60bo
4bo65bo3bo$63b3o66b3o$45bo16bo$45b3o14b2o56b2o$48bo72bo$47b2o69b3o14b
2o$118bo16bo$59b3o74b3o$58bo3bo70bo4bo$57bo5bo68b2o$57bo5bo68bobo$46b
2o9bo5bo$45bobo10bo3bo2b2o$45bo13b3o3bobo$44b2o21bo$49bo17b2o$47b3o$
46bo$40b2o4b2o$40b2o2$37b2o$38bo$35b3o7b2o$35bo9b2o8bo2bo$55b4o$52bo$
52b5o$56bo95b2o$54bo97bobo$54b2o96bo$76bo$77b2o$76b2o15$172b2o$171b2o$
173bo$97bo$95bobo$96b2o15$191b3o$191bo$192bo$115bobo$116b2o$111bo4bo$
111b3o$114bo16bo$113b2o14b3o$128bo$128b2o2$115b3o$114bo3bo$113bo5bo$
113bo5bo$113bo5bo9b2o$110b2o2bo3bo10bobo$109bobo3b3o13bo$109bo21b2o79b
o$108b2o17bo83b2o$127b3o81bobo$130bo$129b2o4b2o$135b2o4$130b2o$118bo2b
o8b2o$118b4o$124bo$120b5o$120bo$122bo$121b2o$100bo$98b2o$99b2o$231b2o$
231bobo$231bo15$79bo$79bobo$79b2o$251b2o$250b2o$252bo15$59bobo$59b2o$
60bo$270b3o$270bo$271bo14$40bo$39bo$39b3o$291bo$290b2o$290bobo15$21bo$
19b2o$20b2o$310b2o$310bobo$310bo15$o$obo$2o$330b2o$329b2o$331bo18$349b
3o$349bo$350bo17$370bo$369b2o$369bobo18$389b2o$389bobo$389bo18$409b2o$
408b2o$410bo18$428b3o$428bo$429bo17$449bo$448b2o$448bobo38$488b2o$487b
2o$489bo18$507b3o$507bo$508bo17$528bo$527b2o$527bobo18$547b2o$547bobo$
547bo18$567b2o$566b2o$568bo18$586b3o$586bo$587bo17$607bo$606b2o$606bob
o18$626b2o$626bobo$626bo18$646b2o$645b2o$647bo18$665b3o$665bo$666bo17$
686bo$685b2o$685bobo18$705b2o$705bobo$705bo18$725b2o$724b2o$726bo18$
744b3o$744bo$745bo!

As long as the incoming p79 glider stream is regular without holes / mistimed gliders, here's what happens:

- The incoming glider stream is redirected to a new glider stream, which enters the region surrounded by the blocks.
- The stream is further redirected twice. After that, gliders leave the part surrounded by blocks.
- The stream is further redirected twice, before leaving the engineered p79 dependent reflector.

However, when there is a hole in the incoming glider stream, the failure of the engineered p79 dependent reflector proceeds in a different way:

- A stray glider is created during destruction of the first elementary reflector in the chain.
- That stray glider creates a hole in the stream returning from the region surrounded by blocks.
- That hole reaches the penultimate elementary reflector in the chain, which fails.
- The stream going to the last elementary reflector in the chain ends, and the last elementary reflector fails.

Now, the seemingly paradoxical part is that

- As long as you use the device as a dependent reflector, you can increase or decrease the "total path length" (assuming that's meaningful), by moving the "trombone slide" part surrounded by blocks by 79 diagonals and adding/removing intermediate gliders as needed.
- However, when the device fails, it always fails the same way with the same timing, regardless of any "trombone slide adjustment".

In other words, it is confusing (if not outright contradictory) to view individual gliders as "signals" in this example.

To me, this example indicates that period-N dependent reflectors shouldn't be described as "signal devices". They are merely devices capable of redirecting uninterrupted period-N glider streams. The specific way of failing is not (and should not be) part of the definition of dependent reflector.

[Scorbie]

Note: My stance is essentially Pavgran's post. Pavgran may or may not agree with my post though.

Now, the seemingly paradoxical part is that

- As long as you use the device as a dependent reflector, you can increase or decrease the "total path length" (assuming that's meaningful), by moving the "trombone slide" part surrounded by blocks by 79 diagonals and adding/removing intermediate gliders as needed.
- However, when the device fails, it always fails the same way with the same timing, regardless of any "trombone slide adjustment".

In other words, it is confusing (if not outright contradictory) to view individual gliders as "signals" in this example.

I think information theory defines these as signals (precisely, each bit in a unary encoding has nonzero entropy) so I disagree.

Personally I don't think "total path length" is meaningful in general.(*) I think the reflectors with different trombone slides should be considered as different reflectors (Edit: I wrongly thought that they behave differently inside a black box. In general I don't care they belong to the same class, their behavior inside a black box is important to me in the black box approach)
Edit: Wait they work the same when put inside a black box don't they? I don't see what's contradictory about this.

Edit:
(*) To clarify, there could be black box (works on every dependent reflector) and white box (works on only some dependent reflectors) approaches to analyzing dependent reflectors. I'm talking about the black box approach

Edit:

The specific way of failing is not (and should not be) part of the definition of dependent reflector

I had thought the same but I heard that in signal circuitry timers are used and this has the ability to be used as a timer, so I disagree.
Edit: The specific method of failing, I don't care. The fact that it fails at some adjustable time, I do care.

