Poll: should the LifeWiki definition of the term "signal" be adjusted?

[dvgrn]

Okay, this is the next attempted step at gathering community opinion on a topic that confocaloid and I have been having a lot of difficulty with over the past several weeks or so. See here and maybe here and here, if you want to work through all the details.

One part of the topic that currently has us at an impasse is the question of whether or not to make a specific edit to the signal article on the LifeWiki. Here are the two options in their current form, with a little more detail than what fits in the poll questions:

• Option 1:
  The p43 Snark loop contains no signals. Circulating gliders in a context like this should not be referred to as "signals" in LifeWiki articles, because they do not encode any information. The current LifeWiki definition of "signal" is complete and correct as it stands:

  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.

  LifeWiki and Life Lexicon articles should be edited to remove uses of "signal" that imply that the p43 Snark loop can contain signals.
  ​
• Option 2:
  The p43 Snark loop contains eight circulating signals. The LifeWiki "signal" article should be adjusted slightly so that it's clear that this is a mainstream use of the term "signal":

  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. "Signal" can also refer directly to active objects traveling through circuitry -- gliders, Herschels, B-heptominoes, etc.

If neither of the above two options looks right, please respond with a post stating an option that you would vote for -- or, e.g., state your agreement with an option described in someone else's post.

Confocaloid and I will attempt to leave space for other people to discuss the issue here, though we are certainly likely to respond to questions if anyone asks us directly for clarification.

[confocaloid]

For the record, I do not believe the poll is helpful. There are multiple different issues, several of which are lumped together in this poll. Several people already replied in previous forum threads.

There is the concept of signals, and there is the word 'signal'.
The concept of signal has to be well-defined and unambiguous.
The word 'signal' is ambiguous; different meanings should be listed separately from each other.

In places where clarity is expected (definitions, introductory text on LifeWiki), the distinction between a signal and an active object should be maintained, because this distinction is important for understanding. An active object does not have to carry a signal.

Oscillators like the p43 Snark loop or the capped period-256 glider gun contain active objects (gliders / Herschels), but there are no choices to be made that would be somehow represented or encoded in those active objects. It is misleading to claim existence of 'signals' in those patterns (a reader is misled to look for something that is not there). Terminology should be helpful.

Regarding existing terminology and existing distinctions between related-but-different concepts:

As far as I can tell, the current definition of signal comes from (no later than) 2003 version of Life Lexicon:

:signal 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.

The use of 'active object' is a long-standing terminology and distinction. It can be found for example on the page Stephen Silver on Stable Reflectors --

[...]

It is a fact of Life that a given R-to-Herschel conduit almost never works with any given source of R-pentominoes. There is a good reason for this. While Herschels are usually produced by a standard evolutionary sequence which throws the Herschel away from its source, R-pentominoes (and most other common active objects) tend to be produced in relatively inaccessible places. I was, therefore, not very hopeful about the usefulness of the new conduit. [...]

[...]
It should perhaps be pointed out that all these reflectors and converters are not quite as bad as their large recovery times make them appear. It is possible, for certain periods, to pack multiple active objects into the reflector/converter with each one deleting the beehive (or, in the case of Dave Buckingham's reflector, replacing the boat) that would normally be dealt with by a later one. [...]

These distinctions existed long before this discussion, and they are reflected in the existing definitions. My preference here is to keep the existing definition of the concept of signal unchanged. Any ambiguous usage of the word should be explained separately.

[confocaloid]

Crosspost from Life Lexicon update collection thread

Creation of the poll is misleading. I do not consider the poll legitimate. Instead of creating even more and more scattered threads (misleading/diverting attention of community to irrelevant issues), it would be much more reasonable and productive to keep the discussion in a single place, and take into account already stated replies/opinions/feedback.

The above quotations are from sources that are two decades old or more. That certainly establishes their antiquity. However, their relevance to a debate that is happening in 2023 is questionable at best.

[...] It just plain isn't going to work (in my opinion) to pretend that it's still 2003 and that the definition of "signal" has stayed the same for the last twenty years.

The definition of signal did in fact stay the same since 2003.
You are proposing to re-define signals in your preferred way.

The existing definitions are still fine as they are. Just because the word is ambiguous, does not mean there is anything wrong about the concept.

