ConwayLife.com

Posted: **April 4th, 2025, 8:17 am**

dvgrn wrote: April 4th, 2025, 7:51 am
confocaloid wrote: April 4th, 2025, 4:01 am I disagree. That edit summary isn't really accurate.

First, there is currently no "standard" on this that would be specific to LifeWiki.
I have now taken the time to click through and look at all eighty-some guns in the LifeWiki “Guns” category, one at a time.

As you say, there definitely is no consistently applied standard that is maintained across all of the gun pages at the moment. The form of each gun is an individual choice made by each article’s author — and those choices are all over the map. There are guns with gliders well outside the bounding box, slightly outside the bounding box, well inside the bounding box, and just after the glider first appears. Occasionally an RLE got copied from the gun collection so it’s in canonical form — especially for guns in the mid-double digits, for some reason.

But more often, no output is shown at all. So that edit summary still seems accurate to me. There is a LifeWiki standard that I think is implied by the no-output patterns.

A sometimes-followed LifeWiki standard
When the LifeWiki displays and describes a pattern, it’s trying to be accurate and not confusing. Gun patterns without outputs are the miminal form that can accurately be labeled a “gun”. Output gliders are certainly handy for some purposes, but they’re not part of the gun, and they’re easy to obtain by running the pattern.

I haven’t surveyed all the no-output guns to this level of detail, but it looks like they’re very often in a low-population phase.

Check the period-690 glider gun article, for example. I find the chosen phase to be pretty irritating, really, because it takes so long for the glider to escape that I think I must have missed it! However, the infobox lists a population of 62 — and the T=0 form shown has a population of 62.

That was a deliberate choice, made to align with a LifeWiki-specific standard. It shows the form that we need to know about to correctly fill in the infobox population parameter.

Gun-collection canonical form, by contrast, has some number of extra gliders included, bulking up the population. Those patterns usually aren’t the minimal form of the gun at all— there’s a widely varying number of extra gliders, plus an arbitrary-looking non-minpop phase of the gun that happens to align a glider with the gun’s all-phases bounding box — which may not match the current phase’s bounding box. It’s a complicated standard. It might be better to stick with a simple minpop standard on the LifeWiki.

The minpop state of a gun often doesn’t give much hint about where the output will appear, but it often does a good job of showcasing the active objects that will interact to create the output.

Long story short
There are advantages to both of these competing standards. If we want to go around standardizing gun articles, it’s not clear to me why a non-LifeWiki standard should necessarily be chosen over an existing LifeWiki-specific standard— given that neither standard has been followed particularly well to date!

Next steps
If someone wants to take on this gun standardization problem as an editing project, I’d suggest starting a discussion thread first to see if there’s a consensus about which standard we should consistently use. Otherwise this is in the category where it’s easy to get into accidental edit wars, because there are perfectly decent reasons to make changes to individual articles in either direction.

Unless we’re going to decide to pick one standard and stick with it, there’s no strong reason either to make changes (in either direction) to any one individual gun page, or to undo such a change once it is made.

This thread aims to address the issue of non-standardized gun RLEs on the Wiki. Discussion about this issue can be done here.

(sorry about violating Long Quote, Short Reply)

EDIT: Here is more on the discussion.

confocaloid wrote: March 26th, 2025, 2:28 pm https://conwaylife.com/w/index.php?diff ... did=159486
Not an improvement. Removing output gliders from a glider gun (which was previously shown in the canonical form, as far as I can tell).

Glider guns should be shown in the canonical form. See the readme file:
https://entropymine.com/jason/life/dpguns/readme.txt
https://entropymine.com/jason/life/dpguns/

EDIT by dvgrn: Reviewed this item. WhiteHawk's edit summary was

It's more standard on the wiki to show the gun without any emitted gliders (see p15, p16, p60 guns, GGG, etc.

which seems accurate. As far as I can recall, it has been quite seldom that we've used gun-collection canonical form for gun patterns in LifeWiki infoboxes. That has always obscurely annoyed me, but it is what it is; the README from Dieter and Peter's gun collection isn't particularly relevant.

