Dependent Reflector Terminology and Analysis

[dvgrn]

By definition, the p43 Snark loop is a relay. It is an oscillator in which gliders travel in a loop. So it is an example where a relay has multiple spaceships.

Here's some more background on the term "relay". In 1995 when Alan Hensel was compiling the glossary of Life terms that would later become the Life Lexicon, David Bell provided this definition

relay
An oscillator in which gliders or spaceships travel in a loop.

-- but then second-guessed himself a few weeks later:

relay
I said a relay was one which spaceships travel in a loop. But is this
true? It is certainly true about gliders traveling back and forth between
two oscillators. But how about a four sided path??? No one has challenged
this.

I haven't found any responses, so this has been kind of an open question for almost thirty years... and meanwhile, the original off-the-cuff definition got copied into the Life Lexicon, and then into the LifeWiki.

I've been able to find a couple of uses of the term "relay" to mean something other than one spaceship traveling back and forth -- where there was an actual loop where the backward and forward paths don't overlap each other. But the great majority of the early uses, across several different rules, are just one spaceship traveling back and forth.

Maybe partly as a result of this ambiguity, "relay" seems to have gotten relatively little use in the last decade or so, as the original users of the word have become less active.

[Jormungant]

That second guessing bit seems to suggest what would boil down to be the difference between "shuttle" and "relay". I mean, as it stands, I think the "all shuttles are relays" is correct, but for the fact some objects are not spaceships in there. I am not sure why we need a object class that worries about the symmetry class "back-and-forth" or "rotation symmetry", I am pretty sure many "relay" are not pure back and forth anyways, and moves in 4 directions:

Code: Select all

x = 47, y = 16, rule = B3/S23
4$20b3o$5bo2bo4bo2bo3bo$3b3o2b6o2b3o2bo$5bo2bo4bo2bo2$30bo2bo4bo2bo$
28b3o2b6o2b3o$30bo2bo4bo2bo!

that glider is not on the same lane on the return, so it is not 100% "back and forth", and has "4 sides" in a sense. I mean that's seems ambiguous, so I would be strange to constrain the notion of relay in such a manner... I mean, what about this, 2 sides or 4 sides?

Code: Select all

x = 86, y = 77, rule = LifeHistory
5$74.2B$64.2B6.6B$63.6B.2A6B2A$62.6BA4BA2BA4BA$61.7BA4BA2BA4BAB$60.7B
.A4BA2BA4BA$59.7B4.2A6B2A$58.7B7.6B$57.7B10.2B$56.7B$55.7B$54.7B6.2A$
53.7B7.A$53.9B2.BA.A$52.11B.B2A$52.9BA3B7.A$52.8BABA2B5.3A$51.9BABA2B
4.A$50.4B.B2.3BA3B4.A.2A$49.4B6.7B2.2A.A.A$48.4B8.5B2.AB2.B2.A$47.4B
9.6B.B7A$46.4B11.10B$45.4B8.11B.B3A$44.4B8.14BAB.A$23.2B18.4B8.3B2A9B
3.2A$21.6B15.4B9.2BA2BA8B$19.10B12.D3B9.4B2A9B$17.BA2BA4BA2BAB9.BDBD
9.14B.B2A$16.3A2B6A2B3A7.2B2D9.15B.BA.A$17.BA2BA4BA2BAB7.4B9.4B.B2.8B
4.A$19.10B8.4B9.4B6.6B5.2A$21.6B9.4B9.4B7.5B$23.3B9.4B9.4B9.3B$22.5B
7.4B9.4B9.6B$14.2A5.6B6.4B9.4B8.10B$15.A4.8B2.B.4B9.4B7.BA2BA4BA2BAB$
15.A.AB.15B9.2A2B7.3A2B6A2B3A$16.2AB.14B9.ABAB9.BA2BA4BA2BAB$18.9B2A
4B9.3BA12.10B$18.8BA2BA2B9.4B15.6B$13.2A3.9B2A3B8.4B18.2B$13.A.BA14B
8.4B$14.3AB.11B8.4B$16.10B11.4B$12.7AB.6B9.4B$12.A2.B2.BA2.5B8.4B$13.
A.A.2A2.7B6.4B$14.2A.A4.3BA3B2.B.4B$17.A4.2BABA9B$14.3A5.2BABA8B$14.A
7.3BA9B$20.2AB.11B$19.A.AB2.9B$19.A7.7B$18.2A6.7B$25.7B$24.7B$11.2B
10.7B$9.6B7.7B$7.2A6B2A4.7B$5.A4BA2BA4BA.7B$4.BA4BA2BA4BA7B$5.A4BA2BA
4BA6B$7.2A6B2A.6B$9.6B6.2B$11.2B!

[dvgrn]

that glider is not on the same lane on the return, so it is not 100% "back and forth", and has "4 sides" in a sense. I mean that's seems ambiguous, so I would be strange to constrain the notion of relay in such a manner... I mean, what about this, 2 sides or 4 sides?

I absolutely agree that either adding new constraints on "relay", or trying to enforce constraints that may have existed in the past, would be a little strange at this point.

My own solution will be to avoid using "relay" at all, since it doesn't seem to be a term that's needed any more.

For these dependent-reflector-based oscillators that are suddenly showing up in incredible profusion, my preference would definitely be to use an uncontroversial standard descriptive name for all of them, even if it's a little longer. So far there haven't been any suggestions that I like better than "p{N} dependent reflector loop".

(I mentioned "p{N} dependent glider loop" a few posts up, to save a syllable and get the idea across that it's gliders that are doing the support work. But "dependent" doesn't really belong with "glider" -- it needs to be part of "dependent reflector" to keep it from being confusing -- so I don't think that "dependent glider loop" would really work as a standard name.)

-- All of that said, of course if enough people start using some new shorthand term for these dependent reflector glider loops, I'll happily go along and call them all "drefgloops" or whatever. Just want to be careful not to further confuse any existing terminology that is already fairly confusing.

[confocaloid]

For these dependent-reflector-based oscillators that are suddenly showing up in incredible profusion, my preference would definitely be to use an uncontroversial standard descriptive name for all of them, even if it's a little longer. So far there haven't been any suggestions that I like better than "p{N} dependent reflector loop".

Well, this term is already in use outside this thread. No objections here, as long as everyone clearly understands what counts/what does not count as a concrete example.
I agree that many of these things look like loops visually.

My own solution will be to avoid using "relay" at all, since it doesn't seem to be a term that's needed any more.

I think the word "relay" is helpful (outside of this dependent-reflector-related discussion). The word doesn't require a specific geometric movement/symmetry. As far as I understand from the above, there were no actual symmetry constraints in the past (so nothing to enforce).

I mean, if a glider circulates through four Snarks, then each of those Snarks actually relays that glider (as in: "I was told the news first and then I relayed it to the others.")
The intuition here is that the glider is news to the Snark.
Nothing changes if there are multiple active objects (in the p43 Snark loop, each glider is individually relayed through).
Nothing changes if the geometric shape of the envelope changes (it could be back-and-forth, or four-sided, or more complicated). The geometry/symmetry of a relay oscillator is not important for understanding why exactly the oscillator counts as a relay.

On the other hand, in a "dependent reflector loop" oscillator, a dependent reflector doesn't relay anything. The input glider must arrive in time every N ticks, or else the dependent reflector fails. Again intuitively, the input glider is not news to the dependent reflector in a "dependent reflector loop" oscillator. So nothing is relayed there, and nothing travels in the loop, and there's no Traveling Something.
(So the word "loop" describes the geometric shape of many of these oscillators -- but it isn't a good description of the overall movement of gliders.)

