Poll: Should the term "dependent reflector" be made to include reflectors that accept a range of timings?

[Scorbie]

Hello, this poll is intended to collect the current semantics of the term "dependent reflector".

- The LifeWiki definition of "dependent reflector" is based on periodic dependent reflectors. According to the current definition a dependent reflector requires the next active object to arrive at exactly P generations, where P is the dependent reflector's period.

- In other words, if we define X as a reflector that requires the next active object to arrive at a..b generations, X and "dependent reflector" are different terms. viewtopic.php?p=165398#p165398

- There is prior usage of quicksilver reflector included as dependent reflectorsviewtopic.php?p=163340#p163340. The current definition only allows quicksilver reflectors as dependent reflectors if it's used in a constant period stream.

This poll was made to check the community's opinion of the term "dependent reflector" coincides with this aspect of the LifeWiki definition.

[Scorbie]

Update: confocaloid noted that there are two possible different interpretations of my wording

In other words, if we define X as a reflector that requires the next active object to arrive at a..b generations, X and "dependent reflector" are different terms.

viewtopic.php?f=2&t=6059&p=165420#p165415
Specifically, `a..b generations` can be interpreted either as `a..b generations after the last active object that triggered the reflector` or as `k th object appears in generations pk+a .. pk+b`. If you have differing opinions between the two interpretations feel free to let us know here (if you already voted kindy let us know that you voted and optionally what you voted if you're okay) and maybe we could make another poll when needed.

[MathAndCode]

Why must we have such a binary mindset? Why not call them semidependent reflectors or something?

[confocaloid]

Why must we have such a binary mindset? Why not call them semidependent reflectors or something?

I liked the suggestion to use 'constrained reflector' in viewtopic.php?p=165413#p165413 -- timings between consecutive input gliders are constrained. I'm not sure how to describe periodic reflectors where the k-th input glider is required to appear in generations pk+a .. pk+b, regardless of timings of other gliders -- probably these could be another type of 'constrained reflector'.

I think I would keep 'dependent reflector' referring to periodic dependent reflectors that require a period-N input stream (i.e. keep the existing definition). Maybe also include reflectors that were (edit: or could be) previously believed to be dependent, until it was discovered that they can accept different timings of gliders (but the timing of output gliders is not affected by the change in timing of input gliders) -- i.e. 'pN gun engines that can be supported by a pN glider stream with two or more different timings mod N'.

edit: here's an example that could be believed to be a periodic dependent reflector, until noticing that the input glider stream can be advanced/delayed by one tick:

Code: Select all

x = 57, y = 42, rule = B3/S23
55bo$54bo$54b3o6$47bo$47bobo$47b2o5$40bo$39bo$39b3o5$24bo$22bobo$12b2o
6b2o$11bo3bo4b2o$2o8bo5bo3b2o$2o8bo3bob2o4bobo$10bo5bo7bo$11bo3bo$12b
2o$23bo$24b2o$23b2o6$30bobo$31b2o$31bo!

[dvgrn]

I like "dependent reflector" as a general term for reflectors that are dependent on their input signal streams -- i.e., you can't just stop sending dvgrn-signals in whenever you want, you have to keep up a steady stream.

The relatively rare subcategory of dependent reflector that allows some adjustability in its inputs could perfectly well be called "constrained dependent reflectors" or something along those lines.

Up to now it seems like "constrained dependent reflectors" haven't necessarily gotten talked about enough that they need their own terminology. We could invent a term, but then if and when people ever actually start talking about them, they might end up using some other term.

As a side note, the current definition of "dependent reflector" doesn't actually exclude the QuickSilver type of "constrained dependent reflector" -- it doesn't really mention them:

A periodic reflector is described as dependent if it requires that the incoming stream repeat with the same period as the reflector, or independent otherwise.

That doesn't say anything about whether a stable reflector can be described as "dependent", in the context of overclocking. We could add a paragraph talking about that case, and the overall definition would still work just fine -- except maybe for the edge case confocaloid just brought up, where the reflector is periodic but the suppressing input can be advanced or delayed a little bit.