The gallery items for the p14 glider gun article definitely are in canonical form, but that doesn't seem to set a precedent for the infobox -- where, for that particular article, the LifeViewer display was disabled quite a long time ago, and was only restored recently when the p14 gun was reduced to a reasonable size so that it worked to display it in the infobox after all.

confocaloid wrote: April 4th, 2025, 4:01 am
post208236 wrote: March 26th, 2025, 2:28 pm https://conwaylife.com/w/index.php?diff ... did=159486
Not an improvement. Removing output gliders from a glider gun (which was previously shown in the canonical form, as far as I can tell).
[...]
EDIT by dvgrn: Reviewed this item. WhiteHawk's edit summary was

It's more standard on the wiki to show the gun without any emitted gliders (see p15, p16, p60 guns, GGG, etc.

which seems accurate. [...]
I disagree. That edit summary isn't really accurate.

First, there is currently no "standard" on this that would be specific to LifeWiki.
There is readme.txt from the old glider gun collection, and people continued and continue to follow the same guidelines for glider guns in other contexts.
Other than those guidelines from Dieter and Peter's gun collection, I'm not aware of anything else relevant to the question that would be anywhere close to a "standard".

Second, showing the gun without any emitted gliders leads to particularly awkward display in this case:
Code: Select all
x = 72, y = 64, rule = B3/S23
39b2o$38bo2bob2o$37bo2b2obobo$37b2obo4bo$38bo2b2o2bob2o$36bo2bobo2b2o
2bo$35bob3obobo2bo$26bo9bo5bob2o$26b3o8b3o$4b2o23bo10b6o$4bobo4b2o3b2o
6b4o2bo4b3ob2o4bo$7bo3b2o2bobo6bo3b2o4bo3bobob2o$2o3b2obo6bobob2ob2o3b
o4bo2b4obobo$bo2bobob7obobobobo5b6obobo2bo3bo$o3bo2bo8bo3bo2bo4bo6bo2b
o2b4o$4obo2b6o2b2ob6o5b5o3b2o$5b3o4bo2b3obo4bo8bo5bob2o$2o3bobobo4bo5b
ob2o10b2obobobobo$o4bobobob10obo4bo6bo3bobo2bo$b3ob2o2bo12bo3bobo5bo3b
obobo$4bo4b4o2b2o2b2obob3obo4b5o3bo$b2obo7bo2bo4bo2bo4b2o2b5o$bobo7bo
4bo3bo3b4o2bo5bo$10bob4o2b2ob3o4b2o5bo$11bo4b2obobo3b2obobo3b2o3b2o$12b
3obo9bobo2b2obo3bo2bo$14bo2b4o4bo2bo4bo4b3o$15b2o4bo4b3ob3obo$17b5o7b
ob2o3b3o$17bob4o3b2obo4b2o3bo$4b2o8b2obo5bobobob2o4bo3bo$5bo8bo2bo2bo
2bobo3bo2b2o3bob2o$5bobo7b2o2bobobo2bo2bo6b2o3bo$6b3ob2o5b5ob3o2b2obo
6b3o3b2o$8b2obo5bo4bo7b2obo4bo5bo2b2o$8b3o7b3obob6o2b2o11bobo$4b2o3bo
10bobobo9b2o8b2obo2b2o10b2o$4bo2bobo17b2obo2bobo7bo3bobobo10bobob2o$6b
2obo2bo8b3o3bo4bo2bo6b3ob2obo14bobobob2o$10b2o16bo3bo2b2ob2o23b2o3bob
o$29b2ob2o6bo5b2o4b2ob2o7bo3bo2bo$22bo9b2o3bo2bo4b3o3b6o2b2o2bo5b2o$21b
3o13bo2bo3bo13bo5b5o$11b2o7b5o12b2obo3b2o4bobo2bobo3bo6bo$11b2o6b2o3b
2o12bobo3bo3bobob2obob11o$10bo8b2ob3o13bobobobobobobo14bo$9b3o8b4o15b
obob2ob2o2bobob9o$19bo2bo18bo3bo4bob2o5bo2bo$40b2o3bob3o5b2o$14b3o7bo
21b2o8bo$15bo7bobo28bo$13b2o9bo29b2o$13b2o4$14b2o$14bobo$15b2o2$12b2o
$13bo$10b3o$10bo!
The previous revision of the RLE in question ( https://conwaylife.com/w/index.php?titl ... did=159486 ) was the canonical form, and it was much easier to see where are the actual output gliders:
Code: Select all
x = 72, y = 64, rule = B3/S23
39b2o$38bo2bob2o$37bo2b2obobo$37b2obo4bo$38bo2b2o2bob2o$36bo2bobo2b2o
2bo$35bob3obobo2bo$26bo9bo5bob2o$26b3o8b3o$4b2o23bo10b6o$4bobo4b2o3b2o
6b4o2bo4b3ob2o4bo$7bo3b2o2bobo6bo3b2o4bo3bobob2o$2o3b2obo6bobob2ob2o3b
o4bo2b4obobo$bo2bobob7obobobobo5b6obobo2bo3bo$o3bo2bo8bo3bo2bo4bo6bo2b
o2b4o$4obo2b6o2b2ob6o5b5o3b2o$5b3o4bo2b3obo4bo8bo5bob2o$2o3bobobo4bo5b
ob2o10b2obobobobo$o4bobobob10obo4bo6bo3bobo2bo$b3ob2o2bo12bo3bobo5bo3b
obobo$4bo4b4o2b2o2b2obob3obo4b5o3bo$b2obo7bo2bo4bo2bo4b2o2b5o$bobo7bo
4bo3bo3b4o2bo5bo$10bob4o2b2ob3o4b2o5bo$11bo4b2obobo3b2obobo3b2o3b2o$12b
3obo9bobo2b2obo3bo2bo$14bo2b4o4bo2bo4bo4b3o$15b2o4bo4b3ob3obo$17b5o7b
ob2o3b3o$17bob4o3b2obo4b2o3bo$4b2o8b2obo5bobobob2o4bo3bo$5bo8bo2bo2bo
2bobo3bo2b2o3bob2o$5bobo7b2o2bobobo2bo2bo6b2o3bo$6b3ob2o5b5ob3o2b2obo
6b3o3b2o$8b2obo5bo4bo7b2obo4bo5bo2b2o$8b3o7b3obob6o2b2o11bobo$4b2o3bo
10bobobo9b2o8b2obo2b2o10b2o$4bo2bobo17b2obo2bobo7bo3bobobo10bobob2o$6b
2obo2bo8b3o3bo4bo2bo6b3ob2obo14bobobob2o$10b2o16bo3bo2b2ob2o23b2o3bob
o$29b2ob2o6bo5b2o4b2ob2o7bo3bo2bo$22bo9b2o3bo2bo4b3o3b6o2b2o2bo5b2o$21b
3o13bo2bo3bo13bo5b5o$11b2o7b5o12b2obo3b2o4bobo2bobo3bo6bo$11b2o6b2o3b
2o4bo7bobo3bo3bobob2obob11o$10bo8b2ob3o3bobo7bobobobobobobo14bo$9b3o8b
4o5b2o8bobob2ob2o2bobob9o$19bo2bo10bo7bo3bo4bob2o5bo2bo$34bo5b2o3bob3o
5b2o$14b3o7bo7b3o11b2o8bo$15bo7bobo28bo$13b2o9bo12bo16b2o$13b2o20bobo
$36b2o$40bo$41bo$14b2o23b3o$14bobo$15b2o27bo$42bobo$12b2o29b2o$13bo33b
o$10b3o35bo$10bo35b3o!
This is why I wrote "Not an improvement. Removing output gliders from a glider gun (which was previously shown in the canonical form, as far as I can tell)."