I think this intuition is helpful for understanding, and this distinction quickly shows as soon as one attempts to understand how these oscillators work.

[confocaloid]

So far there haven't been any suggestions that I like better than "p{N} dependent reflector loop".

(So the word "loop" describes the geometric shape of many of these oscillators -- but it isn't a good description of the overall movement of gliders.)

FWIW, if it was up to me to invent a new term, then maybe I would choose "closed p(N) dependent reflector chain".

Intuitively, there are several dependent reflectors, and there are segments of spaceship streams. These are linked together into a single structure, forming a closed polygonal chain.
There would be no sneaky implications about (existence/non-existence) of trackable traveling entities. The term would describe the structure.

But since it seems like everyone is already using "dependent reflector loop" to describe newly discovered oscillators of this type, that's going to be the documented term, as long as LifeWiki describes existing usage instead of prescribing it.

[dvgrn]

I think the word "relay" is helpful (outside of this dependent-reflector-related discussion). The word doesn't require a specific geometric movement/symmetry. As far as I understand from the above, there were no actual symmetry constraints in the past (so nothing to enforce).

I mean, if a glider circulates through four Snarks, then each of those Snarks actually relays that glider (as in: "I was told the news first and then I relayed it to the others.")
The intuition here is that the glider is news to the Snark.
...
On the other hand, in a "dependent reflector loop" oscillator, a dependent reflector doesn't relay anything. The input glider must arrive in time every N ticks, or else the dependent reflector fails. Again intuitively, the input glider is not news to the dependent reflector in a "dependent reflector loop" oscillator. So nothing is relayed there, and nothing travels in the loop, and there's no Traveling Something.

Sadly, this intuition doesn't work particularly well for me. It's not terrible, but it isn't something I'd use... partly because I still completely disagree with this idea that you keep stating as if it's true, that "a dependent reflector doesn't relay anything".

Just because a continuous period-N input stream is required for a dependent reflector to work, does not in the least imply to me that no information is passing through the reflector. To the contrary, information very clearly seems to be getting relayed through a dependent reflector. The continued presence of the input stream is reliably reported to the output side of the reflector, every N ticks. That is perfectly valid information, moving through the Life universe.

"Information moving through the Life universe" is the LifeWiki definition of "signal". That is why it is also still making no sense to me to artificially avoid the term "signal circuitry" for dependent reflectors, but allow it for independent reflectors. When fed with the same continuous input stream, dependent and independent period-N reflectors can be arranged to produce identical output signals. So "signal" vs. "not-signal" does not seem like a helpful distinction here, where dependent reflectors are concerned.

"Dependent" vs. "independent" is the helpful distinction, but we already have that information, right there in the name. There doesn't seem to be any need to try to re-define "signal" at this late date, to somehow mean "free choice of presence or absence".

Dean Hickerson and others have way too many uses of "signal" starting from a quarter century or more ago, that don't line up at all with that attempted re-definition. Low-period signal injectors don't produce "signals" according to your definition, because there's no choice about whether to continue sending them to keep the whole mechanism functioning.

But signal injectors do produce signals according to my definition -- and given the name "signal injector", that doesn't surprise me. They're sending information, reporting to the signal sink every N ticks that the signal injector is still sending signals.

[confocaloid]

On the other hand, in a "dependent reflector loop" oscillator, a dependent reflector doesn't relay anything.

Sadly, this intuition doesn't work particularly well for me. It's not terrible, but it isn't something I'd use... partly because I still completely disagree with this idea that you keep stating as if it's true, that "a dependent reflector doesn't relay anything".

I think "a dependent reflector doesn't relay anything" does not accurately reflect the idea I'm trying to convey.
I wrote "On the other hand, in a "dependent reflector loop" oscillator, a dependent reflector doesn't relay anything."

So it is already known that the pattern under discussion is a "dependent reflector loop" oscillator. In such an oscillator, the dependent reflectors only link together oscillating streams that are needed to support the active reactions in the dependent reflectors. In these oscillators, nothing of interest travels all the way around. In these oscillators, there is no "moving thing" to be tracked all the way around the closed chain.

I think the whole discussion of "traveling information" is confusingly redundant in the discussion of "dependent reflector loop" oscillators.
(edited to attempt to clarify, last edit 2023-07-14 05:46 UTC)

[dvgrn]

In these oscillators, nothing of interest travels all the way around.

Nothing of interest to you travels all the way around, perhaps. Could you maybe try to qualify your statements a little more clearly? When there's clear disagreement on a point, it might be helpful if everyone is careful to state their opinions as opinions rather than as facts. I apologize if I've also failed to do that consistently, but I'm definitely making the attempt... sometimes, anyway!

In these oscillators, there is no "moving thing" to be tracked all the way around the closed chain.

The chain of causality that travels all the way around dependent reflector loops is very much of interest to me. My understanding of the definition of "signal" is that a moving pattern with this kind of clear chain of causality does in fact count as a (dependent, period N) signal stream. These are not signal mechanisms that are capable of carrying binary information, no -- but information is still traveling from one place in the Life universe to another.

(I do remember the objection that the animation in that link is not two-state Life so it doesn't count. But experiments with two-state Life show the same chain of causality, so the objection doesn't seem relevant to me.)

An eater of an input stream, disconnected from a nearby gun producing an output stream, doesn't carry any information from one place to another. But a dependent reflector does maintain a causal connection between input and output. I still think it should continue to be okay to say that it carries signals, and therefore it counts as signal circuitry.

Is This Important? Why Is This Important?
I think that the LifeWiki's specific definition of "signal" is subtly important in several other places besides the "dependent reflector" article -- not just on the LifeWiki, but also in the Life textbook and elsewhere. It's subtle because the great majority of signals are perfectly capable of carrying binary information. But the definition of "signal" does not require this.

Here's Gabriel Nivasch's article on lightspeed signals. Almost all of the patterns shown there, including most of the periodic signal injectors, are producing signals that could theoretically carry binary data. But right in the middle, there's a short section on signal "sources and sinks" for the more general type of signal:

It is also possible to build oscillators that emit signals at some specific period. The pattern then needs another oscillator of the same period at the other end, to absorb the signals.

Discussions on these topics have used "signal" in this way, without thinking twice about it -- not often, of course, because these kinds of "dependent signals" are quite rare -- but it does happen:

>>I found a source for this period 6 lightspeed signal. The sink has
>>been known for some time, but I believe the source had been unknown.
>>I have attached a .rle file of the entire oscillator.
>>Sincerely,
>>Josh Ball

Code: Select all

x = 90, y = 16, rule = B3/S23
20boboo53boo$3bobobboo10boobbobooboo31boo3boo7bobbo$bboboobbobo3boo7bo
3bobo33bobbobbo7boobo$bbo3boobbobbobobb5oboo5bobbobbobbobbobbobbobbobb
obbobbobb3obobboob3o3bo8bo$oob3obb3obobbobobbo3bob38o3boboobobobbobob
oo4b3o$bobobboo4boboboo4boo40boobobobbo3boobobobobbo$obb3obboobobboo6b
oo5b4obb4obb4obb4obb4obb4o4bobb3o7boobobboo$boobboboo3boo7boobb3o4boo
4boo4boo4boo4boo4bo3bo6bobo4b3o$3boo3boobo9boboboobb4obb4obb4obb4obb4o
bb4o3boboboob4o3boboboo$3bo3bo4boo4boboobo40b3obobobo7bobo3bo$bobo4b3o
bo4bo5bob39o3bobo4b5obobb4o$boo7bobobb3oboboboo6bobbobbobbobbobbobbobb
obbobbobbo3bo3bob3o5bobobo$12bobo3boboo4b3o34boob3obobobbobbobbobbo$
11booboobbo3boboo3bo34bobobbo4boobboobboo$17boo3boobobboo34bobbo$65boo!

