Suggested patterns for Golly

[Andrew]

I've noticed that comments in the pattern files seem to use three different formats for paragraphs:

1. Blank line before each paragraph:

Code: Select all

#C 
#C This is a paragraph.
#C This is the next line in this paragraph.
#C This is the last line in this paragraph.
#C 
#C This is another paragraph.
#C This is the next line in this paragraph.
#C This is the last line in this paragraph.

2. Indentation on the 1st line of a paragraph:

Code: Select all

#C    This is a paragraph.
#C This is the next line in this paragraph.
#C This is the last line in this paragraph.
#C    This is another paragraph.
#C This is the next line in this paragraph.
#C This is the last line in this paragraph.

3. Indentation on all lines of a paragraph except the 1st:

Code: Select all

#C This is a paragraph.
#C   This is the next line in this paragraph.
#C   This is the last line in this paragraph.
#C This is another paragraph.
#C   This is the next line in this paragraph.
#C   This is the last line in this paragraph.

We seem to be moving to #3 as the preferred format. In my eyes it's the ugliest of the 3 formats. Is there some reason for that weird indentation? Format #1 would be my preference (easiest to read, and to write).

[dvgrn]

I've noticed that comments in the pattern files seem to use three different formats for paragraphs:

1. Blank line before each paragraph...
2. Indentation on the 1st line of a paragraph...
3. Indentation on all lines of a paragraph except the 1st...

We seem to be moving to #3 as the preferred format. In my eyes it's the ugliest of the 3 formats. Is there some reason for that weird indentation? Format #1 would be my preference (easiest to read, and to write).

Well... there's definitely a reason. It just might not be a _good_ reason.

The reason is simply that I do pretty much all of the checkins for Golly's pattern collection, and for a while now I've been putting everything in format #3, with hanging-indent paragraphs. Definitely this is just a personal preference on my part, and not a particularly strong personal preference.

Mostly I think that hanging indents solved a problem I was having with deciding whether to include blank lines in comments. Recent comments almost always have a "Reference" section at the end with a link to a relevant location on the forums, and/or to a LifeWiki page. If "Reference:" goes on the same line as a forums URL, then the line is a little bit longer than all the other comment lines. So I've been moving the URL to its own line.

Then I can never decide whether blank #C lines are needed before and after that "Reference" section. For the paragraph format separated by blank #C lines, seems cleanest to include them, like in this re-work of muzik's latest pattern:

Code: Select all

#N noninteger-replication-order.rle
#O wwei23, 23 October 2019
#C
#C This anomalous fractal does not appear to contain an integer
#C count of sub-units at any scale, unlike the other fractals we've
#C seen so far.
#C
#C Reference:
#C https://conwaylife.com/forums/viewtopic.php?p=84439#p84439
#C
x = 29, y = 24, rule = B2e3aiky4jnqtwz5cjqry6aei/S1c2-k3ajry4cjkrt5-eqr6i
7b3o9b3o$8bo11bo$8bo11bo$7b3o9b3o$12bo3bo$11bobobobo$11bobobobo$12bo3b
o7$8bo11bo$7b3o9b3o$7bobo9bobo2$12bo3bo3$b2o23b2o$o2bo21bo2bo$b2o8b2o
3b2o8b2o!

So I probably chose hanging indents just because I could safely get rid of all of the empty-and-yet-not-empty #C lines. More actual comments will fit in the pop-up window in Golly without any scrolling, and yet it's still clear where each new paragraph starts.

So -- long story short, it's an arbitrary decision on my part, and I don't really care what the standard is, only that there _is_ a clear standard. If it's important enough to enough people that comment formatting should be done in some different way, then I don't mind switching to format #1 for future check-ins.

Are there more people that dislike the hanging-indent standard, enough to say so here? Or, even better, is there anyone who dislikes the hanging-indent so much that they're willing to go through the whole pattern collection and change the comments so they're in some other consistent format?