Removing the gliders fails to make a better display, and also changes the display away from the canonical form of the gun.

Posted: **April 4th, 2025, 9:08 am**

WhiteHawk wrote: April 4th, 2025, 8:17 am [...] (sorry about violating Long Quote, Short Reply)

It isn't even about "long quote short reply". It is more about quoting one side of the argument, and failing to take into account objections, earlier discussions, sources, etc.

The edit summary isn't accurate. Despite the claim made in the edit summary, there is currently no standard on the wiki, other than the old traditional approach of measuring guns by the bounding box area and putting them into the canonical phase and orientation.

confocaloid wrote: April 4th, 2025, 4:01 am [...] there is currently no "standard" on this that would be specific to LifeWiki.
There is readme.txt from the old glider gun collection, and people continued and continue to follow the same guidelines for glider guns in other contexts.
Other than those guidelines from Dieter and Peter's gun collection, I'm not aware of anything else relevant to the question that would be anywhere close to a "standard".
[...]

confocaloid wrote: April 4th, 2025, 8:33 am [...] glider guns are traditionally measured / reduced by the bounding box area (rather than population). The readme.txt file linked above explains good reasons for that preference. Various people on the forums also repeatedly explained reasons for measuring guns by the bounding box area. Catagolue also measures glider guns by the bounding box area. If there is anything close to a standard at all, then it is this approach (measure by the bounding box area, and put the gun into the form where the output glider is about to leave the bounding box; more details can be found in readme.txt).