I updated my page on lightspeed signals and added Josh's signal source. I
also added Hartmut's 2c/3 against-the-grain signals.

Regards,
Gabriel

If some LifeWiki editors now suddenly start thinking that support for encoding binary data is a make-or-break criterion for the presence of signals, then it seems like we'd suddenly need footnotes here and there saying that "these things that have traditionally been called 'signals' aren't really signals". In other words, I think we'd end up with a big confusing mess.

I think the whole discussion of "traveling information" is confusingly redundant in the discussion of "dependent reflector loop" oscillators.

It seems highly relevant to me. The basic problem we're trying to solve is whether there's a justification for removing uses of the word "signal" from the dependent reflector article -- or whether in fact the underlying confusion is with conflicting interpretations of the word "signal" -- in which case maybe it's the signal article that actually needs a clarification.

If it's the latter, then the question of what that clarification should be ... is definitely not settled yet!

[confocaloid]

The chain of causality that travels all the way around dependent reflector loops is very much of interest to me. (...)
(I do remember the objection that the animation in that link is not two-state Life so it doesn't count. But experiments with two-state Life show the same chain of causality, so the objection doesn't seem relevant to me.)

When I convert the linked pattern to plain two-state B3/S23 and run it, I only see a p31 oscillator. I do not see anything "moving all the way around" the working oscillator.
Can you please re-state your argument into another form, so that it can be understood within plain two-state B3/S23 (i.e. without using non-Life "visual effects" implemented as multi-state ruletables)?

Here's Gabriel Nivasch's article on lightspeed signals. ...

The linked article discusses objects (perturbations) that move through a non-empty background pattern. Such objects (perturbations) are traditionally called "signals". In the book Conway's Game of Life: Mathematics and Construction, the corresponding section is 4.5.1 "Wires and Signals". Related terms are defined there.

The other common meaning of "signal" distinguishes between the moving information and the object (or set of objects) that carries that information.
For example, a single signal can be carried by a tandem glider. There are two gliders, but there is only one signal.

If some LifeWiki editors now suddenly start thinking that support for encoding binary data is a make-or-break criterion for the presence of signals, then it seems like we'd suddenly need footnotes here and there saying that "these things that have traditionally been called 'signals' aren't really signals". In other words, I think we'd end up with a big confusing mess.

I think such footnotes would be much more helpful, and much less confusing, in comparison to any attempts to invent a single "one-size-fits-all" definition of "signal" that somehow manages to combine two different meanings.

In this case, there is existing ambiguity -- when taken out of context, the word "signal" can mean different things:
(1) An object (perturbation, deformation) that moves through a non-empty background pattern. This is the meaning discussed in Wire.
(2) Information that moves through the Life universe, and is carried by some object(s). This is the meaning discussed in Signal.
Depending on context, both meanings are useful.
Normally, the context allows to distinguish which of the two meanings is intended.

The basic problem we're trying to solve is whether there's a justification for removing uses of the word "signal" from the dependent reflector article -- or whether in fact the underlying confusion is with conflicting interpretations of the word "signal" -- in which case maybe it's the signal article that actually needs a clarification.

If it's the latter, then the question of what that clarification should be ... is definitely not settled yet!

I think it is both.

The justification for avoiding the word "signal" in Dependent reflector is that it is unnecessary to refer to either of existing meanings of "signal" to explain dependent reflectors. Uses of the word in that article increase confusion for no added benefit -- the explanation is simpler without them. The simplest/least confusing working explanation is preferred.

The underlying confusion is with conflicting meanings of the ambiguous word "signal". So the page Signal should really be a disambiguation page.

[dvgrn]

The chain of causality that travels all the way around dependent reflector loops is very much of interest to me. (...)

When I convert the linked pattern to plain two-state B3/S23 and run it, I only see a p31 oscillator. I do not see anything "moving all the way around" the working oscillator.
Can you please re-state your argument into another form, so that it can be understood within plain two-state B3/S23 (i.e. without using non-Life "visual effects" implemented as multi-state ruletables)?

Sure. I'll retain LifeSuper just for convenience in labeling:

Code: Select all

x = 78, y = 76, rule = LifeSuper
33.3R.R.R.3R$35.R.R.R.R.R$33.3R.3R.3R$33.R5.R.R.R$33.3R3.R.3R3$38.T$
37.T.T$36.T3.T$36.T3.T$36.5T$36.T3.T$36.T3.T$36.T3.T$36.T3.T2$46.2M$
45.2M.M5.2M$46.M.2M4.2M$47.2M2$20.2M8.M6.2G3.3M$20.2M6.3M7.2G.M3.M$
27.M9.G3.M3.M$27.2M12.M.2M7.2M$52.M.M$47.M6.M$48.M5.2M$21.2M23.3M$18.
2M.2M6.3M$20.M5.M4.M$26.M3.M31.4T$25.M36.T3.T$24.M3.M33.T3.T$24.M.2M
26.G7.T3.T4.3R.3R$13.4T8.M29.2G5.4T5.R5.R$13.T3.T36.2G6.T3.T4.3R.3R$R
.3R.3R4.T3.T44.T3.T4.R.R.R$R.R.R.R6.T3.T7.2G35.T3.T4.3R.3R$R.3R.3R4.T
3.T6.2G29.M6.4T$R.R.R.R.R4.T3.T8.G26.2M.M$R.3R.3R4.T3.T34.M3.M$13.T3.
T37.M$13.4T33.M3.M$49.M4.M5.M$49.3M6.2M.2M$32.3M23.2M$25.2M5.M$26.M6.
M$26.M.M$27.2M7.2M.M12.2M$35.M3.M3.G9.M$35.M3.M.2G7.3M6.2M$36.3M3.2G
6.M8.2M2$32.2M$25.2M4.2M.M$25.2M5.M.2M$33.2M6.3T$40.T3.T$40.T$40.T$
40.T$40.T$40.T$40.T3.T$41.3T4$38.R.3R.R.R$38.R3.R.R.R$38.R.3R.3R$38.R
.R5.R$38.R.3R3.R!

We're talking about a p31 oscillator here, so we know that it's going to keep doing the same thing -- there aren't any surprises where some part of the structure might have just gotten constructed, or some glider might have come flying in from somewhere else.

In the following, I'll be using "signal" in the same sense that I've been using "signal" whenever I've talked about signal circuitry for the last two decades or so. There shouldn't be any need to object to every individual usage. One objection to the whole post will be clear enough. I'll re-state the definition of "signal" as I'm using it, below.

Run the pattern to T=248. It ends up looking just the same as it started. Not surprising, it's an oscillator. But now let's look closely at what happend during those 248 ticks.

- At T=248, we have a glider in pink location A. It reached that location from the western dependent reflector's output signal stream, only because a previous glider in that reflector's input signal stream was present at location D at T=186. If that input glider had not been present at location D at T=186, the output glider would not be present at location A.

So there's a causal connection between the two sides of the reflector. This is why it's called a "reflector" and not a "gun": guns don't need external support, but this reflector needs an input signal stream to produce its output signal stream.

The key piece of information -- the fact that a glider was present at D at T=186 -- has been communicated across the Life universe from location D to location A in 62 ticks, through the dependent reflector mechanism. The input signal has produced an output signal.

