Rule definition terminology

[confocaloid]

This term is only to be used in the time of this edit war, that is, when the edit war ends all instances of "law" are replaced with the preferred term.

Can you write down a definition? What it is supposed to mean, exactly?

[HerscheltheHerschel]

I propose the term "law" for conditions (as confocaloid says)/transitions (as the world says). This term is only to be used in the time of this edit war, that is, when the edit war ends all instances of "law" are replaced with the preferred term.

Can you write down a definition? What it is supposed to mean, exactly?

"Law" refers to birth and survival conditions.

[confocaloid]

Can you write down a definition? What it is supposed to mean, exactly?

"Law" refers to birth and survival conditions.

• How to describe 'B6' when it means birth on 6 alive neighbours?
• How to describe the digit '6' in 'B36/S23'?
• How to explain the situation when two distinct CA rules have similar-looking rulestrings (e.g. B3/S23 and B36/S23)?

[HerscheltheHerschel]

Can you write down a definition? What it is supposed to mean, exactly?

"Law" refers to birth and survival conditions.

• How to describe 'B6' when it means birth on 6 alive neighbours?
• How to describe the digit '6' in 'B36/S23'?
• How to explain the situation when two distinct CA rules have similar-looking rulestrings (e.g. B3/S23 and B36/S23)?

B6 is a birth law.
The digit 6 in B36/S23 is a law.
B3/S23 and B36/S23 have similar laws.

[confocaloid]

"Law" refers to birth and survival conditions.

• How to describe 'B6' when it means birth on 6 alive neighbours?
• How to describe the digit '6' in 'B36/S23'?
• How to explain the situation when two distinct CA rules have similar-looking rulestrings (e.g. B3/S23 and B36/S23)?

B6 is a birth law.
The digit 6 in B36/S23 is a law.
B3/S23 and B36/S23 have similar laws.

I like the basic idea to avoid controversial terminology (at least temporarily) altogether.
However, instead of coining a new term ('law') specifically for this purpose, I think one could (attempt to) reword relevant parts to explain the idea differently, while still (hopefully) keeping the wording both technically correct and clear for newcomers/LifeWiki readers.

[HerscheltheHerschel]

• How to describe 'B6' when it means birth on 6 alive neighbours?
• How to describe the digit '6' in 'B36/S23'?
• How to explain the situation when two distinct CA rules have similar-looking rulestrings (e.g. B3/S23 and B36/S23)?

B6 is a birth law.
The digit 6 in B36/S23 is a law.
B3/S23 and B36/S23 have similar laws.

I like the basic idea to avoid controversial terminology (at least temporarily) altogether.
However, instead of coining a new term ('law') specifically for this purpose, I think one could (attempt to) reword relevant parts to explain the idea differently, while still (hopefully) keeping the wording both technically correct and clear for newcomers/LifeWiki readers.

How about using the term "law/condition"? This way newcomers can know that it refers to conditions while still (mostly) preventing controversial terminology.

[confocaloid]

I think 'transition' is correct and clear, when this word is used to refer to an event. For example, in a 2-state CA rule, a cell birth is an event (a state-0 cell transitions into a state-1 cell). Birth (0->1), death (1->0), survival (1->1), stagnation (0->0, "abstain") are the four possible transitions in a 2-state cellular automaton.

I think 'condition' is correct and clear, when this word is used to refer to a situation (specific environment of a cell, or a constraint on the environment of a cell, possibly including the cell itself). B6, "birth on six alive neighbours", "a cell survives when it has one orthogonal neighbour", "two diagonal neighbours opposite each other and no orthogonal neighbours", "2n" are examples of conditions.

A phrase 'transition rule' can also be used to refer to "a single row of a ruletable" or something equivalent. When one writes e.g. "a cell is born when it has three alive neighbours" or "B3", one is specifying a transition rule.

I'm not everyone, but this is basically what I think on this. And I believe the above is supported by prior existing usage (i.e. not my invented terminology).

[Haycat2009]

Transition vs condition.

Transitions are conditions put into practice like (B2a), and they affect rule behaviour.

Conditions are possible arrangements of cells like 2a, but it has no effect unless put into practice, which makes it a transition.

Now stop arguing and look at this.

[confocaloid]

Transitions are conditions put into practice like (B2a), and they affect rule behaviour.

Conditions are possible arrangements of cells like 2a, but it has no effect unless put into practice, which makes it a transition.