The above quotations are directly relevant. Those quotations show that 'active object' was used long time ago for things moving through circuits, and 'signal' was defined the same way as it is still defined today. There are also more recent examples of the same usage.

Existing definitions explicitly distinguish between signals and carrying active objects. This is a useful distinction important for understanding -- not something to be discarded.

[Haycat2009]

Something to challenge that definition: Does Nihonium have 0 signals (As it is a loop), 2 signals (The herschels) or 4 signals (The Herschels+Gliders)? If I insert a Bx106 between the halves, does the number of signals change?

[confocaloid]

Something to challenge that definition: Does Nihonium have 0 signals (As it is a loop), 2 signals (The herschels) or 4 signals (The Herschels+Gliders)? If I insert a Bx106 between the halves, does the number of signals change?

I did not understand what you mean by inserting a Bx106 between the halves.

Nihonium is a p113 oscillator made out of two R64 and two R49, with two Herschels. An 'interesting thing' about this is that the first natural gliders from the two Herschels collide and delete each other; if only one Herschel is left, then the FNG will damage a R64.
If for some reason one wants to ignore that 'interesting thing', then one could say that it is a closed Herschel track with two active objects.
Signals are not relevant. There is no movement of information across the Life universe, there is no communication, no encoding, no attached meaning, no logic operations.

[dvgrn]

Just for the record, I'd be very interested to hear from people who haven't spoken up on this thread yet, especially if they voted to keep the definition the way it is.

For example, was the choice of Option 1 more of a vote against changing the definition in the specific way I suggested in Option 2? Or was it really a nice un-complicated vote in favor of preserving the current implied restriction on the meaning of "signal" (to information-theory signals only)?

EDIT: The discussion, such as it is, has most recently been happening in the Life Lexicon update collection thread, which is not represented in confocaloid's links below. I've been keeping my responses mostly in that thread recently, more or less for the reasons confocaloid has given.

I'm hoping to reserve this thread for a broader cross-section of the community to express their opinions, if they're willing to do so. It would be very helpful to me in resolving this rather sticky issue.

[confocaloid]

Either of two "options" mixes several different issues in a misleading way.

The question of whether or not the p43 Snark loop contains any signals is different from the questions regarding the content of the entry signal. What should or should not be considered a "mainstream use" of words is another different question. How to word articles that rely on the concept of signals is yet another different question.

This poll thread fails to distinguish different issues, and it ignores previous discussion with existing feedback. You cannot resolve these issues by making a poll and scattering discussion over multiple places. It would be more sensible to keep discussion in a single place.
viewtopic.php?f=2&t=6100
viewtopic.php?f=2&t=6059
viewtopic.php?f=2&t=6099

[silversmith]

Just for the record, I'd be very interested to hear from people who haven't spoken up on this thread yet, especially if they voted to keep the definition the way it is.

For example, was the choice of Option 1 more of a vote against changing the definition in the specific way I suggested in Option 2? Or was it really a nice un-complicated vote in favor of preserving the current implied restriction on the meaning of "signal" (to information-theory signals only)?

I voted to keep the definition of signal the same for a few reasons.

1. My main reason, we already have a term for the things which would be included in the change, aka. "active objects traveling through circuitry -- gliders, Herschels, B-heptominoes, etc." They are drifters.

   A drifter is a perturbation moving within a stable pattern.

   Unless we change the definition of a drifter as well, a glider moving through vacuum is a drifter, a glider bouncing off snarks is a drifter, a glider bouncing off buckaroos is a drifter, and Herschels, B-heptominos, etc... moving though conduits are drifters.
   
   Based off of this, it just makes sense to standarize these moving objects which may or may not carry information as "drifters".
2. Additionally, "signals" as things carrying information do not necessarily need to move relative to the grid. For example, in the caterloopillar information is carried by loafs, which as far as I can tell would fall under "signal", but not under "active objects traveling through circuitry -- gliders, Herschels, B-heptominoes, etc."

In summary, a "thing which carries information" is separate from "a thing which moves", and an object can be both, neither, one, or the other. Therefore, they should have two independent terms to refer to them.