- At T=186, we had a glider in location D that reached that location from the southern dependent reflector, only because a glider was present at location C at T=124.
- At T=124, we had a glider in location C that reached that location from the eastern dependent reflector, only because a glider was present at location B at T=62.
- At T=62, we had a glider in location B that reached that location from the northern dependent reflector, only because a glider was present at location A at T=0.

Re-stating all this starting from T=0 and moving forward, the causal chain extends all the way around the loop, from A back to A again. At the end of the cycle, the information about the glider's presence at location A at T=0 has traveled all the way around the loop in 248 ticks.

A slightly different phrasing would be that "each input signal allows each dependent reflector to function for another cycle, and to produce its next output signal".

In many of these dependent-reflector cases, it makes good intuitive sense to describe this as the input signal passing through the reflector and coming out the other side -- just the same as with any other reflector. The Life Lexicon and LifeWiki definition of "signal" only requires that information has to be moving... and information is in fact moving.

In this case, there is existing ambiguity -- when taken out of context, the word "signal" can mean different things:
(1) An object (perturbation, deformation) that moves through a non-empty background pattern. This is the meaning discussed in Wire.
(2) Information that moves through the Life universe, and is carried by some object(s). This is the meaning discussed in Signal.
Depending on context, both meanings are useful.
Normally, the context allows to distinguish which of the two meanings is intended.

I certainly agree with the last two sentences.

However, the level of ambiguity you describe seems normal to me for any useful piece of terminology. It does not imply that there's any irreconcilable conflict between the two meanings.

Most generally, my understanding is that a signal is a set of active cells that moves through a stationary background pattern. The background pattern may be empty or stable or periodic, but must not be chaotic. If part of the background is temporarily disturbed, it will return to its original configuration after the signal passes.

It continues to seem to me like a painful and confusing exercise, to try to avoid the use of my standard words for signals passing through signal circuitry, for the reason that they're "more confusing" than some alternative. I can see the potential for confusion, sure, but it seems to me that the confusion is handled by keeping a clear distinction between "dependent" and "independent" signals. The alternatives you're suggesting seem much more confusing than that, while also not lining up well with existing usage.

I can see how the definition of "signal" should maybe get a footnote to state clearly that a signal can still be a signal even if you can't encode binary information in it -- specifically because some structures such as dependent reflectors require a constant periodic stream of input signals to produce their output signals.

But I don't see a reasonable way to consistently document this claim that dependent reflectors don't count as signal circuitry, and loops of them don't contain any signals... without also having to tie ourselves in terminological knots to explain why things like the periodic signal sources and sinks quoted in my previous post still somehow do constitute "signals" after all.

Long story short, I think things remain much simpler if gliders (and *WSSes, H, B, R, pi, 2c/3 wire signals, and so on) can always be called "signals" if it's useful to do so.

[confocaloid]

We're talking about a p31 oscillator here, so we know that it's going to keep doing the same thing -- there aren't any surprises where some part of the structure might have just gotten constructed, or some glider might have come flying in from somewhere else.

In fact, we know more than that. We're talking about a p31 oscillator consisting of four period-31 dependent reflectors and four period-31 glider streams. Since the reflectors are dependent, we know that there will be no surprises where some glider is missing but the oscillator still works.
(For comparison, note that such surprises would be possible with p43 Snark loop. You could remove one glider and the oscillator would function correctly in essentially the same way.)

At T=248, we have a glider in pink location A. It reached that location from the western dependent reflector's output signal stream, only because a previous glider in that reflector's input signal stream was present at location D at T=186.

The part "only because" is an error.

Specifically, the glider in location D at T=186 is not the only cause for the glider in location A at T=248.

• The glider in location D at T=155 is a cause for the glider in location A at T=248. In other words, if the glider in location D at T=155 is missing, then the glider in location A at T=248 does not appear.
• The glider in location D at T=124 is a cause for the glider in location A at T=248. In other words, if the glider in location D at T=124 is missing, then the glider in location A at T=248 does not appear.
• The glider in location D at T=93 is a cause for the glider in location A at T=248. In other words, if the glider in location D at T=93 is missing, then the glider in location A at T=248 does not appear.
• (and so on)

In other words, as soon as you run the oscillator for sufficiently many ticks (e.g. until T=248), you will get as many "causes" for the presence of every chosen glider as desired.

In fact, the unique cause for the presence of the glider in location A is that

• we're talking about a p31 oscillator consisting of four period-31 dependent reflectors and four period-31 glider streams.

So all these gliders are part of the rotor of the oscillator. We know that there are no surprises where the rotor is touched and the oscillator is not an oscillator anymore.

The key piece of information -- the fact that a glider was present at D at T=186 -- has been communicated across the Life universe from location D to location A in 62 ticks, through the dependent reflector mechanism. The input signal has produced an output signal.

The unique cause for the presence of every single glider is the structural integrity of the oscillator. At the brginning of the argument, we already know that the pattern is an oscillator. Its structural integrity holds everywhere through the pattern.

The structural integrity of the oscillator guarantees that a glider will be present at D every 31 tick. Similarly the structural integrity of the oscillator guarantees that a glider will be present at A every 31 tick. Nothing traveled around the pattern here. There is no communication of any kind.

(Compare this with p43 Snark loop. Since individual gliders can be added or removed at will without disrupting the basic mechanism of the oscillator, it is easy to see that in the p43 Snark loop, the active objects do travel all the way around.)

The Life Lexicon and LifeWiki definition of "signal" only requires that information has to be moving... and information is in fact moving.

The information is not moving, as I explained above. The unique cause for the presence of every glider is the structural integrity of the oscillator. Hence there are no signals.

Most generally, my understanding is that a signal is a set of active cells that moves through a stationary background pattern. The background pattern may be empty or stable or periodic, but must not be chaotic. If part of the background is temporarily disturbed, it will return to its original configuration after the signal passes.

It continues to seem to me like a painful and confusing exercise, to try to avoid the use of my standard words for signals passing through signal circuitry, for the reason that they're "more confusing" than some alternative. I can see the potential for confusion, sure, but it seems to me that the confusion is handled by keeping a clear distinction between "dependent" and "independent" signals. The alternatives you're suggesting seem much more confusing than that, while also not lining up well with existing usage.

I can see how the definition of "signal" should maybe get a footnote to state clearly that a signal can still be a signal even if you can't encode binary information in it -- specifically because some structures such as dependent reflectors require a constant periodic stream of input signals to produce their output signals.

But I don't see a reasonable way to consistently document this claim that dependent reflectors don't count as signal circuitry, and loops of them don't contain any signals... without also having to tie ourselves in terminological knots to explain why things like the periodic signal sources and sinks quoted in my previous post still somehow do constitute "signals" after all.

Long story short, I think things remain much simpler if gliders (and *WSSes, H, B, R, pi, 2c/3 wire signals, and so on) can always be called "signals" if it's useful to do so.

My understanding is that the word "signal" is ambiguous, and this existing ambiguity should be documented in LifeWiki.
Mixing two different meanings of the word "signal" will only create more confusion, both short-term and long-term.

I strongly disagree with the idea to use the word "signal" to refer to any active objects, regardless of whether or not they transmit information. This does not line up with either of two meanings of the word "signal".

[dvgrn]

My understanding is that the word "signal" is ambiguous, and this existing ambiguity should be documented in LifeWiki.
Mixing two different meanings of the word "signal" will only create more confusion, both short-term and long-term.

I strongly disagree with the idea to use the word "signal" to refer to any active objects, regardless of whether or not they transmit information. This does not line up with either of two meanings of the word "signal".