... Come to think of it, we could easily engineer a periodic dependent reflector where the suppressing input could be advanced or delayed by hundreds or thousands of ticks -- e.g., if the oscillating part is a Herschel loop where the only support needed is the periodic removal of an extra block.

I don't think that artificial edge cases like that should invalidate the definition of "dependent reflector", however. It seems better to continue to use the existing terminology, and adjust the LifeWiki definition slightly so that it continues to mean what we mean it to mean. It doesn't have to cover all the weird edge cases in practice, because nobody really seems interested in applying the label "dependent reflector" to big Herschel loops with an input glider stream suppressing a stray block.

[Scorbie]

Why must we have such a binary mindset? Why not call them semidependent reflectors or something?

I want to clarify the intention of the poll, which I have purposefully left unclear because I wanted to minimize its effect to the result of the poll.

The term "dependent reflector" was used before its definition on the wiki. Sokwe's first post was from 2013 and uses the term "dependent".
viewtopic.php?f=2&t=1082&p=7822&hilit=D ... ctor#p7822

Because the adjective "dependent" has prior use in "dependent conduit" with the meaning `(dependent X) := X that depends on some Y inferred from context` although this wasn't on the wiki, I believe there was widespread consensus that "dependent reflector" meant `exists T, T reflector that depends on the next T` (Edit: fixed wrong quantification)

Calcyman's usage of this term supports this idea as well. The results of this poll also supports this despite the fact that this is different from the LifeWiki definition. I believe this is enough to show that the majority of the members have thought about it this way but one could ask members explicitly on how they perceive this term.

Recent edits of the term "dependent reflector" is trying to make the term non-confusing to newcomers which I think is pointing to the right direction. However, as a side effect the LifeWiki definition departed from what the majority of the community's perception of "dependent reflector"s which seemed to be a red flag to me.

I know there has been a lot of debate over the precise semantics of "dependent reflector" which can be hard and does not necessarily have an answer, but this is a single aspect of the definition and does have a well-defined answer to choose from.

[confocaloid]

Because the adjective "dependent" has prior use in "dependent conduit" with the meaning [ ... ]

A dependent conduit is a different thing and a different term. Those do not add constraints on the input Herschel stream, and can be used to transmit arbitrarily-timed (above the repeat time) signals carried by Herschels, the same way as one could do with an independent Herschel conduit. In that case, the word 'dependent' refers to [constraints on] possible connections between consecutive conduits in the track.

Recent edits of the term "dependent reflector" is trying to make the term non-confusing to newcomers which I think is pointing to the right direction.

I don't agree with 'recent'. The current definition was already there for some time before the current discussion.

[Scorbie]

A dependent conduit is a different thing and a different term.

I'm referring to Sokwe's first usage of the term where it originated. You can go to the link in my last post to see that it originated from the same adjective.

Edit:

I don't agree with 'recent'. The current definition was already there for some time before the current discussion.

I'll clarify that I meant the lifewiki definitions were recent compared to usage of the term since 2013.

[Sokwe]

I'm referring to Sokwe's first usage of the term where it originated.

I didn't originate the term "dependent reflector". I just copied it from Jason Summers' February 24, 2012 update to his Game of Life Status page. I looked for an earlier occurrence of the term in the LifeCA email archives but couldn't find one.

[Scorbie]

I stand corrected. I'm sorry about the misinformation and I can see that people can perceive this term both ways in existing usage (either accepting a range or requiring the stream to be strict)

[confocaloid]

Bump due to activity in forum thread viewtopic.php?p=167112#p167112

[dvgrn]

Bump due to activity in forum thread viewtopic.php?p=167112#p167112

I'm a bit puzzled by the response there:

Okay, so from what I understand, a dependent reflector is a glider reflector that explodes if it isn't fed a constant stream of input gliders.

In that state they really can only transmit information once: they signal the end of the input stream by exploding.