[bubblegum]

For posterity, so silversmith's post from a good few days ago doesn't appear like the last word on the subject: the majority view and only sense of the term for most of its existence is that the LifeWiki definition of the term "drifter" is, strictly speaking, incorrect and is generally understood as referring to perturbations within still, non-vacuum media.

"Drifter" originated as the name of a search program written by Dean Hickerson, in common parlance shortened to dr to avoid ambiguity, which was designed to search for active perturbations within a stable background. Searches using dr began with a specified small cutout of a still-life background pattern, upon which a small initial perturbation would be specified. The term "drifter" thus began to be applied to patterns with the implication that dr would probably be good at finding them. dr was not intended to search for spaceships or conduits, so they were never referred to as drifters.

So by giving a cursory glance at some of the patterns that have been referred to as drifters, you come out with the impression that drifters are perturbations within dense stable media where both the medium and the perturbation have splits between OFF and ON cells that are somewhat close to 50/50. Or rather, the perturbation is often dominated by OFF cells, effectively "moving" as cells on the leading edge of the perturbation die off and ones on the trailing edge are born.

Of course, with a definition as vague as "something this program would be good at finding", you could definitely find edge cases where a perturbation is mostly ON and travelling through a mostly OFF medium. You could draw a bright-line division between what is and what isn't a drifter, but it'd be a lost cause - someone's bound to disagree with you. On the other hand, as you get closer along the line of active regions to a free-floating signal travelling through conduits, what you'll notice is that people start telling you to stop calling the things you encounter along the way drifters.

So what's wrong about just using the term "drifter" exactly how it's currently given on LifeWiki? Well, aside from the fact the wiki should reflect established usage and not the other way around, "drifter" is a nice, clear term to use because of its specificity - the single word conjures up a reasonably well-defined mental image of what it refers to. Generalise it and that image is destroyed, and you end up with a term about as concrete as confocal's "active object". Or, no, even more vague, since "object" typically refers to some congregation of ON cells, while a drifter is free to incorporate a much more significant amount of active OFF cells, so essentially a synonym of the "active region" I gave as a hypernym above.

The question of whether or not the p43 Snark loop contains any signals is different from the questions regarding the content of the entry signal. What should or should not be considered a "mainstream use" of words is another different question. How to word articles that rely on the concept of signals is yet another different question.

Perfect is the enemy of the good. Whether or not the two options given by the poll actually represent a conflation of multiple independent and increasingly general questions, debating whether the poll should be completely disregarded due to the answer to that serves little in the way of progress.

I personally don't see how the different questions you propose don't answer themselves, however. LifeWiki glossary articles must reflect actual mainstream usage, and whether the interpretation of the p43 as containing eight circulating signals is a mainstream use of the term "signal" is to be determined by the poll. Hence I also don't see why you have such an issue with the existence of the poll, so I'm all ears if you'd care to elaborate.

You claim that "the existing definitions are still fine as they are", I'd like to explicitly and directly disagree with that. Of no value is a definition which reflects incorrectly upon actual contemporaneous usage of the term it defines, especially in this community which has proven time and time again that they will happily reject any perfectly logical term definition in favour of whatever they first came up with when the concept the term referred to started being a thing, subtly misinterpreted over a number of years until no accurate written definition exists. The 2003 definition still holds water perfectly well, yes, but it takes a few seconds to change a definition, compared to the mountain of luck and sea of patience required to change general usage, i.e. the term's actual meaning.

Just for the record, I'd be very interested to hear from people who haven't spoken up on this thread yet, especially if they voted to keep the definition the way it is.

I voted to update the definition, to generalise it. In the case of "signal", unlike for "drifter" which could draw out a defined mental image, the term has never referred to a specific pattern of patterns, merely a possible usage of objects in relation to other patterns. It's always been a relatively general term, and as is the tendency for such terms has evolved to serve as a catch-all for any object that has the potential to fill in the role, even if it doesn't in that specific instance, suggesting that it just isn't that useful to have a dedicated term for the more specific sense especially without a similar dedicated term for the more general sense.

None of the suggested candidates for replacing "signal" in that more general sense seem particularly compelling either. "Drifter" is unsuitable and also would need to be replaced itself, which just sounds like too much trouble. "Active object" is far too vague - you would never describe an active object in some generic reaction as a signal - and it also doesn't seem like all patterns falling under "signal" would be considered objects, such as the drifters in 2c/3 wire and similar. "Active region" is even more vague; you might as well just call them "things".