Hmm. Well, the impasse continues, I suppose. We're both just making predictions about what "will only create more confusion" and therefore should be avoided, with not much evidence in either direction -- except that you don't want to use my terminology, and I don't want to use yours.

And yet this problem may be a lot smaller than we're making it out to be. It seems to me I've been using "signal" in my preferred sense, since the early 2000s when I added the "Signal-Circuitry" folder to Golly's pattern collection. Nobody has ever mentioned that they find my use of the word "signal" confusing, in the context of actual discussions about actual patterns.

Conversely, while a few people have chimed in on these various questions in this thread, most people are wisely avoiding a discussion in which we've restated our opinions in half a dozen ways without getting anywhere. Heh, I can see why it would seem like a good idea to stay away.

I'm still in agreement with Scorbie's point from a while back:

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.

It seems to me that your method of applying terminology requires far too much inspection of context. Signals may either be present or absent from any given chunk of circuitry -- even identical pieces of circuitry -- depending on the behavior of other parts of the pattern.

By contrast, with my terminology, if I see a glider moving, I can call it a "signal" if that seems like an appropriate label -- e.g., if it's getting converted from one active form to another, getting reflected (by any kind of reflector), getting duplicated, etc. I've given several examples where "signal" is handy general terminology for whatever moves through signal circuitry (and that includes moving through empty spaces between pieces of circuitry.) If I want to highlight the fact that a continuous period-N signal stream is required somewhere, then I'll say "dependent".

So far, proposals to restrict the term "signal" or use some other word for dependent loops or the things circulating in them, seem to me like a big headache with no actual benefit.

The information is not moving, as I explained above. The unique cause for the presence of every glider is the structural integrity of the oscillator. Hence there are no signals.

You're welcome to say that here as often as you want, of course, but I just don't think it's time to start editing LifeWiki articles based on those interpretations. I can see how you believe all of your statements, but I don't find them compelling in the least, in terms of a useful way to describe signal circuitry in practice.

[confocaloid]

(edited 2023-07-15 22:43 UTC)

It seems to me that your method of applying terminology requires far too much inspection of context. Signals may either be present or absent from any given chunk of circuitry -- even identical pieces of circuitry -- depending on the behavior of other parts of the pattern.

The glider stream from a periodic gun is never a signal stream, regardless of surrounding context.
The same holds for a dependent reflector. Neither the input glider stream nor the output glider stream is a signal stream.

For something like a Snark, it depends on your intentions when you're describing some engineered pattern (either a working completed device, or something partially finished).
If the Snark is, in fact, supposed to transmit information in the large pattern (e.g. by making some gliders present and some absent, or by adjusting timing of individual gliders), then it is fair play to say that those gliders carry signals.
If the Snark is, in fact, only supposed to redirect a strictly periodic glider stream with fixed timing, then it is dishonest to claim that "those gliders carry signals" (because those gliders don't carry signals).

And so on. I think this approach is much simpler and less subjective.
Sometimes a glider is a glider (and nothing more than that).

By contrast, with my terminology, if I see a glider moving, I can call it a "signal" if that seems like an appropriate label -- e.g., if it's getting converted from one active form to another, getting reflected (by any kind of reflector), getting duplicated, etc.

This approach to wiki editing is subjective, because people disagree on the exact set of cases where "signal" is an appropriate label. (See replies from several members in this discussion.)

This approach is also very confusing in the sense of being hard to understand for a reader, especially when someone is trying to understand basic concepts. From the common-sense viewpoint, it's obvious that dependent reflectors don't transmit signals. A dependent reflector is not "signal circuitry" in any helpful way.

[dvgrn]

This approach is also very confusing in the sense of being hard to understand for a reader, especially when someone is trying to understand basic concepts. From the common-sense viewpoint, it's obvious that dependent reflectors don't transmit signals. A dependent reflector is not "signal circuitry" in any helpful way.

You keep on making these confident pronouncements about what is confusing to a reader, without any additional evidence. I think by this time we've established that I don't agree with you about the relative levels of confusingness of the options we're considering. If we're the only ones talking to each other on this thread, and nobody new chimes in, we might as well just stop saying the same things over and over.

Here you've linked to a dictionary definition of "signal", not the LifeWiki's definition. You seem to be really stuck on trying to make the usage of "signal" on the LifeWiki match this definition that you've gotten from elsewhere -- i.e., the idea that all signals must be able to carry binary information, such that each signalling object can be either present or absent with no catastrophic effects on the circuitry.

It should not be surprising that the LifeWiki definition of "signal" is a bit different from the average dictionary definition, just like the definition of "circuitry" and "speed of light" and lots of other terms that have been carefully redefined for a CA context.

Your insistence on this non-LifeWiki definition of "signal" makes you repeatedly say "there are no signals" about dependent reflector loops. That looks to me like a dangerous thing to be insisting on, coming from someone who is trying to make LifeWiki edits on the subject. I think that, at the very least, each dependent reflector can be said to have input signals and output signals (the gliders). That's not "no signals".

I'm less concerned about the philosophical point of whether or not it's a reasonable shorthand to say that dependent reflectors reflect signals. I'd certainly like to be able to say that they reflect period-N signal streams. If they don't, then why are they called "reflectors"? But they're dependent reflectors, so they're dependent on the stream of input signals being continuous. I think people are mostly okay with that terminology.

I could find out that I'm wrong about this, if a lot of people speak up and say that they find the current situation to be terribly confusing, and that they would be much happier if no one ever used "signal" in the context of a dependent reflector loop.

[confocaloid]

Here you've linked to a dictionary definition of "signal", not the LifeWiki's definition.

That's intentional.
I understand that the Life-specific definition of "signal" is bound to differ from the dictionary definition. However, I already attempted to discuss Life-specific meaning before in this discussion.
My point here is that it might be a good idea to keep meanings of terms sufficiently close to the "common-sense" meaning of the words.

You seem to be really stuck on trying to make the usage of "signal" on the LifeWiki match this definition that you've gotten from elsewhere -- i.e., the idea that all signals must be able to carry binary information, such that each signalling object can be either present or absent with no catastrophic effects on the circuitry.

You're misrepresenting my position.

The current definition of "signal" states:
"A signal is the movement of information through the Life universe. Signals can be carried by spaceships, fuses, drifters, or conduits."

My objection is to your proposal to describe all those gliders, spaceships, active objects going through conduits as "signals". That's not what is in the LifeWiki definition.

I'd certainly like to be able to say that they reflect period-N signal streams. If they don't, then why are they called "reflectors"?

They're called reflectors, because they are reflectors -- i.e., devices that can reflect spaceships without suffering permanent damage.

I could find out that I'm wrong about this, if a lot of people speak up and say that they find the current situation to be terribly confusing, and that they would be much happier if no one ever used "signal" in the context of a dependent reflector loop.

Several people already did post here in this thread.
At least as far as my understanding goes, several people already did disagree with you.

I don't really understand why it needs to be "a lot of people".

[dvgrn]

You seem to be really stuck on trying to make the usage of "signal" on the LifeWiki match this definition that you've gotten from elsewhere -- i.e., the idea that all signals must be able to carry binary information, such that each signalling object can be either present or absent with no catastrophic effects on the circuitry.

You're misrepresenting my position.

Sorry if so. I'm doing my best to state your position accurately. But statements like "there are no signals" are really confusing for me.

The current definition of "signal" states:
"A signal is the movement of information through the Life universe. Signals can be carried by spaceships, fuses, drifters, or conduits."

Yup. I like the current definition of "signal", though it seems like maybe a footnote should be added to make it clearer that "movement of information" doesn't have to mean the transmission of nice clean ones and zeroes.