The whole point is that conditions are part of the rule definition. For example, the rulestring 'B3/S23' is a rule definition. It defines a rule with the following conditions:
* B3: a new cell is born when there are three alive neighbours
* S2: an alive cell survives when there are two alive neighbours
* S3: an alive cell survives when there are three alive neighbours
* in all other cases, new cells are not born, and alive cells die

When patterns evolve according to some rule (e.g. the rule defined by 'B3/S23'), cells in those patterns undergo transitions (0 -> 1, 1 -> 0, 1 -> 1, 0 -> 0). However, the rule definition itself does not change.

B2a is a condition (a new cell is born when it has two neighbours orthogonally adjacent to each other).

[confocaloid]

Someone said that I'm overusing the term 'jargon'. If that helps, here is my attempt to restate my thoughts on this terminology issue without using that term.
(Note: this post assumes there are only two cellstates, 0 = dead and 1 = alive. With three or more cellstates, 'alive', 'dead', 'birth', 'death', 'survival', 'stagnation' may need to change their meanings, or to be replaced by a different way of describing things.)

For an average wiki reader/newcomer,

• in phrases like "birth condition" or "death condition", the word 'condition' likely means something close to "a requirement that must be met in order for something (birth, death) to happen"
• in phrases like "birth transition" or "death transition", the word 'transition' likely means something close to "a change or shift from one state to another"

(At least, unless and until said reader/newcomer is brainwashed enough to believe otherwise.)

I think the plausible intuitive meanings of phrases (again, for someone who is new to the hobby) are likely to be:

• condition: how many neighbours have to be alive / which neighbours have to be alive, for a cell to be born / to survive / to die
• transition: an event of a cell going from some state to some state. Such events happen when some pattern is evolved.
• transition rule: a rule prescribing that some transition will happen whenever a certain condition is met
• birth condition: something that must be true for a dead cell (state 0) to be born (to go to state 1); part of a CA rule definition
• birth transition: an event of a cell changing from "dead" to "alive", when some pattern is evolved
• survival condition: something that must hold for an alive cell to remain alive; part of a CA rule definition
• survival transition: an event of a cell surviving (remaining alive), when some pattern is evolved
• B3 condition: the part of a CA rule definition that says "a cell is born when it has three alive neighbours"
• B3 transition: an event of a cell being born because it had three alive neighbours
• S2n condition: the part of a CA rule definition (written using Hensel notation) that says "a cell is born when it has two alive neighbours, which are diagonal neighbours opposite each other"
• S2n transition: an event of a cell surviving because the S2n condition was met for this cell

LifeWiki is supposed to collect knowledge, primarily in the form of articles (about topics that are considered notable). Articles are written primarily for newcomers / for readers who do not yet know the stuff that is explained.
On the other hand, most forum posts are written with assumption that people reading the post already know/understand most of the basic concepts and ideas and terminology.
This difference is the reason why explainations in the wiki articles have to differ from discussions in the forum posts. LifeWiki is aimed at a different (overlapping, but very different) set of people.

In this case, I believe that means that writing in a LifeWiki article something like "HighLife differs from Life in the B6 condition" is better than writing "HighLife differs from Life in the B6 transition" in the same place.
The part 'B6' means "a cell is born when it has 6 alive neighbours". This condition applies in every generation, to every cell with 6 alive neighbours. Here 'B6' is part of a rule definition, which is a "static thing" that determines evolution of patterns, but does not change itself, unless one is using an alternating rule.

(And if one is actually using an alternating rule, then things become more interesting. In addition to pattern cells transitioning from a cellstate to a cellstate according to the evolution rule, some of the conditions that are part of the evolution rule will also transition between being present and being absent.)

Likewise, it is confusing to write in a LifeWiki article "Pedestrian Life differs from Life in one transition" or "LeapLife differs from Life in two transitions". The difference between rules is measured in added/removed conditions (i.e. in modifications to the rule definition), rather than in transitions ("events of a cell being born or surviving").

---

Please note that all the above is unrelated to the challenge of documenting the way how words are used on the forums.
If one wants to document common meanings of some word in forum discussions, then the natural place to do that would be a disambiguation page for that word. That disambiguation page would mention and describe common CA-related meanings of that word, with relevant links to other pages where an interested reader can find out more information about one of listed meanings of the word.

However, if there is a common non-intuitive meaning, and a less common but more intuitive meaning, then the more intuitive meaning should be preferred, whenever the word is used in articles to explain other topics. In those situations, correctness and intuitiveness are more important than commonness. If you care about understanding, do not contradict the intuition without necessity.

[dvgrn]

It's been several months since the "transition" vs. "condition" argument ran out of steam. This seems like the thread that most succinctly collected the relevant explanations of the different definitions of "transition", along with the differing philosophies of LifeWiki explication that might make one definition or another seem to be better.