It is also significantly easier to manually double-check that the gun is in the canonical form, compared to double-checking that the population is stated correctly. (Especially for high-period guns.)
[...]

Posted: **April 4th, 2025, 2:29 pm**

confocaloid wrote: April 4th, 2025, 8:33 am The first question is: where that "a LifeWiki-specific standard" was previously documented? Maybe I simply forgot about its existence, but I don't remember seeing where that was written.
To me, it looks like there is simply no such "standard".

I certainly don't know of any current clear written LifeWiki standard for guns. The point of this thread is to find out if it's possible for the community to agree on a standard along those lines -- and then, presumably, document it!

That said -- quite a number of people have deliberately chosen a minpop phase as the LifeWiki-canonical phase of a glider gun, such that that's the phase that is displayed in many gun articles' infoboxes. Editors' previous collective decisions are a valid type of de facto documentation.

Guns in infoboxes seem to show a minpop phase more often than they show a gun-collection canonical phase -- though I've only done an informal survey, not a thorough census as yet, so I don't have final statistics on this!

Posted: **April 4th, 2025, 2:30 pm**

confocaloid wrote: April 4th, 2025, 9:08 am The edit summary isn't accurate. Despite the claim made in the edit summary, there is currently no standard on the wiki, other than the old traditional approach of measuring guns by the bounding box area and putting them into the canonical phase and orientation.

It seems like the edit summary provided some perfectly valid information about what is currently more common in LifeWiki gun articles. It gave several specific examples. The old traditional approach that you mention simply doesn't seem to be a very well established tradition for LifeWiki infoboxes. It seems like there are valid reasons why the LifeWiki hasn't picked up this tradition.

However, maybe there are also good reasons to switch to gun-collection canonical phases after all! It's not clear which reasons are more important. Hence this discussion.

@confocaloid, here's the key detail that I haven't heard your opinion about yet:

What do you suggest doing, to check in with the rest of the community about whether gun-collection canonical phase should become the official LifeWiki infobox standard?

I suggested opening this discussion, and then waiting to see what other people think. This is my best effort at avoiding unnecessary future complaints and edit wars.

Up to a point, I agree that it is indeed a little bit difficult to figure out what the minpop is for any given gun. Still, that's what people seem to have done, in many cases, as part of the process of documenting these guns for LifeWiki articles.

The minpop phase has to be figured out anyway, to populate the "c" parameter in the infoboxes. Once that is done, it's really not particularly difficult to use that phase as the LifeWiki standard phase -- if we collectively decide here that we want to document the minpop phase as the LifeWiki standard.

Possibilities
If a lot of people would prefer to use the gun-collection canonical phase going forward, and there's no similar number of people who prefer the minpop phase, then I certainly won't have any objection if someone wants to take on the project of changing all the gun patterns over to gun-collection canonical.

Conversely, if more people say they prefer the minpop phase, and not so many people prefer the gun-collection standard, then maybe we'll convert everything over to minpop phase instead

And (maybe the most likely outcome!) if people seem to be about evenly split between those two opinions, then probably we'll just keep muddling along doing the same thing we've been doing for the last several years. Each editor makes their own decision -- and we all just try to avoid getting into pointless edit wars about it.

Posted: **April 4th, 2025, 3:06 pm**

dvgrn wrote: April 4th, 2025, 2:29 pm [...] Editors' previous collective decisions are a valid type of de facto documentation. [...]

Then it's unclear why there are undiscussed changes away from the traditional canonical form (as in the case mentioned before):
https://conwaylife.com/w/index.php?titl ... did=159486
The revision before the linked edit contained the canonical form of the gun, added by a wiki editor.
The linked edit arbitrarily replaced that (claiming existence of "standard" where no such standard currently exists, and mentioning a few pages instead of linking to any preceding relevant discussion).