I hope my interpretation is pretty clear by now, that the presence of a moving glider constitutes the movement of information -- and thus that a glider can be called a "signal" whenever it is convenient to do so.

My objection is to your proposal to describe all those gliders, spaceships, active objects going through conduits as "signals". That's not what is in the LifeWiki definition.

Okay. That's a good clear statement. It seems totally wrong to me, for the reasons stated above, but it's clear. I think that my meaning of "signal" _is_ implicit in the LifeWiki definition.

I would like to continue to describe "all those gliders, spaceships, active objects going through conduits" as signals -- without looking at each case to see if it's really a "signal" in this other sense that you seem to want to be the exclusive meaning.

Several people already did post here in this thread.
At least as far as my understanding goes, several people already did disagree with you.
I don't really understand why it needs to be "a lot of people".

I suppose I should do a formal catalogue, then.
- Scorbie's post attempted to define a one-to-one correspondence between dependent reflector inputs and outputs, and voted in favor of "loop". Didn't mention the word "signal".

- hotdogPi seemed to be in favor of dependent reflector loops being called loops. Didn't mention "signal".

- It seems like calcyman didn't express an opinion one way or another on whether "signal" is a good general term for "active objects going through conduits" -- that post was mostly a mild vote cast in favor of the term "loop".

- johamwit proposed a redefinition of "signal" which I think matches your preferred definition. Or possibly johamwit believed that it was a summary of the current usage of "signal", I don't know for sure (but examples like the dependent p5 and p6 signal-source-and-sink oscillators make me think that in fact it is not a summary of current usage.)

- Scorbie mentions signals here, saying "no preference" and mentioning that alternative wording "signal object" might not be the best:

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)"...

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 [of] 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.)

- The next post by Scorbie could be called a vote in favor of current usage of "signal" --

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

-- but this was a response to Just Me Saying Things, and it has "if" at the beginning, so it doesn't really count as a vote in favor of my position. (I've given some older historical examples of "signal" used by people other than me in a non-bit-stream-carrying sense. I can certainly find more if I hunt for them, but there aren't a whole lot of them out there, of course, because the great majority of things that have been called signals do happen to have the capacity to encode bit streams.)

- GUYTU6J said, among other things:

In my opinion, it is reasonable to say that a dependent glider reflector transmits microscopic information but not mesoscopic information.

... a "signal" should have some entropy, that is, it would be allowed to convey information that one could then use to perform logical operations... I disagree with the "dependent reflector is a signal reflector" statement...

- roolif said:

IMHO, a reflection is a change of direction of a straight beam, so it's not applicable to Herschels, Bs, etc, only to spaceships traveling in straight paths.

One of Jormungant's follow-ups said both "I agree with confocaloid" and

That being said, if you ask me to find a jargon term for "an object that moves through a non-empty pattern", I have nothing obvious that pops in my mind at all, it is a very vague concept that has no clear equivalency in our physical world. So I guess one could call that a signal, it is better than "linkable pattern with deterministic evolution in bounded structure", for instance. From this point of view, I guess there is nothing wrong with calling this a signal...

- Next was glider_rider:

I generally think they should be considered signals, or at least that there's nothing wrong with calling them that. They don't carry as much information as a stream of bit-carrying signals, and said information is generally far less useful, but they still carry information.

That gets us as far as the discussion about the most recent decade's non-traditional use of "shuttle" to mean "loop". I won't quote any of the posts responding to that, but I think we could say that there are three implied votes in there for not using "signal circuitry" for dependent reflectors, one new one from AlbertArmstain and two from johamwit and Jormungant, who had already expressed similar sentiments farther back.

... This all doesn't seem to add up to any preponderance of strong opinion one way or another.

I mentioned "a lot of people" just in an attempt to describe the kind of response that would be likely to convince me that I should stop using "signal" in the way I've used it for the last couple of decades -- which I thought I had picked up correctly from context from Dean Hickerson's periodic signal injectors and suchlike.

A few people giving possible alternate definitions of "signal" doesn't seem like a big enough response to justify such a big change -- especially when a few people also support (what I think of as) the existing definition. A lot of people strongly supporting one alternate definition, and nobody defending the existing definition, would be a lot more convincing.

[confocaloid]

dvgrn, you're wrong about this.

I don't understand why exactly it has to be "a lot of people speak up to say you're wrong about this", to stop this ridiculous back-and-forth.

Not only my position, but I think in the above post you're also misrepresenting positions of other people who posted here.

As far as I can follow this whole discussion (starting at the talk page Talk:Dependent reflector and the Oscillator Discussion thread, and then continuing here as a separate thread), dvgrn is consistently trying to push their point of view, effectively ignoring feedback from other members who already posted here.

[dvgrn]

dvgrn, you're wrong about this.

I don't understand why exactly it has to be "a lot of people speak up to say you're wrong about this", to stop this ridiculous back-and-forth.

You already asked that exact question. I answered it in the last two paragraphs of my previous post: the responses so far seem to show that broad agreement does not exist on this point. The LifeWiki's quality tends to suffer when it starts getting used to document the opinions of the most energetic editor, or the even the majority view in a contested issue. So I've asked for a pause in editing on this particular topic, until the disagreement gets sorted out.

If we can't make any progress here, maybe we could take the route of documenting the conflicting opinions about the proper usage of "signal". That's a little different from documenting non-overlapping meanings in the current usage of "signal", and it will tend to add a lot of potentially confusing verbiage (as we've seen very well here!)

So I'd prefer to avoid that if possible, and just leave the LifeWiki in more or less its current state for a while, until a solution presents itself.

Not only my position, but I think in the above post you're also misrepresenting positions of other people who posted here.

I'm certainly very interested in finding out how I've misrepresented anyone's position. It's very easy to do accidentally. Of course it's best if I find out from the people in question, not by anyone trying to speak for them.

I spent a lot of time combing through this thread and finding representative quotes from everyone, and carefully including a sample from everyone who has been in favor of limiting "signal" to the non-LifeWiki information theory sense of the term. I also very deliberately did not editorialize about the quotes that I didn't agree with.

It seemed to me that that level of careful review wouldn't leave any possible room for any accusations of misrepresentation ... but here we are anyway.

[confocaloid]

Please let some another uninvolved long-term member of the community to moderate this conflict.

The LifeWiki's quality tends to suffer when it starts getting used to document the opinions of the most energetic editor, or the even the majority view in a contested issue. So I've asked for a pause in editing on this particular topic, until the disagreement gets sorted out.
...

You're trying to "control" a conflict where you're an involved side. (Instead of letting someone else moderate the conflict.)
Your posts are consistently longer than my posts, and are rather attacking/aggressive.
That's not counting other things I already mentioned.

Of course it's best if I find out from the people in question, not by anyone trying to speak for them.

I spent a lot of time combing through this thread and finding representative quotes from everyone, ...

In my previous post, I intentionally did not clarify in any way what was (in my opinion) the misrepresentation. I didn't "representative quote" anything, nor rephrase anything. That's because I didn't think that could possibly be productive.

[dvgrn]

Please let some another uninvolved long-term member of the community to moderate this conflict.

I'll second that request -- with apologies for the inconvenience to anyone who wants to get involved.

In my previous post, I intentionally did not clarify in any way what was (in my opinion) the misrepresentation.

That's a fine idea in some respects ... but it does have the awkward side-effect that I have absolutely no idea what this misrepresentation might be.

Maybe a private message to me would help me sort that out, without appearing to contribute to the argument here? I won't take offense.

