Apple Bottom wrote:I must admit I'm not entirely convinced, in light of what Mark wrote there...
... using Mark's terminology at least (which is probably as authoritative as it gets), traffic lights are NOT, in fact, constellations, but rather a "formation" of oscillators, to use the LifeWiki's term (and a "pattern", of course, but that goes without saying).
Mark also replies to Calcyman's assertion (that quasi-objects are to be considered constellations) saying that "
Under existing definitions, that do not distinguish quasi-objects, this is true" (emphasis mine) -- which I would understand to mean that under newer/better/more accurate definitions, it is not true.
Hmm. It seems to me you might be getting a little too theoretical there, with your hypothetical newer/better/more accurate definitions.
I'd say the Lexicon
should not even try to limit itself to the one best possible definition of every term. Its job is to document common usage, not to be a mathematically perfect textbook.
So here's my chain of crackpot logic:
Consider, say, two blocks sharing a diagonal neighbor cell.
It's
much more common usage to label those two blocks as a "constellation", than to say "no they're not, they're a single quasi still life!" Most people probably don't even know what quasi still lifes are -- those only made it into the Lexicon because we had some nice statistics about them! -- so that second option can't be common usage yet.
Therefore the Lexicon definition of "constellation" has to be allowed to overlap with "quasi still life".
Otherwise we'd be trying to legislate a significant restriction to the term "constellation". And since the restriction would be for what looks like boring technical reasons, quite justifiably nobody will pay any attention -- everyone will keep using "constellation" to mean things like those two blocks, whenever they darn well please. And then the Lexicon will have to say annoying pedantic things like "many people use {constellation} when they're talking about {quasi still life}s".
Seems to me that's all just asking for trouble.
I don't think it even makes sense for the Lexicon to mention the possibility that "constellation" might be considered to exclude quasi still lifes. From a classification point of view, that use of "constellation" is certainly cleaner, but it's just plain not common usage at the moment. I figure if the concept turns out to be needed, someone will start referring to it as a "strict constellation" or whatever, and people will start using that, and maybe it will get into the 2030 edition of the Lexicon...!
One more quote from Mark 39453:12, while we're at it:
mniemiec wrote:(Originally, pseudo-objects were also considered constellations. In fact, apgsearch currently still treats them as such.)
This "original" usage seems like another reason why :constellation: should be documented as a simple all-inclusive term. If you can see that a pattern is made up of a bunch of blocks and beehives and blinkers and so forth, you can safely call it a constellation. Even if the pieces are kind of close together, you don't have to stop and look to see exactly how close they are.
Apple Bottom wrote:(Do you ever feel like you're drowning in terminology?)
Nah. Well, maybe a little, but as soon as the feeling starts I just say to myself, "Eh, who really cares?" -- and that floats me right back to the surface again.
EDIT: All of this is just my opinion, of course, and as the previous paragraph implies it's not a terribly strongly held opinion. Sufficient objections will certainly change my mind -- sufficient either in quality or in quantity, I don't need both... At the moment I'm thinking maybe the following adjustment would be good:
For this reason a stricter definition is often used, counting a stable pattern as a {strict still life} only if its {island}s cannot be divided into two or more nonempty sets both of which are stable in their own right. If such a subdivision can be made, the pattern can referred to as a {constellation}. If its cells form a single {cluster} it is also, more specifically, either a {pseudo still life} or a {quasi still life}.
I think that makes a clear and consistent definition that aligns reasonably well with common usage. Alternative suggestions are welcome, especially if they include complete specific wording for any and all related definitions that would have to be changed, like for
:constellation:...!