Out of the few pages mentioned in the edit summary, period-60 glider gun currently does in fact show the canonical form of a glider gun:

Code: Select all

x = 39, y = 27, rule = B3/S23
27bo$27b4o$11bo16b4o$10bobo5b2o8bo2bo5b2o$3b2o3b2o3bo14b4o5b2o$3b2o3b
2o3bo4bobob2o3b4o$8b2o3bo5b2o3bo2bo$10bobo10bo$11bo8bo2bo2$26bobo$28b
o$24bo$26bo$25bo2$11b2o$11b2o4bo$2o6b2o6b5ob2o$2o5b3o5bo2b2o4bo$8b2o5b
2o8bo12bo$11b2o4bo7bo10bobo$11b2o12bo11b2o$24bo8b2o$22b2o9bobo$35bo$35b
2o!

dvgrn wrote: April 4th, 2025, 2:30 pm [...] It seems like the edit summary provided some perfectly valid information about what is currently more common in LifeWiki gun articles. It gave several specific examples. [...]

It is not "perfectly valid information", until and unless there is review of the current situation (giving more than several specific examples), and until and unless there is discussion with some observable agreement on

what the current situation is (some evidence of people understanding what is going on),
what can be done in this case,
what should be done in this case.

Before that happens, that edit summary is an unsubstantiated claim, which is at least partially incorrect. (There is currently no LifeWiki-specific standard for glider guns. As far as I know, the only documented convention for glider guns is to use the last generation before an output glider leaves the bounding box moving in the lower-right direction, with width >= height.)

dvgrn wrote: April 4th, 2025, 2:30 pm [...] What do you suggest doing, to check in with the rest of the community about whether gun-collection canonical phase should become the official LifeWiki infobox standard? [...]

I don't really object to "opening the discussion, and waiting to see what other people think".

However I definitely do object to making unsubstantiated claims that can mislead people about what is agreed, what is a point of contention, what is the case and what isn't the case, what is an actual long-standing convention and what was never quite well documented or understood, what is borrowed from reliable sources and what amounts to mostly-arbitrary local choices that weren't understood well enough.

I don't object to discussion. I object to the attitude of avoiding any kind of interesting ontopic discussion that could lead to better understanding of a topic or an issue.

Posted: **April 4th, 2025, 3:38 pm**

Quoting to "ping" some of people who (as far as I know) are likely to be interested in this discussion (of course if anyone else is interested, please feel free to participate)

ElijahKen wrote: November 30th, 2024, 8:46 pm[...]

johamwit wrote: March 27th, 2025, 4:57 pm[...] BTW, It's nice to have agreed on the standard output of the guns being SE. It would be useful too to agree in the preferred display of the pattern. In rectangular patterns, we would choose between a "standard monitor" aspect ratio, being wider than tall, or a - ehem - modern, cellular display, taller than wide.

Scorbie wrote: March 28th, 2025, 5:28 pm[...] This topic could serve as the possible options to explore.

Sokwe wrote: December 4th, 2024, 7:22 am[...]

Vort wrote: December 4th, 2024, 4:01 pm[...]

Posted: **April 5th, 2025, 3:01 am**

dvgrn wrote: April 4th, 2025, 2:30 pm And (maybe the most likely outcome!) if people seem to be about evenly split between those two opinions, then probably we'll just keep muddling along doing the same thing we've been doing for the last several years. Each editor makes their own decision -- and we all just try to avoid getting into pointless edit wars about it.

This is what I do when I make contribution to open source software:
First, I try to match style of my edits with the style which is already present in surrounding code.
Second, if I think that present style must be changed, I make such change, but only around the code I touched.

With first approach, I always have justification about why I chose specific style: because this is how it was done before.
With second approach, it is possible to make style better, but rate of change is limited by amount of clearly useful (not style-related) modifications I can make. If all contributors follow this method, it also prevents edit wars - while style can be changed back and forth unlimited number of times, useful changes are usually limited.

Third method, with discussions and mass edits is possible as well, but in my experience, it usually wastes too much time and leaves lots of people dissatisfied with the outcome.

Posted: **April 7th, 2025, 7:37 am**

confocaloid wrote: April 4th, 2025, 3:06 pm
dvgrn wrote: April 4th, 2025, 2:30 pm [...] It seems like the edit summary provided some perfectly valid information about what is currently more common in LifeWiki gun articles. It gave several specific examples. [...]
It is not "perfectly valid information", until and unless there is review of the current situation (giving more than several specific examples), and until and unless there is discussion...

Right! This is the thread where that review and discussion is supposed to be happening. Let's discuss.

I've gone through the full list of 83 links in Category:Guns again, and have compiled statistics on how often the infobox gun patterns are presented in various formats. Here's the breakdown:

NOTE: see confocaloid's corrections to these categorizations, below.