No one should feel the need to re-post anything that's already stated in this thread. However, the discussion burned itself out in such a way that nobody seems to have dared to try to clean up any of the leftover difficulties in LifeWiki articles that were in contention.

For example, this edit changed one usage of "outer-totalistic transition" to "outer-totalistic condition", but left the other use of that term unchanged. It seems like a good idea to use either one term or the other.

Not surprisingly, I still prefer "outer-totalistic transition", myself. More importantly, I'd like to fix the redlink on "transition", in that first usage. If we want newcomers to the LifeWiki to know what a "transition" is in current common usage, it seems like it would be good to have a glossary article for the term, rather than just leaving it as a mystery to be gleaned from context.

So: what should the "transition" glossary article say?

In the absence of other suggestions, I can put something together with calcyman's summary as a reference.

[confocaloid]

So: what should the "transition" glossary article say?

It cannot be an article about a single well-defined concept, because it's basically forum jargon, rather than a single well-defined universally-understood term.
Different people use the word with different meanings.

Sometimes the intended meaning is "a state-to-state transition" (birth, death, survival, stagnation).
Other times, the intended meaning is "the act of going from one rulestring to another rulestring while rulegolfing".
Other times, the intended meaning is "a shorthand for the term 'transition rule'".

By itself the word 'transition' is too ambiguous to be useful on LifeWiki, unless context makes it totally clear what is the exact intended meaning.

A consequence is that if you want to make a wiki page for this word, then it should be a disambiguation page, not an article.
Another consequence is that actual articles discussing other topics should avoid this word, unless it is necessary; it is too ambiguous to be helpful when explaining things to readers.

For example, this edit changed one usage of "outer-totalistic transition" to "outer-totalistic condition", but left the other use of that term unchanged. It seems like a good idea to use either one term or the other.

The bracketed digits are conditions. They are neither transitions nor transition rules. A digit states (part of) what is supposed to hold in the current generation, but does not specify the future states of cells.

[dvgrn]

It cannot be an article about a single well-defined concept, because it's basically forum jargon, rather than a single well-defined universally-understood term.

What would you suggest as the content of a "transition" disambiguation page?

Of the three definitions you mention, the first two don't seem to me to get used often enough to warrant a separate mention on a disambiguation page. Only the third definition is relevant for potential uses like the current OCA:HighFlock redlink, or -- just for example --for the first of the seventeen uses in Hensel notation.

I'd like to be able to link to "transition" from the "Hensel notation" article, and have the "transition" article explain what "transition" means as it is used in the "Hensel notation" article.

If there's only one primary definition in the context of the LifeWiki articles that will link to "transition", then a disambiguation page doesn't seem like the right idea. It doesn't seem useful to list every possible meaning of "transition" on a disambiguation page, because that tends to imply that those definitions are all equally likely. I don't think that that's the case here.

Disclaimer
I'm not generally a big fan of disambiguation pages in cases like this. I'd rather state one primary common meaning of "transition", and then maybe have a subsidiary "Other uses" section that lists other usages that people can look through, just in the rare cases when the primary common meaning fails to make any sense.

[confocaloid]

What would you suggest as the content of a "transition" disambiguation page?

In my previous post, I already listed three meanings to be listed on such page, if it is created. All of them did occur in past discussions. If one wants to document the word then one needs to list them all.

the first of the seventeen uses in Hensel notation

That paragraph should be rewritten to say instead

For instance, B2-a/S12 (Just Friends) indicates that a dead cell will be born with 2 neighbours, except when they are adjacent (indicated by the "-a"), and that a live cell will survive with 1 or 2 neighbours in any configuration. This exclusion of the "B2a" birth condition prevents patterns from exploding in a similar manner to patterns in Seeds.

In particular, "B2a" expresses a birth condition that is either included or excluded. It's not a transition.

If you wanted to discuss transition rules instead, then it would be incorrect to say "exclusion of the B2a transition rule", because it is a change from having the rule B2a (a dead cell gets born) to having the rule A2a (a dead cell remains dead), rather than a mere exclusion.