However, I thought: is there a dependent reflector that can accept gliders at multiple slightly offset input timings (i.e. if the period is p, it successfully accepts gliders at times t = p*n and t = p*n + k where 0 < k << p), and each input timing affects the output glider timing in the same way.

So you would actually be able to transmit a continuous stream of bits this way, by shifting the glider timings, even though if you omit any one glider the whole thing explodes (i.e. it's dependent).

Does something like this exist?

AFAIK the word 'dependent' is usually used only for periodic dependent reflector reactions (emitting output gliders at a fixed schedule and supported by an input glider stream with the same period), like those shown on the page Dependent reflector.

There are engineered reflectors (e.g. an overclocked Silver's reflector) that can work when the separations between consecutive input gliders are restricted to a set of values. This is described as overclocking.

See also the posts by several people in the forum thread viewtopic.php?f=2&t=6099

My interpretation of the intentions of the fifteen people (including me) who voted for the first option, was that they would want the term "dependent reflector" to include things like Silver reflectors, not just the strictly periodic stuff. Only four people voted for the second option.

That doesn't seem to match up very well at all with "the word 'dependent' is usually used only for periodic dependent reflector reactions". (?)

[confocaloid]

My interpretation of the intentions of the fifteen people (including me) who voted for the first option, was that they would want the term "dependent reflector" to include things like Silver reflectors, not just the strictly periodic stuff. Only four people voted for the second option.

That doesn't seem to match up very well at all with "the word 'dependent' is usually used only for periodic dependent reflector reactions". (?)

There are at least two errors in the above:
(1) Consensus is not reducible to voting; specific arguments matter. [Even considering only the number of votes, 15 versus 4+4=8 is far from unanimous support. See the posts in the forum threads for more relevant discussion.]
(2) The question "what 'dependent reflector' should mean in the future" is obviously different from the question "what 'dependent reflector' meant in the past, until the recent discussion".

As I wrote,

AFAIK the word 'dependent' is usually used only for periodic dependent reflector reactions (emitting output gliders at a fixed schedule and supported by an input glider stream with the same period), like those shown on the page Dependent reflector.

There are engineered reflectors (e.g. an overclocked Silver's reflector) that can work when the separations between consecutive input gliders are restricted to a set of values. This is described as overclocking.

See also the posts by several people in the forum thread viewtopic.php?f=2&t=6099

[dvgrn]

My interpretation of the intentions of the fifteen people (including me) who voted for the first option, was that they would want the term "dependent reflector" to include things like Silver reflectors, not just the strictly periodic stuff. Only four people voted for the second option.

That doesn't seem to match up very well at all with "the word 'dependent' is usually used only for periodic dependent reflector reactions". (?)

There are at least two errors in the above:
(1) Consensus is not reducible to voting; specific arguments matter. [Even considering only the number of votes, 15 versus 4+4=8 is far from unanimous support. See the posts in the forum threads for more relevant discussion.]

I did not mention the word "consensus", nor did I intend to refer to the concept. I'm not sure why you brought it up here.

I did not claim "unanimous support", or anything like it. I do think that "15 versus 4" would be a more accurate description of the poll than "15 versus 8".

However, either of those counts would still seem to suggest that fifteen people would probably consider this to be a misleading summary --

"the word 'dependent' is usually used only for periodic dependent reflector reactions"

-- and it's not even clear that the other four or eight voters would consider it to be an accurate description of how "dependent" is currently used, or has been used in the past. The poll started with the word "Should". It seems possible that some of those four voters might have been stating a preference about how the term should be used in the future.

[confocaloid]

[...] It seems possible that some of those four voters might have been stating a preference about how the term should be used in the future.

All that is an attempt to "interpret" votes of other people in some way, while there is no evidence that such interpretations would be correct.

Further, the whole poll is not well-defined, and misses/hides several issues/distinctions/possibilities. Consensus is not reducible to voting. Specific arguments matter.