Gun-collection canonical form
These patterns show the last generation before an output glider leaves the bounding box moving in the lower-right direction, with width >= height.

Articles with subject-of-article gun patterns in infobox in gun-collection canonical form: 20
Articles with subject-of-article gun patterns in gun-collection canonical form, but no infobox: 4

AK-94, medusa, p24 gliderless LWSS gun, period-143 glider gun (but no infobox, and displayed with STARTFROM 143), period-144 glider gun, period-18 glider gun, period-20 glider gun, period-24 glider gun, period-33 glider gun, period-34 glider gun, period-35 glider gun, period-435 glider gun (no pattern in infobox, moved out in June 2023), period-45 glider gun, period-47 glider gun, period-48 glider gun, period-49 glider gun (April 2023), period-52 glider gun, period-54 glider gun, period-55 glider gun, period-57 glider gun, period-60 glider gun, period-61 glider gun (2023), period-74 glider gun (no infobox, has Book theme, STARTFROM 74), Quetzal (no infobox)

Output shown beyond edge of bounding box
Gun patterns that have been run some number of ticks beyond canonical form: 4
gunstar, gunstar 2, period-41 glider gun, period-68 glider gun (no infobox, never had one)

Contains gliders/output, but not run out to edge of bounding box
Gun patterns containing some additional gliders, but displayed at some number of ticks before canonical form: 12
Same as above but with no infobox: 2

maximum volatility gun (no infobox), New gun 1, p13-assisted period-52 glider gun, p46 gliderless LWSS gun, P94S, Big_A#Guns (p690 gun not in infobox, contains a few gliders, not minimal), period-156 glider gun, period-22 glider gun, period-256 glider gun, period-51 glider gun, period-53 glider gun, period-59 glider gun (static image, 160697 cells in RLE, 167863 cells reported so that's wrong), period-61 glider gun (2016) (5403 cells reported, matches pattern once 2 gliders are removed, may or may not be minimal), period-88 glider gun

No gliders/output, not in minimum phase
Articles with subject-of-article guns in infoboxes containing no additional gliders/output, gun not in minimum phase: 11
Articles with subject-of-article guns but with no infoboxes, containing no additional gliders/output, gun not in minimum phase: 4

edge shooter (no infobox, KILLGLIDERS), p16-assisted period-32 glider gun, p16-assisted period-32 glider gun (pop count wrong, 143 should be 134), period-113 glider gun (there is a p62 phase, but a nearby p66 phase is given), period-14 glider gun (change to this article started this discussion, c=706 is given, but a 676-cell phase is shown, may not be minimal), period-15 glider gun, period-25 glider gun, period-30 glider gun (no infobox), period-40 glider gun (no LifeViewer in infobox, removed in August 2023), period-43 glider gun (86-cell phase shown, c=66), period-563 glider gun (in "Bronco", no infobox for gun), period-808 glider gun (c=548 is correct for phase shown, which is close to minimal but not quite, no more than 542), period-90 glider gun, period-92 glider gun (c=97 not quite the same as phase shown which is pop 100), slide gun (no infobox).

No gliders/output, in minimum phase
Articles with subject-of-article guns in infobox, no gliders/output shown, in minimum phase: 20
Articles with subject-of-article guns with no infobox, no gliders/output shown, in minimum phase: 2

B-52 bomber, bi-gun, Gosper glider gun, hotcrystal0 reaction (has Book theme, KILLGIDERS), new gun 2, p4-assisted period-28 glider gun (number of cells listed as 157, should be 159 I think), p80 gliderless HWSS gun, period-117 glider gun, period-16 glider gun, period-184 glider gun, period-201 glider gun, period-246 glider gun, period-27 glider gun (c=323, but 316-cell phase shown, may not be minimal), period-36 glider gun, period-44 glider gun (c=140, but 115-cell phase is shown, may not be minimal), period-44 MWSS gun, period-50 glider gun, p67 Herschel loop (c=294, correct for the no-output loop shown)[/wiki], period-690 glider gun, period-856 glider gun (c=275 matches phase shown), Simkin glider gun, vacuum (no infobox)

I left off a few articles that didn't seem to fit anywhere -- 6-engine Cordership gun, Dieter and Peter's glider gun collection, HBK gun. That still seems to add up to 82 not 83, so there must be one more odd case -- maybe something like not counting one of the transclusions. But the conclusions don't change much if one article is added or subtracted here and there, so the statistics seem like they're pretty good.