[wwei47]

I think it goes both ways. A birth transition is a cell turning on. A birth condition describes when a birth transition happens. They are paired together very tightly; every birth transition that ever happens happens because of a birth condition. Something like "B3a" is both a birth condition in the rulestring as well as a type of birth transition that happens when that birth condition is met.
EDIT: Wording

[confocaloid]

I think it goes both ways. A birth transition is a cell turning on. A birth condition describes when a birth transition happens. They are paired together very tightly; every birth transition that ever happens happens because of a birth condition. Something like "B3a" is both a birth condition in the rulestring as well as a type of birth transition that happens when that birth condition is met.
EDIT: Wording

Relevant: Lifeline_Volume_5#Page_2

If written as a single line in a ruletable, that would be something like

Code: Select all

neighborhood:Moore
symmetries:rotate4reflect
0, 1, 1, 1, 0, 0, 0, 0, 0, 1  # the B3a rule
0, 1, 1, 1, 0, 0, 0, 0, 0, 0  # the A3a rule
1, 1, 1, 1, 0, 0, 0, 0, 0, 0  # the D3a rule
1, 1, 1, 1, 0, 0, 0, 0, 0, 1  # the S3a rule

Although all four lines would not appear in one ruletable. (One would have at most one of B3a, A3a written explicitly; similarly at most one of D3a, S3a written explicitly.) Each line gives one rule "when such-and-such condition is met, the middle cell remains or becomes a state-N cell in the next generation".

Transitions either happen or do not happen. Conditions either are met or are not met. Rules (= transition rules) either are followed or are not followed.
I think that is fairly clear and intuitive.

[bubblegum]

That paragraph should be rewritten to say instead...

I think that article should continue to refer to B2a as a transition.

[confocaloid]

That paragraph should be rewritten to say instead...

I think that article should continue to refer to B2a as a transition.

Why? Do you think that's particularly intuitive or readable, from the viewpoint of a LifeWiki reader / beginner?

The word 'transition' is supposed to be a word for something that happens. How do you "exclude" "something that happens", exactly? What would that mean?

Two-state cellular automata can be defined by sets of birth/survival conditions. Those conditions can be included or excluded. That is intuitive and agrees with the general ideas of sets and conditions.

When patterns evolve according to some ruleset, individual cells follow the rules, which are usually of the form "when such-and-such condition is met, the middle cell goes to such-and-such state or remains in that state". Again, that can be understood intuitively on first reading.

[silversmith]

I don't think I have seen mentioned the fact that websites like Lifewiki are generally descriptive, rather than prescriptive (eg. citing other sources for info). Currently, "transition" is by far the default way to describe rules. I agree that "condition" makes more sense, but as long as the purpose of the Lifiwiki is to describe the current state of CA exploration, replacing "transition" with "condition" would make it a less accurate description.

Unless "condition" is used regularly by a non-trivial portion of the community, I believe Lifiwiki should continue match the community's use of "transition".

[confocaloid]

I don't think I have seen mentioned the fact that websites like Lifewiki are generally descriptive, rather than prescriptive (eg. citing other sources for info).

The problem with this is that the words/phrases are not synonyms. 'Condition', 'transition', 'transition rule' are all used, with different meanings.

Therefore the question is not simply "which makes more sense", but "what each term actually means". One should not replace one with another merely on the basis of the number of occurrences in a text search.

Further, bringing everything said on the forums to the wiki would make the wiki into useless mess. The wiki should be helpful (i.e. it should be easy to understand the intended meaning when reading a page). A reader should not be forced into back-and-forth jumping between different pages with subtly different definitions (or even sometimes lacking any meaningful definitions).

LifeWiki:Style guide says: "The purpose of LifeWiki is to provide encyclopedia-style content describing Conway's Game of Life and similar cellular automata. All articles should be informative and targeted at the site's audience. Information that is only of interest to the writer or to other editors should not be included in articles. Articles should not assume knowledge that someone with a basic understanding of the Game of Life would not possess; relevant terminology should be explained or linked to."

[silversmith]

The passage in the LifeWiki Style Guide does in fact mean that my "descriptive, not prescriptive" statement does not apply to the LifeWiki.

Which, in turn, implies confocaloid is completely correct in arguing that LifeWiki articles should be edited for ease of understanding, even if it means replacing "transition" with "condition" wherever it would make the meaning more intuitive.

[confocaloid]

I'll add that the wider Life/CA community does in fact speak about conditions. (Conditions for a cell to be born in the next generation; conditions for a cell to survive; conditions for a cell to be alive in the next generation; in general, conditions for something to happen -- the underlying idea remains the same.)

I already posted many quotes, and there are many more CA-related texts and CA-related discussions that are relevant.
viewtopic.php?p=171596#p171596
viewtopic.php?p=171671#p171671
viewtopic.php?p=172091#p172091

The word 'condition' is a word for something that must be met before something happens.
The word 'transition' is a word for something that happens.
The phrase 'transition rule' (or simply 'rule') means some way to prescribe when something happens.