Update: confocaloid noted that there are two possible different interpretations of my wording

In other words, if we define X as a reflector that requires the next active object to arrive at a..b generations, X and "dependent reflector" are different terms.

viewtopic.php?f=2&t=6059&p=165420#p165415
Specifically, `a..b generations` can be interpreted either as `a..b generations after the last active object that triggered the reflector` or as `k th object appears in generations pk+a .. pk+b`. If you have differing opinions between the two interpretations feel free to let us know here (if you already voted kindy let us know that you voted and optionally what you voted if you're okay) and maybe we could make another poll when needed.

I would say that QuickSilver is not a dependent periodic reflector (it's obviously not periodic), but is a dependent stable reflector (it's locally-stable circuitry that puts constraint on what the input signals might be to allow it to continue functioning as a reflector). Maybe it's a good idea to reserve 'dependent reflector' only for periodic dependent reflectors, and use something like 'constrained reflector' for QuickSilver, but I'm fine either way, with slight preference to extending terminology instead of introducing new one in this case.

Why must we have such a binary mindset? Why not call them semidependent reflectors or something?

[dvgrn]

[...] It seems possible that some of those four voters might have been stating a preference about how the term should be used in the future.

All that is an attempt to "interpret" votes of other people in some way, while there is no evidence that such interpretations would be correct.

Further, the whole poll is not well-defined, and misses/hides several issues/distinctions/possibilities.

My statement said specifically "It is possible that some of those four voters ...". I'm allowed to say that. I think it is possible. As you say, the poll was not absolutely perfectly designed to remove all possible ambiguity from everyone's vote.

If you want to run another poll that removes all possible ambiguity, please go ahead. If you succeed in doing that, then we can all stop speculating about what we all really think.

However, in the meantime, the existing poll actually seems to me to have given a reasonably clear and unambiguous result: the community mostly seems to believe that the term "dependent reflector" should be allowed to include reflectors that accept a range of timings. This doesn't seem terribly surprising, since we have well-documented existing usage by calcyman and others, of "dependent reflector" being used in exactly that way.

Consensus is not reducible to voting. Specific arguments matter.

Once specific arguments have been made and either accepted or rejected, however, repeating those specific arguments without adding any new information ... does not prevent the larger group from eventually reaching a consensus, however strongly the objectors might continue to object.

[confocaloid]

[...] However, in the meantime, the existing poll actually seems to me to have given a reasonably clear and unambiguous result: the community mostly seems to believe [...]

If anything, I think the community mostly seems to believe that the question in the poll is either unimportant or ill-defined. 23 people in total are a small fraction of the community.

Even looking at the results of the poll as they are, it is incorrect to describe 15/23 as "mostly seems to believe", especially given that there are specific arguments by several people in the forum posts that are largely ignored / were not discussed in depth.

[dvgrn]

If anything, I think the community mostly seems to believe that the question in the poll is either unimportant or ill-defined. 23 people in total are a small fraction of the community.

It's a mighty small community, though! I would say that 23 is a very significant fraction of the people that we're going to be able to find who have any kind of opinion on this subject. As you say, most of the community seems to believe that the question is unimportant -- and that's perfectly fine. What we're looking at is results from people who thought the question was important enough to vote on.

If we're unlikely to get better data than this poll has provided, then it seems like a reasonable plan to proceed with some minor LifeWiki editing, on the basis of the preference that has been expressed in the poll -- rather than declaring that this is some kind of impasse. If the poll turns out to be wrong, then ... well, it's a wiki, so we can really easily fix whatever we might do at this point.

Acknowledging the "dependent stable reflector" case in the article, as the rather weird edge case that it is, does not seem like it has to be particularly difficult or controversial. For example, how about a second paragraph saying

"A stable reflector can also be described as dependent if it is overclocked, so that it depends on a steady stream of inputs to continue to function."

-- and then clarify in the third paragraph that

"A honey farm predecessor is one of the most commonly used active reactions in a periodic dependent reflector."

?