Thanks goes out to dvgrn for giving me the kick in the pants necessary for me to finally say something here about this, and also for coming up with more than a handful of lines I stole directly from to put the least amount of conscious effort into this I possibly could.

[Haycat2009]

Something to challenge that definition: Does Nihonium have 0 signals (As it is a loop), 2 signals (The herschels) or 4 signals (The Herschels+Gliders)? If I insert a Bx106 between the halves, does the number of signals change?

I did not understand what you mean by inserting a Bx106 between the halves.

Nihonium is a p113 oscillator made out of two R64 and two R49, with two Herschels. An 'interesting thing' about this is that the first natural gliders from the two Herschels collide and delete each other; if only one Herschel is left, then the FNG will damage a R64.
If for some reason one wants to ignore that 'interesting thing', then one could say that it is a closed Herschel track with two active objects.
Signals are not relevant. There is no movement of information across the Life universe, there is no communication, no encoding, no attached meaning, no logic operations.

Here's something to get you thinking: Does a R64, Fx119 and F166 loop have 1 signal or none? If I replace a Fx119 with a Bx106, does the number of signals changes?

[confocaloid]

[...] confocal's "active object".

Apparently you ascribe this phrase to me. I'd like to explicitly and directly disagree with that. I did not invent that phrase.

The phrase 'active object' was used before (and is still used) for a long time (example 1, example 2; example 3 in the definition of "track" in 2003 version of Life Lexicon, example 4, example 5, example 6).
Some other examples where 'active object' is used in the context of conduits and circuits, to refer to the input/output object:

I know I've written all of this up before, but not for U-turners. Still, the same classification system applies to any elementary active object.

My idea would be that a conduit should be cross-posted to this thread as soon as there are multiple distinct ways to start from active object X, convert it to some other known active object already used in the ECC, and then convert back again to object X -- in such a way that arbitrarily long chains can be created.

I think this choice is fairly consistent, and it avoids confusion between the signals (which are carried by some mechanism and are encoded in some way) and the (not necessarily signal-carrying) active objects.

You claim that "the existing definitions are still fine as they are", I'd like to explicitly and directly disagree with that. Of no value is a definition which reflects incorrectly upon actual contemporaneous usage of the term it defines, especially in this community which has proven time and time again that they will happily reject any perfectly logical term definition in favour of whatever they first came up with when the concept the term referred to started being a thing, subtly misinterpreted over a number of years until no accurate written definition exists. The 2003 definition still holds water perfectly well, yes, but it takes a few seconds to change a definition, compared to the mountain of luck and sea of patience required to change general usage, i.e. the term's actual meaning.

Destruction is easier than construction.

If an idea was "subtly misinterpreted over a number of years until no accurate written definition exists", then the next question would be: why even one would choose to have an inaccurate "definition" that does not actually define a single concept (as opposed to having a disambiguation page, that could actually help in such cases, by distinguishing existing meanings of the word or the phrase)?

IMO when there are two or more different meanings, a disambiguation page would be "the way to go".

[silversmith]

..."drifter" is a nice, clear term to use because of its specificity - the single word conjures up a reasonably well-defined mental image of what it refers to. Generalise it and that image is destroyed, and you end up with a term about as concrete as confocal's "active object".

I don't believe that is a valid argument, since a term is fully capable of being defined as a general class of things, and still conjuring a well defined image of a subset of things which it refers to in conversation.

• Geometry: a rectangle refers to a generalized set of things, yet it still conjures up a well-defined image of a rectangle. If I said football is played on a rectangular field, it wouldn't lead to confusion on whether I mean a square field.
• Set therory: a subset refers to a any set whose elements are contained in another set, yet it still conjures up the well-defined image of the subset being smaller. If I said still lifes are a subset of oscillators, it wouldn't lead to confusion on whether I mean all oscillators are still lifes.

I could give more examples, but given this, I don't see any reason why your ability to form a "mental image" of a drifter would be harmed if something like a Herschel would also count as a drifter.