(Examples of conditions: a cell is currently alive; a cell currently has exactly three alive neighbours; a cell is currently alive and has exactly two alive neighbours that are side-to-side adjacent to each other.
Examples of transitions: birth; death; survival.
An example of a rule: if a cell currently has exactly three alive neighbours, then the cell will be alive in the next generation.
A rule could be represented as a pair (condition; transition).)

That can be understood intuitively, even before the reader gets to the definitions. When you're writing for other people to read, do not contradict intuition (especially considering that there are many texts and discussions that in fact follow the same ideas and use the same words and phrases with the same meanings).

[dvgrn]

... confocaloid is completely correct in arguing that LifeWiki articles should be edited for ease of understanding, even if it means replacing "transition" with "condition" wherever it would make the meaning more intuitive.

Sure, I can agree completely with that -- nothing wrong with ease of understanding!

So now the question is: are there in fact any uses of "transition" on the LifeWiki, where the article's meaning would be more intuitive if they were replaced with "condition"?

I haven't found any instances where the change seems like an improvement to me. So I'm currently planning to try some other options to improve ease of understanding.

The thing is, the community's current common usage of the word "transition" seems perfectly understandable already. The term has a very good clear unambiguous technical definition, as given in calcyman's summary.

It seems like it's eminently worth explaining that definition clearly on the LifeWiki, so that any hypothetical people who don't understand that sense of "transition" can click on the link and see the definition.

It's true that calcyman's technical definition doesn't match the 'intuitive' definition...
It's absolutely true that "transition" in that particular CA context, referring to an outer-totalistic rulestring element or an isotropic rulestring element or a MAP rulestring element or what have you, doesn't have exactly the same meaning as a standard dictionary definition --

"the process or a period of changing from one state or condition to another"

That's perfectly okay, though! Technical definitions very often don't match dictionary definitions.

... and confocaloid's proposed definition also doesn't match the 'intuitive' definition
confocaloid's proposed definition of "transition" doesn't match the dictionary definition all that well either. "Survive" and "Abstain" are claimed to be transitions, but those don't actually involve changing from one state to another. The state just stays the same -- no transition is happening at all in those two cases.

So ... who is going to make the judgment as to which of these technical definitions is closer to the common meaning and therefore "more intuitive"?

-- Or does that even actually matter? Isn't it more important to define commonly-used technical terms clearly, so that newcomers can understand existing technical conversations, and participate in new conversations whenever they're ready to do that? How well is that going to work if the LifeWiki uses a completely distinct vocabulary from the predominant usage on the forums and Discord?

[confocaloid]

So now the question is: are there in fact any uses of "transition" on the LifeWiki, where the article's meaning would be more intuitive if they were replaced with "condition"?

OCA:tlife and Hensel notation are two examples of pages, where the word 'transition' is currently used in an incorrect and confusing way, where the described thing is a condition, rather than a transition.

The thing is, the community's current common usage of the word "transition" seems perfectly understandable already. The term has a very good clear unambiguous technical definition, as given in calcyman's summary.

That's counterintuitive and confusing for LifeWiki readers. A cellular automaton is defined by listing rules. Each line of a ruletable defines a rule.
Transitions are events that happen -- not rules that prescribe when some event will happen.

confocaloid's proposed definition of "transition" doesn't match the dictionary definition all that well either. "Survive" and "Abstain" are claimed to be transitions, but those don't actually involve changing from one state to another. The state just stays the same -- no transition is happening at all in those two cases.

I disagree. A transition is an event, something that happens. There's no requirement of change.
Survival is an event that happens: an alive cell remains alive in the next generation.
Likewise, stagnation is an event that happens: a dead cell remains dead in the next generation.

Isn't it more important to define commonly-used technical terms clearly [...]?

The word 'transition', as used on these forums, is local forum jargon. It's not a technical term, and doesn't have a single well-defined universal meaning. Different people mean different things in different contexts when they use the word. Therefore, there should be no such article you're suggesting.

Again, if you want to document forum jargon, then there should be a disambiguation page listing existing meanings of the word or phrase.

[dvgrn]

The word 'transition', as used on these forums, is local forum jargon. It's not a technical term, and doesn't have a single well-defined universal meaning. Different people mean different things in different contexts when they use the word.

I don't see any evidence at all of any of these statements. Every time that a phrase such as " 'B3j' transition " is used, the term "transition" always means exactly the same thing -- the meaning that calcyman summarized. That's a valid technical definition of the term.

How would you suggest that the decision should be made, about which technical definition of "transition" should be documented on the LifeWiki?