[confocaloid]

I disagree. Since there were no prior uses, before/outside this discussion, of the phrase 'dependent reflector' to describe patterns that are not periodic dependent reflectors, such meanings should not be documented on LifeWiki at this point.

Whatever new uncommon terminology might be invented here in these recent terminology-related threads, may fail to be useful or convenient or reasonable.

In this particular case, I think writing 'overclocked reflector' directly is clearer, compared to inventing a new meaning of 'dependent' only to be able to write the same thing in an indirect and less clear way.

"Dependent reflector" has quite regularly been used as a description of overclocked Silver reflectors.

This is inaccurate. I cannot find prior uses of the phrase 'dependent reflector' in that way, before/outside this forum thread and the recent terminology-related discussion.

That's true -- I'm not finding a lot of support for my "quite regularly" characterization, so I'll withdraw those two words. [...]
[...]

[...]
Acknowledging the "dependent stable reflector" case in the article, as the rather weird edge case that it is, does not seem like it has to be particularly difficult or controversial. [...]

[dvgrn]

I disagree. Since there were no prior uses, before/outside this discussion, of the phrase 'dependent reflector' to describe patterns that are not periodic dependent reflectors, such meanings should not be documented on LifeWiki at this point.

Hmm. Well, I don't really mind if my suggestion doesn't get implemented in this case. Obviously I think it's still a true statement even if it doesn't make it into the LifeWiki article.

However, it also doesn't seem like there's any particular reason to ignore valid uses just because they happened in this discussion. How should the opinions of the fifteen people who voted in favor of "including reflectors that accept a range of timings" be acknowledged in the article, then?

On the other hand, overclocked Silver reflectors are definitely a fairly trivial weird edge case in the modern world of dependent reflectors. There are lots of weird edge cases for all kinds of definitions that we don't mention in LifeWiki articles -- and that's perfectly fine.

I'm actually quite a bit more concerned by these other non-QuickSilver-Demonoid cases, where something can get called a "dependent reflector" in good faith, until somebody suddenly notices that it also works if you vary the input timing a little bit:

here's an example that could be believed to be a periodic dependent reflector, until noticing that the input glider stream can be advanced/delayed by one tick...

That is, in fact, a periodic dependent reflector. It's periodic, and it's a dependent reflector. It doesn't seem at all acceptable to say that it's not a dependent reflector just because several distinct different dependent reflectors also exist using the same basic mechanism, where the input comes in with a slightly different timing.

That seems like it would be a mistake due to trying to enforce an overly literal reading of the current definition of "dependent reflector" -- which was only invented quite recently. That definition was written to fit the particular types of dependent reflectors that have been getting discovered recently, maybe without quite enough thought about the various edge cases. The current definition shouldn't be considered to be set in stone for quite a while yet.

[confocaloid]

here's an example that could be believed to be a periodic dependent reflector, until noticing that the input glider stream can be advanced/delayed by one tick...

That is, in fact, a periodic dependent reflector. It's periodic, and it's a dependent reflector. It doesn't seem at all acceptable to say that it's not a dependent reflector just because several distinct different dependent reflectors also exist using the same basic mechanism, where the input comes in with a slightly different timing.

That seems like it would be a mistake due to trying to enforce an overly literal reading of the current definition of "dependent reflector" -- which was only invented quite recently. That definition was written to fit the particular types of dependent reflectors that have been getting discovered recently, maybe without quite enough thought about the various edge cases. The current definition shouldn't be considered to be set in stone for quite a while yet.

Any newer modification of the definition would be even more recent in comparison.

That p30 example can actually be advertised and viewed as a p30 dependent reflector. Such advertisement does not mention that the input gliders can come at different timing (but that does not affect timing of the output gliders).
Another example where the input gliders are all delayed by one tick, can also be described as a p30 dependent reflector, in the same way.
The two examples are "essentially the same", in the sense that the core mechanism (the reaction and the removal of beehives with incoming gliders) is the same.

[skomick]

Sorry for any inconvenience, but for the record I meant to vote for INCLUDE, not EXCLUDE.