[confocaloid]

The reason why I believe it is confusing (if not outright contradictory) to view individual gliders as "signals" in my example above, is that the "black box" approach conflicts with the "white box" approach.
My example contraption is (basically) an engineered p79 dependent reflector that is made out of five elementary p79 dependent reflectors and some additions.
With the "black box" approach, you see when the output stream ends (relative to the arrival of a hole in the input stream), but you cannot distinguish between "trombone slide adjustments".
With the "white box" approach, you see how it works internally, and you can see that during normal operation, the gliders move back and forth a significant distance, and that distance is different for different "trombone slide adjustments".
You can also add a block to cleanly absorb that "stray glider" from the first elementary reflector in the chain, and that minor modification will have the effect of making the family of "trombone slide adjusted" variants into different dependent reflectors that can be distinguished even in the "black box" approach.

I think information theory defines these as signals (precisely, each bit in a unary encoding has nonzero entropy) so I disagree.

They could be viewed as signals that way in the "black box" approach, or alternatively they could be viewed as signals in the "white box" approach, but in the "white box" approach, the gliders going in/out the five internal elementary dependent reflectors form segments of streams which don't correspond with the movement of the "black box" unary encoded signal. There lies potential and actual confusion.

Either way, I believe one shouldn't have to learn signals to understand reflectors. Reflectors are much simpler than signals.

Edit:

The reason why the two approaches look contradictory is because you're introducing gliders while joining them. Adding gliders have a nonzero effect on the reflector.(just as any other signal circuitry) You can think of that part as a degenerate kind of 0-degree dependent reflector, or you can redefine the boundary of a dependent reflector to include the gliders so that you don't require gliders to join them. Either way the whole thing works like a pipeline and the whole is equal to the sum of parts.

Sorry, I don't follow.

For the above example, I meant that the "signals" apparently move differently depending on the approach.
- If you see the entire contraption as a "black box" (i.e. the five internal reflectors are invisible), then you cannot distinguish between "trombone slide adjustments" by feeding a hole instead of a glider -- the whole family appears to fail in the same way with the same timing.
- If you see the entire contraption as a "white box" (but presumably the five internal reflectors are still "black boxes" because they're elementary p79 reactions), then you can see that the Euclidean distance traveled by some of the gliders is different for different "trombone slide adjustments" during normal operation. You can see that the failure propagates differently along a shorter route -- but you could prevent that shorter propagation by adding an extra block to catch the "stray glider" from the failing first elementary dependent reflector -- but that block addition is merely a stator variant of the whole contraption, and shouldn't be considered as something that changes the nature of the whole engineered p79 dependent reflector.

[Scorbie]

The reason why the two approaches look contradictory is because you're introducing gliders while joining them. Adding gliders have a nonzero effect on the reflector.(just as any other signal circuitry) You can think of that part as a degenerate kind of 0-degree dependent reflector, or you can redefine the boundary of a dependent reflector to include the gliders so that you don't require gliders to join them. Either way the whole thing works like a pipeline and the whole is equal to the sum of parts.

[Scorbie]

(consecutive post because confocaloid replied with an edit)
First things first I need some time because I had a precise mathematical idea of a "black box" and "white box" terminology when I thought about this (Edit: I had thought about this before your post) and I need some time to analyze your statement based on the idea.

but that block addition is merely a stator variant of the whole contraption, and shouldn't be considered as something that changes the nature of the whole engineered p79 dependent reflector.

For a contraption that can fail I don't think the term "stator variant" is meaningfully defined.

[confocaloid]

The specific way of failing is not (and should not be) part of the definition of dependent reflector

I had thought the same but I heard that in signal circuitry timers are used and this has the ability to be used as a timer, so I disagree.

but that block addition is merely a stator variant of the whole contraption, and shouldn't be considered as something that changes the nature of the whole engineered p79 dependent reflector.

For a contraption that can fail I don't think the term "stator variant" is meaningfully defined.

If someone is using some specific dependent reflector [dependent reflector variant] as some other device, then they are not using it as a dependent reflector anymore -- it becomes a more complicated device, a device that takes more effort to explain/understand.

My stance on this is that it would be misleading/confusing, if someone decided to
(a) use a periodic dependent reflector (like the one below) in some nontrivial way in logic circuitry, but
(b) still continue to describe/advertise it as a dependent reflector.

As soon as one relies on some specific way of failure, describing the device as "dependent reflector" stops being honest advertising.
As long as a dependent reflector is used for what it is, stator variants (like adding a block or replacing a fishhook by an elevener) do not change the reflector -- it is still compatible with the same input/output periodic glider stream / glider track placement.

Code: Select all

x = 66, y = 96, rule = B3/S23
65bo$63b2o$64b2o18$44bo$44bobo$44b2o18$24bobo$24b2o$25bo4bo$28b3o$10bo
16bo$10b3o14b2o$13bo$12b2o2$24b3o$23bo3bo$22bo5bo$22bo5bo$11b2o9bo5bo$
10bobo10bo3bo2b2o$10bo13b3o3bobo$9b2o21bo$14bo17b2o$12b3o$11bo$5b2o4b
2o$5b2o2$2b2o$3bo$3o7b2o$o9b2o8bo2bo$20b4o$17bo$17b5o$21bo$19bo$19b2o$
41bo$42b2o$41b2o18$62bo$60bobo$61b2o!