[haaaaaands]

bruh i see 8 gliders there and if signals are transmitted by that type of shape then it has 8 signals

[confocaloid]

In addition to my reply above ( viewtopic.php?p=167779#p167779 ):

Hence I also don't see why you have such an issue with the existence of the poll, so I'm all ears if you'd care to elaborate.

There are several different underlying issues, that were mentioned or partially discussed in the related forum threads.

• Whether or not the moving active objects can consistently be identified with carried signals. (The likely answer is "no" for several reasons: a single signal can be carried by an assembly of two or more moving objects; a signal can be carried by a temporary stationary object near a sparker that will "revive" it later; a moving object does not necessarily carry a signal; you will sometimes want to ignore part of the active reaction when deciding on the "current location" of the signal; fundamentally, the signals remain unchanged as they move, while the active objects evolve into different active objects.)
• Whether or not reflectors, conduits, converters should be considered as things that necessarily contain some signals. Is it really necessary to invoke the concept of signals when explaining conduits? Or maybe it would suffice to have a simpler introduction, in which a conduit is a passing reaction controlled by catalysts/sparkers, which can be used to transmit information but does not have to be used to transmit information?
• Whether or not dependent reflectors should be considered as things that have something to do with the idea of signals. Or maybe they are just glider gun engines supported by glider streams, working at a specific period and not adjustable to any other period.
• Whether or not it is a good idea to semi-mechanically insert the word 'signal' in many LifeWiki articles, without considering the question "Does the addition of this word help to explain this particular topic, or in this case it is mostly an irrelevant distraction for the majority of readers from the target audience?"

There should be several other related issues that I forgot to mention above, but these should be already enough to explain why I believe that this poll thread conflates several different issues in a confusing way, to the point that I do not believe that the results can be meaningfully interpreted.

(For example, just because a glider can sometimes be used to transmit a signal, and there are 8 gliders in a glider loop oscillator, does not mean that those gliders in that glider loop oscillator actually transmit signals. Even assuming that, it still would not follow that there are exactly 8 of them -- you could view pairs of consecutive gliders as "signals", or you could view all 8 gliders together as a single "signal", etc.)

[silversmith]

• Whether or not the moving active objects can consistently be identified with carried signals. (The likely answer is "no" for several reasons: a single signal can be carried by an assembly of two or more moving objects; a signal can be carried by a temporary stationary object near a sparker that will "revive" it later; a moving object does not necessarily carry a signal; you will sometimes want to ignore part of the active reaction when deciding on the "current location" of the signal; fundamentally, the signals remain unchanged as they move, while the active objects evolve into different active objects.)

One more concrete example where "dvgrn-signal" means something independent of "confocaloid-signal" is when one "dvgrn-signal" encodes multiple "confocaloid-signals". Specifically, the High-bandwidth Telegraph produces one "dvgrn-signal" each 960 generations, but also produces 5 "confocaloid-signals" in the same timespan.

Put simply, the High-bandwidth Telegraph emits one drifter with 5 signals every 960 generations.

Edit: The first two sentences are inaccurate. Instead, one "drifter" (a perturbation moving within a stable pattern) can contain multiple "confocaloid-signals".

[confocaloid]

For what it is worth, I don't think those "placeholders" are helpful in this discussion. They were invented by dvgrn, and not me. Hence, I cannot be totally sure whether I understand what is meant by 'confocaloid-signal' or 'dvgrn-signal'.

Signals (as they are currently defined, and were defined before) are information that moves through the Life plane, carried by some mechanism and encoded in some way.
A single drifter can carry multiple signals at once. For example, the signals could be independent ternary choices, and one could send two signals with one drifter, encoded as one of 3^2 = 9 possible timings.

[silversmith]

For what it is worth, I don't think those "placeholders" are helpful in this discussion. They were invented by dvgrn, and not me. Hence, I cannot be totally sure whether I understand what is meant by 'confocaloid-signal' or 'dvgrn-signal'.

I am using them as shorter placeholders for [my understanding of confocaloid's currently preferred definition of signal] and [my understanding of dvgrn's currently preferred definition of signal] respectively. For me:

...one "drifter" (a perturbation moving within a stable pattern) can contain multiple "confocaloid-signals".

is synonymous with:

A single drifter can carry multiple signals at once.