[Jormungant]

Well, I can say that dvgrn definitely misunderstood what confocaloid meant by what is "interesting", because that's not a subjective concept here, we are talking about the characterization that could distinguish between oscillators and circuits. The "interest" just mean say there is "information" or "something to learn" from the input stream, input stream which is assumed to be periodic is not "interesting" in that sense only, but of course there are a large set of patterns that are "not interesting", and that makes calling them "oscillators" to be more appropriate, but these might be interesting to us nonetheless.

I do not like the term "signal stream", and "signal circuitry", as having definitions naively implied from the definition selected for "signal" alone, I see no beauty and it, and no use in conversations. I would always drop the name as it is imprecise, and just say "stream" or "circuit" if I actually mean to be imprecise, but otherwise use either "periodic stream" and "dependent circuits", or "input stream" and "signal circuit" in respective contexts; basically, I ratter have "circuit" has the top class, and "signal circuit" and "periodic/dependent circuit" as subclasses with no intersection. If a reflector demands a periodic stream to survive, it cannot be part of a "signal circuit", in my book (that I do not plan to write). Allowing periodic circuit to be called "signal circuits" makes the line between oscillators and circuits too blurry, to the point I wonder if you have a good argument for a blinker to not classified as "signal circuit", and make sure that argument also works for the extended families of P2 oscillators that could "look like circuits". Assuming you found a convincing argument, and dispelled any possibility that "signal objects" can be lone dot trails, then here is your final challenge:

Code: Select all

x = 68, y = 73, rule = B3/S23
5$39bo$38bobo$37bo3bo$38bo3bo$39bo3bo$18bob2ob2o15bo3bo$17b2ob2o2b2o
15bobo$17bo24bo$24b2o$17b2o5b2o$17b2o$25bo$17b2o2b2ob2o$18b2ob2obo3$
51bob2ob2o$50b2ob2o2b2o$14b2o34bo$6bo6b3o10bob2o27b2o$5bobo6b2o10bo3bo
bo17b2o5b2o$4bo3bo6bo7b2o7bo17b2o$3bo3bo8b2o6bo5bobo25bo$2bo3bo10bo5bo
b2obo21b2o2b2ob2o$bo3bo45b2ob2obo$2bobo$3bo4$14b2o$14b2o33b2o$49b2o4$
61bo$60bobo$7bob2ob2o45bo3bo$6b2ob2o2b2o21bob2obo5bo10bo3bo$6bo25bobo
5bo6b2o8bo3bo$13b2o17bo7b2o7bo6bo3bo$6b2o5b2o17bobo3bo10b2o6bobo$6b2o
27b2obo10b3o6bo$14bo34b2o$6b2o2b2ob2o$7b2ob2obo3$40bob2ob2o$39b2ob2o2b
2o$39bo$46b2o$39b2o5b2o$39b2o$22bo24bo$21bobo15b2o2b2ob2o$20bo3bo15b2o
b2obo$21bo3bo$22bo3bo$23bo3bo$24bobo$25bo!

For all I have known, this is a p128 oscillator and not a signal circuit, but if you beg to differ I would want to know why.

That being say, maybe you believe there is no need for distinguishing between the two. Circuits do not expand (usually) and hence are guarantied to be periodic even if there is no-called "information" or not in the loops. Still, if one wants to make constructions with circuits, knowing what is "a valid signal" for a circuit is critical, and merits to classify circuit differently if require a specific subclass of "signals". So I am fine with "Signal" as a placeholder for "might be a spaceship or not, it moves so surely it is not a still-life or oscillator", but I would not use that definition to define subclasses or further concepts.

[dvgrn]

Allowing periodic circuit to be called "signal circuits" makes the line between oscillators and circuits too blurry, to the point I wonder if you have a good argument for a blinker to not classified as "signal circuit", and make sure that argument also works for the extended families of P2 oscillators that could "look like circuits".

Blinkers are certainly sometimes included as catalysts, in what I would call "signal circuitry". A blinker by itself doesn't constitute either a signal or a signal circuit, of course. There has to be some information moving from the input to the output of the circuit.

Assuming you found a convincing argument, and dispelled any possibility that "signal objects" can be lone dot trails, then here is your final challenge:

Code: Select all

x = 68, y = 73, rule = B3/S23
5$39bo$38bobo$37bo3bo$38bo3bo$39bo3bo$18bob2ob2o15bo3bo$17b2ob2o2b2o
15bobo$17bo24bo$24b2o$17b2o5b2o$17b2o$25bo$17b2o2b2ob2o$18b2ob2obo3$
51bob2ob2o$50b2ob2o2b2o$14b2o34bo$6bo6b3o10bob2o27b2o$5bobo6b2o10bo3bo
bo17b2o5b2o$4bo3bo6bo7b2o7bo17b2o$3bo3bo8b2o6bo5bobo25bo$2bo3bo10bo5bo
b2obo21b2o2b2ob2o$bo3bo45b2ob2obo$2bobo$3bo4$14b2o$14b2o33b2o$49b2o4$
61bo$60bobo$7bob2ob2o45bo3bo$6b2ob2o2b2o21bob2obo5bo10bo3bo$6bo25bobo
5bo6b2o8bo3bo$13b2o17bo7b2o7bo6bo3bo$6b2o5b2o17bobo3bo10b2o6bobo$6b2o
27b2obo10b3o6bo$14bo34b2o$6b2o2b2ob2o$7b2ob2obo3$40bob2ob2o$39b2ob2o2b
2o$39bo$46b2o$39b2o5b2o$39b2o$22bo24bo$21bobo15b2o2b2ob2o$20bo3bo15b2o
b2obo$21bo3bo$22bo3bo$23bo3bo$24bobo$25bo!

For all I have known, this is a p128 oscillator and not a signal circuit, but if you beg to differ I would want to know why.

I'm quite puzzled by this example, actually. I'd certainly say that this is a perfectly valid piece of (periodic) signal circuitry, containing two signals traveling in a loop.

This would be true according to most of the definitions of "signal" that have been suggested, not just the definition I've been using. You can freely remove or add one or both of the signals from the circuitry without damaging anything, so this particular loop can hold two bits of binary data *:

Code: Select all

x = 63, y = 63, rule = B3/S23
38bo$37bobo$36bo3bo$37bo3bo$38bo3bo$17bob2ob2o15bo3bo$16b2ob2o2b2o15bo
bo$16bo24bo$23b2o$16b2o5b2o$16b2o$24bo$16b2o2b2ob2o$17b2ob2obo3$50bob
2ob2o$49b2ob2o2b2o$13b2o34bo$5bo6b3o10bob2o27b2o$4bobo6b2o10bo3bobo17b
2o5b2o$3bo3bo6bo7b2o7bo17b2o$2bo3bo8b2o6bo5bobo25bo$bo3bo10bo5bob2obo
21b2o2b2ob2o$o3bo45b2ob2obo$bobo$2bo4$13b2o$13b2o5$60bo$59bobo$6bob2ob
2o45bo3bo$5b2ob2o2b2o43bo3bo$5bo50bo3bo$12b2o41bo3bo$5b2o5b2o42bobo$5b
2o50bo$13bo$5b2o2b2ob2o$6b2ob2obo3$39bob2ob2o$38b2ob2o2b2o$38bo$45b2o$
38b2o5b2o$38b2o$21bo24bo$20bobo15b2o2b2ob2o$19bo3bo15b2ob2obo$20bo3bo$
21bo3bo$22bo3bo$23bobo$24bo!

Was this intended to be a dependent loop? Maybe some other example would be closer to the intended challenge? Or what am I not understanding here?