[muzik]

Are there more people that dislike the hanging-indent standard, enough to say so here? Or, even better, is there anyone who dislikes the hanging-indent so much that they're willing to go through the whole pattern collection and change the comments so they're in some other consistent format?

I could try this once an actual beta version is released.

[Andrew]

Or, even better, is there anyone who dislikes the hanging-indent so much that they're willing to go through the whole pattern collection and change the comments so they're in some other consistent format?

I'm happy to do that for the 4.3b1 release. (Not sure if you've seen my latest emails -- have you finished pushing all your commits?)

[b-engine]

Code: Select all

#N 2c1042-self-constructing-spaceship.rle
#O b-engine, 4 March
#C This is the first self-constructing spaceship in B-Univ, demonstrating the capabilities of universal construction in the rule.
#C Before the spaceship is engineered the only other known spaceships in B-Univ are photons made of state 3 cells.
#C The top speed of the spaceship is 2c/1042.
#C Reference: https://conwaylife.com/forums/viewtopic.php?f=11&t=6375#p179743
x = 521, y = 15, rule = B-Univ
AGCAGCAGCADCAGCAGCAGCAGCAGCAGCAGCAGCADCAGCADCAFCAFCAFCAGCADCAGCADCAFC
AGCAGCADCAFCAFCAFCAFCAGCAGCAGCAGCADCAGCADCAFCAGCAGCADCAFCAFCAFCAFCAGC
ADCAGCAGCAGCABCADCAFCAFCAFCAFCAGCAGCAGCAGCAGCADCAGCADCAFCAFCAFCAGCADC
AGCADCAFCAGCAGCADCAFCAFCAFCAFCAGCAGCAGCAGCADCAGCADCAFCAGCAGCADCAFCAFC
AFCAFCAGCADCAGCAGCAGCABCADCAFCAFCAFCAFCAGCAGCADCAGCAGCABCADCAFCAFCAFC
AFCAFCAFCAFCAFCAFCAFCAFCAFCAFCAFCAFCAFCAFCAFCAFCAFC124A$.A517.A$.519A
2$.346ACFACFACFACFACFACFACFACFACFACFACFACFACDACBACGACBACGACFACFACFACF
ACFACFACBACDACFACBACDACGACGACGACBACGACGACGACGACGACFACFACFACBACDACFACB
ACDACGACGACGACBACGACGACGACGACBACGACGACGACGA$.A517.A$.519A2$519A$A517.
A$CFACFACFACFACFACFACFACFACFACFACFACFACDACBACGACDACGACFACFACFACFACFAC
FACBACDACFACBACDACGACGACGACDACGACGACGACGACGACFACFACFACBACDACFACBACDAC
GACGACGACDACGACGACGACGACDACGACGACGACG346A2$519A$A517.A$CFACFACFACFACF
ACFACFACFACFACFACFACFACFACFACFACFACFACFACFACFACDACBACGACGACBACGACGACF
ACFACFACFACDACBACGACGACGACBACGACFACFACFACFACDACGACGACFACDACGACBACGACG
ACGACGACFACFACFACFACDACGACGACFACDACGACBACGACFACFACFACDACGACBACGACGACG
ACGACGACFACFACFACFACDACBACGACGACGACBACGACFACFACFACFACDACGACGACFACDACG
ACBACGACGACGACGACFACFACFACFACDACGACGACFACDACGACBACGACFACFACFACDACGACB
ACGACGACGACGACGACGACGACGACBACGACGACG124A!

[confocaloid]

Replying to some earlier posts.

(See also my suggestion to merge the new pattern b3578s23-failed-2d-replicator.rle into pattern comments of b3578s23-2d-replicator.rle and expand pattern comments to mention both "almost-replicators" there. I don't believe the "almost-replicators" are sufficiently notable by themselves to warrant inclusion as separate patterns. They still can go into pattern comments for the working replicator in the same rule, in the same way how there is just one page "Replicators: B3578/S23" mentioning all of them.)