* I suppose if there's an external clock and sufficiently clever measurement devices, there's some sense in which you could store either no signals, or one signal at 256/8 = 32 possible locations, or two signals at 128/8 = 16 possible positions. That's a lot more than two bits. But I hope it's clear what I mean by "two bits" -- there are two active objects (basically Herschels plus some trailing junk) that can each be independently present or absent.

[confocaloid]

Allowing periodic circuit to be called "signal circuits" makes the line between oscillators and circuits too blurry, to the point I wonder if you have a good argument for a blinker to not classified as "signal circuit", and make sure that argument also works for the extended families of P2 oscillators that could "look like circuits". Assuming you found a convincing argument, and dispelled any possibility that "signal objects" can be lone dot trails, then here is your final challenge:

Code: Select all

x = 68, y = 73, rule = B3/S23
5$39bo$38bobo$37bo3bo$38bo3bo$39bo3bo$18bob2ob2o15bo3bo$17b2ob2o2b2o
15bobo$17bo24bo$24b2o$17b2o5b2o$17b2o$25bo$17b2o2b2ob2o$18b2ob2obo3$
51bob2ob2o$50b2ob2o2b2o$14b2o34bo$6bo6b3o10bob2o27b2o$5bobo6b2o10bo3bo
bo17b2o5b2o$4bo3bo6bo7b2o7bo17b2o$3bo3bo8b2o6bo5bobo25bo$2bo3bo10bo5bo
b2obo21b2o2b2ob2o$bo3bo45b2ob2obo$2bobo$3bo4$14b2o$14b2o33b2o$49b2o4$
61bo$60bobo$7bob2ob2o45bo3bo$6b2ob2o2b2o21bob2obo5bo10bo3bo$6bo25bobo
5bo6b2o8bo3bo$13b2o17bo7b2o7bo6bo3bo$6b2o5b2o17bobo3bo10b2o6bobo$6b2o
27b2obo10b3o6bo$14bo34b2o$6b2o2b2ob2o$7b2ob2obo3$40bob2ob2o$39b2ob2o2b
2o$39bo$46b2o$39b2o5b2o$39b2o$22bo24bo$21bobo15b2o2b2ob2o$20bo3bo15b2o
b2obo$21bo3bo$22bo3bo$23bo3bo$24bobo$25bo!

For all I have known, this is a p128 oscillator and not a signal circuit, but if you beg to differ I would want to know why.

I get it that this is not my challenge, but here's my view anyway, for what it's worth.

I think the natural way to describe the pattern is indeed that it is a p128 oscillator. There are two active objects, cycled forever in a closed periodic track(*).
Since each of these active objects can be individually present or absent, with no fatal consequences for the structural integrity, one could describe this as a reasonable way to store some information -- assuming there is a reasonable way to read that information and make use of it somewhere else (so that presence or absence of these active objects determines behaviour of the larger pattern).
As it is, neither of these active objects is a signal. There's no movement of information here, the active objects are cycled forever in the same bounding box, if you try to measure speed then you get zero. I don't see any signals here. The pattern is an oscillator.

I guess that some sufficiently devilish argument could convince me, against my will, that I see "signal circuitry" here (everyone is gullible). I think that would constitute loss of understanding -- when one sees "signals" where there are no signals in any helpful sense, something is very very wrong.
----------
(*) I'm using the following definition of "track", taken from an old version of Life Lexicon:

:track A path made out of conduits, often ending where it begins so that the active object is cycled forever, forming an oscillator or a gun.

Unfortunately, the recent version of Life Lexicon redefined "track", and now it mentions something called "active signal object". There's no such thing as "active signal object". I believe this change constitutes damage.

I believe the old definition of "track" (quoted above) is helpful and accurate, and it should be reverted back to the form it was back in 2002, without any modifications.

Archived copy of the old definition (2002-05-23): https://web.archive.org/web/20020523023 ... .htm#track
Archived copy of the current definition (2023-07-17): https://web.archive.org/web/20230717004 ... .htm#track

[dvgrn]

Unfortunately, the recent version of Life Lexicon redefined "track", and now it mentions something called "active signal object". There's no such thing as "active signal object". I believe this change constitutes damage.

I believe the old definition of "track" (quoted above) is helpful and accurate, and it should be reverted back to the form it was back in 2002, without any modifications.

Archived copy of the old definition (2002-05-23): https://web.archive.org/web/20020523023 ... .htm#track
Archived copy of the current definition (2023-07-17): https://web.archive.org/web/20230717004 ... .htm#track

That change would have been due to the major update to the Life Lexicon that was in the works for several years, mostly 2016-2017. Here's the old definition and new definition a month apart in November and December 2017.

The odds are quite good that it was my idea to add the link to "signal" in the "Herschel track" entry. There wasn't a definition for "active object" to link to, but a definition of "signal" existed. It was uncontroversial at the time that active objects could be described as "signals".

It's not completely clear who made that specific change, though. A very large number of definitions had to be shared and reviewed around that time, since the Lexicon hadn't been updated since 2006. The closest communication I can find is this one:

Hello, Dave:

Hmm, I see there is already an almost empty :Herschel track: entry. You could use that for the
new info, and not have to add another entry.

(But then :track: is a bit weird, since it implies (but not forces) the track to be a loop.)

In my view, a track or a circuit or a path, either has a starting component and an ending component
to make a linear (not geometry wise, but connection wise) object, OR it has no start and stop
components and the components form a loop.

But can a Herschel component have a branch, so for example one Herschel input and two Herschels
as output? In which case 'path' is not such a good term, but 'circuit' could still be.

Also, to mess things up more, I suggest that :Herschel component: be made a synonym
to :Herschel conduit:. I like to think of these things as components of a circuit.

Tell me I'm all wrong, that's OK!

I'm glad I'm not the editor.

BCNU,
-dbell

Possible influence from a real-world definition
I've been thinking about this a bit more, and it occurred to me that there's one real-world sense of the word "signal" that seems like a particularly good match for the p128 oscillator that Jormungant posted.

Here is a sample use of "signal" to describe the repeating cyclic fluctuation in the AC power grid:

Power grids vary by nation. In the US, the grid is based on a highly stable 60-hertz signal, meaning it cycles 60 times per second.

I like the analogy to the cycling of signals in the p128 oscillator. I think this has also always been the "signal" analogy that works best for me, for things like the dependent p5 and p6 signal sources and sinks mentioned above.

[dvgrn]

@confocaloid, I'm not terribly happy about your recent edits to Dove and I-heptomino. These were both uses of the term "signal" by a member of the community who hasn't yet expressed an opinion on this thread. Those uses perfectly matched the widespread usage of "signal" from the 2018 Life Lexicon, as a general term for any active object moving through signal circuitry.

As such, they could be considered to be evidence that this sense of "signal" is in common use. Please continue to hold off for the time being on LifeWiki edits that attempt to fix the problem with this usage of "signal", until there's some broader agreement that this is actually a problem.

I'm sure you have the best of intentions with all this. However, after reviewing everything again, I still personally think that there are much better uses of everyone's time than attempting to patrol every existing and new use of "signal" that refers to active objects traveling in signal circuitry. These uses go back a decade and more. Until this discussion started, there have been no complaints that I've been able to find, about any actual confusion being caused by this.

It's not yet clear to me that there's any support for changing dozens of Life Lexicon entries in the way that you suggest. Most people who actually build complex signal circuitry have not given any opinion on this yet.

On that Life Lexicon thread I'll add a quick explanation of the size of the effort that you're requesting, for a rollback of the change to "track" that you've labeled as "damage".