Would it make sense to include a Replicators folder in Life-Like as well? The presence of b3578s23-2d-replicator.rle, b3578s23-failed-2d-replicator.rle, replicator.rle, sqrt-replicator.rle and triangular-rule-150-replicator.rle might be enough to secure such a spot, especially since four out of five of those are completely new pattern files for 4.3.
[...]
Also, with this folder, we may as well go all in with the 2D parity rules: [...]

One of mentioned patterns is in a triangular rule, and several other patterns suggested down in the post (i.e. von Neumann neighbourhood, hexagonal neighbourhood, tripod neighbourhood) are also in rules that aren't Life-like cellular automata. Hence such a folder should not go into "Life-Like".

Is the idea to have just one subfolder named "Replicators" for all notable replicators in various CA rules? If so, then I think it should go to "Other-Rules" instead.

For the same reason, the new pattern "von-neumann-merge.rle" does not belong in "Life-Like" folder (the rule B2/S23V uses von Neumann neighbourhood). I suggest to move that pattern into the "Other-Rules" folder (which already contains patterns working in rules with von Neumann neighbourhood).

On the topic of creating new directories: Larger-than-Life might benefit from a spaceships folder, [...]

o After moving the bugs into a subfolder, how many patterns remain in "Larger-Than-Life" (and not a subfolder)? Moving most patterns into a nested subfolder does not seem to be an improvement over leaving them all in a single folder.
o I think related patterns "BugCollection.rle", "GunCollection.rle", "BugReactions.rle", "BugGun.rle" should all remain in the same folder.

A different possibility is to group patterns by the CA rule, rather than by the pattern type. I think grouping by the rule is preferable.
For example, the several related patterns in Bosco's Rule would remain all in a single folder, making it easy to explore that rule (without having to constantly switch back and forth between different subfolders).

[erictom333]

<snip>
One of mentioned patterns is in a triangular rule, and several other patterns suggested down in the post (i.e. von Neumann neighbourhood, hexagonal neighbourhood, tripod neighbourhood) are also in rules that aren't Life-like cellular automata. Hence such a folder should not go into "Life-Like".

Is the idea to have just one subfolder named "Replicators" for all notable replicators in various CA rules? If so, then I think it should go to "Other-Rules" instead.

For the same reason, the new pattern "von-neumann-merge.rle" does not belong in "Life-Like" folder (the rule B2/S23V uses von Neumann neighbourhood). I suggest to move that pattern into the "Other-Rules" folder (which already contains patterns working in rules with von Neumann neighbourhood).
<snip>

I have a suggestion: Create a new "Other-Neighbourhoods" folder for Life-like and non-totalistic range-1 rules in the hexagonal, von Neumann and tripod neighborhoods. This should also ideally come with support for non-totalistic hexagonal and von Neumann rules, and a collection of patterns from such rules.

[confocaloid]

[...]
I have a suggestion: Create a new "Other-Neighbourhoods" folder for Life-like and non-totalistic range-1 rules in the hexagonal, von Neumann and tripod neighborhoods. This should also ideally come with support for non-totalistic hexagonal and von Neumann rules, and a collection of patterns from such rules.

Currently:

• there is already a dedicated folder "Patterns/Life-Like" for patterns in Life-like cellular automata. (Two cellstates, square grid, range 1 Moore neighbourhood, outer-totalistic rules.)
• there is also already a folder "Patterns/Non-Totalistic", mostly containing patterns in two-state isotropic CA on the square grid with range-1 Moore neighbourhood.
• there is also a folder "Patterns/Larger-than-Life".

Maybe "Other-Neighbourhoods" would cover any other neighbourhoods (specifically excluding two-state range-1 Moore neighbourhood, which is already covered by "Life-Like" and "Non-Totalistic", and also excluding LtL)?
Or maybe it would cover any other notated neighbourhoods (i.e. with support for a dedicated rulestring notation)?

Where this folder should go, then? A natural place seems to be "Patterns/Other-Rules". The new pattern "von-neumann-merge.rle" would also go in that subfolder (or directly into "Patterns/Other-Rules").