Boiling it all down a bit
There are 20+4 = 24 articles with subject-of-article guns in traditional gun-collection canonical form, either in the infobox or in the article text.

There are 22+11+4 = 37 articles with subject-of-article guns shown with no additional gliders/outputs.
Of these, 22 articles show the minpop form.

There are 4+12+2 = 18 articles where guns currently don't follow either the gun-collection canonical form or the no-additional-outputs convention.

Standards / conventions / competing options
It seems to me that 37, or even 22, is a big enough number that it's worth having a careful look at why all those minpop guns were deliberately uploaded that way. The odds of hitting a minpop phase are low enough that that wouldn't have happened accidentally, at least not often enough to substantially change the statistics.

I haven't gone through yet again and looked at upload dates for these guns, but just going off my vague knowledge of when various guns were discovered, it seems like people might have been following the minpop standard for guns on the LifeWiki fairly consistently at one point, and then that convention was later mostly forgotten.

So we've got two perfectly good valid competing options for standardizing gun phases that I can see -- gun-collection canonical and minpop. Or we might not want to decide to standardize gun phases at all, and leave it up to individual authors.

Any new opinions from anyone, in light of these new statistics?

Posted: **April 7th, 2025, 8:05 am**

Looks like at least some of these links are listed in the wrong category, so the counts are incorrect.

I didn't (yet) double-check all the links. But to give one example, hotcrystal0 reaction currently contains
(permanent link to the current revision: https://conwaylife.com/w/index.php?titl ... did=159743 )

a p174 glider gun in the canonical form (there is animation and KILLGLIDERS, but the stored pattern is the canonical form)
a p49 glider gun which is neither in the canonical form, nor in the minimum-population form (you can delete several output gliders to reduce population)
a p51 glider gun which is neither in the canonical form, nor in the minimum-population form (compare the population in generation 0 vs. the population in generation 4; gen 4 has lower population)
and several other patterns that aren't glider guns

- however, you're putting that page into the category "No gliders/output, in minimum phase", which is incorrect both in that there are gliders/output, and in that the guns aren't in the minimum phase.

dvgrn wrote: April 7th, 2025, 7:37 am [...]
Articles with subject-of-article guns in infobox, no gliders/output shown, in minimum phase: 20
Articles with subject-of-article guns with no infobox, no gliders/output shown, in minimum phase: 2
[...]
[...] hotcrystal0 reaction (has Book theme, KILLGIDERS), [...]

In this particular case (hotcrystal0 reaction), I would certainly prefer putting all of those guns in the canonical form. (Width >= height, gliders go to lower-right, an output glider will leave the bounding box of the gun in the next generation)

Posted: **April 7th, 2025, 9:19 am**

In at least one section, the quoted counts are incorrect: several listed articles fail to show any minimum-population form at all. (Out of 22 listed wiki pages, at least 7 don't belong in that section, counting hotcrystal0 reaction already mentioned above.) See the list below for details.

dvgrn wrote: April 7th, 2025, 7:37 am [...] I haven't gone through yet again and looked at upload dates for these guns, but just going off my vague knowledge of when various guns were discovered, it seems like people might have been following the minpop standard for guns on the LifeWiki fairly consistently at one point, and then that convention was later mostly forgotten. [...]

To me, it still looks like there is no such "minpop standard for guns". At least judging from the current state of those pages I checked so far.
Some guns happened to be in the minimum population, presumably because an editor decided to add them in the minimum population, but it doesn't look like that was ever followed consistently enough to deserve to be called something as grand as "standard", and I'm not aware of any source where that was explicitly written down as an actual convention to be followed.
It is also possible that an editor simply copied the pattern from a forum post or another source, without double-checking which generation and orientation should be shown.

The advantages of https://entropymine.com/jason/life/dpguns/readme.txt are that it is an old written convention, and it was followed by people in different contexts.

bi-gun (permanent link) fails to show the minimum-population form, which would have at most 40 alive cells instead.
In this case, my preference is to show the following form instead (the gliders are about to leave the bounding box, one of them goes lower-right, width >= height):
Code: Select all
```
x = 50, y = 17, rule = B3/S23
22b3o$22bo$23bo13b2o$38b2o$34b5o$34b4o$2o9b2o$2o8b2o22b4o$11b5o18b5o$
12b4o22b2o8b2o$37b2o9b2o$12b4o$11b5o$10b2o$11b2o13bo$27bo$25b3o!
```

new gun 2 (permanent link). Compare generation 1, which has fewer alive cells than generation 0.
In this case, my preference is to show the canonical form:

Code: Select all

x = 51, y = 25, rule = B3/S23
23b2o24b2o$23b2o24b2o2$29b2o$28bob2o$28bo3bo7b3o$29b5o6b2ob2o$19bo10bo
b3o8b2o$19bo13b2o8b2o$18bobo$15bo13b2o8b2o$14b3o2b2o8b2o8b3obo$14bo4b
3o7b2ob2o6b5o$14bob2obo11b3o7bo3bo$15b3o24b2obo$7b2o34b2o$7b2o$49b2o$
49b2o2$4b2o18bo$2o4b4o10b2o2b2ob3o$2o2b2ob3o10b2o4b4o7bo$4bo19b2o12bo$
36b3o!

Period-117 glider gun (permanent link) fails to show a minimum-population form. (Compare generation 32)
In this case, my preference is to show the following form instead (same reasons as previously):

Code: Select all

x = 42, y = 29, rule = B3/S23
7bo7b2o10bob2o4b2obo$2o5b3o5b2o10b2o2bo2bo2b2o$obo7bo14b2o3bo4bo3b2o$o
8b2o13bo3b2ob4ob2o3bo$25b3obobo2bobob3o$27bobobo2bobobo$18b3o9bo4bo$
31bo2bo$16bo5bo8bo2bo$13bo2bo5bo$13bo2bo5bo$13bo$18b3o$5b2o$5b2o28b2o$
35b2o$21b3o$28bo$19bo5bo2bo$19bo5bo2bo$7bo2bo8bo5bo$7bo2bo$6bo4bo9b3o$
3bobobo2bobobo$b3obobo2bobob3o$o3b2ob4ob2o3bo13b2o8bo$b2o3bo4bo3b2o14b
o7bobo$3b2o2bo2bo2b2o10b2o5b3o5b2o$3bob2o4b2obo10b2o7bo!

Period-856 glider gun (permanent link) fails to show a minimum-population form. (Compare generation 6)
In this case, my preference is to show the following form instead (same reasons as previously):

Code: Select all

x = 106, y = 105, rule = B3/S23
45b2o$44b4o$40bob2ob2obo$39bo2b2o$38b3o2b2o2b2o$38b3o2bob4o$39bo2bo3bo
2bo$40b4obo2b3o$40b2o2b2o2b3o$45b2o2bo$40bob2ob2obo$41b4o$42b2o7$83bo$
81b3o$80bo$18b2o60b2o$19bo$19bobo$20b2o3$72bo$54b2o15bo$54b2o14b2o$71b
2o3b2o$72b3ob2o$74bobo$75b2o7$95b2ob6o$95b2ob6o$95b2o$58b2o35b2o5b2o$
47b2o2bo6bo2bo33b2o5b2o$43b2o2b2o3bo42b2o5b2o$43b2o7bo5bo2bo40b2o$48b
4o5bo2bo34b6ob2o$57bobo35b6ob2o$28b2o28bo$28b2o$58b2o$45b2o11b2o$45bo
2bo$75b2o$3b2o40bo2bo26b2o$3bo40bo2bo$2b2o2b3o35bobo9b2o2bo$2bo3bob3o
34bo6b2o2b2o3bo$2b2obo4bo41b2o7bo$4bobo38b2o10b4o$o4bob2o36b2o$3obo3bo
$2b3o2b2o$7bo$6b2o10$49b2o$49b2o54bo$103bobo$104b2o2$83b2o$83bobo$85bo
$23b2o60b2o$24bo$21b3o$21bo9$57bobo2bo$57b3obob2o$56bo6bo$57bo5b2o2$
56b2o5bo$57bo6bo$56b2obob3o$58bo2bobo!

p67 Herschel loop (permanent link) doesn't show any gun in the infobox.

vacuum (permanent link) doesn't have any patterns shown (!).
The page Vacuum (gun) (permanent link) has errors in the infobox, it incorrectly claims population 36. Not sure what was intended here. My preference would be to show the following form instead (same reasons as previously):

Code: Select all

x = 49, y = 43, rule = B3/S23
b2o23b2o$b2o23bo$24bobo$12b3o9b2o$2o9bo4bo$2o8bo5bo4b2o$11bo8bo2bo$12b
2o7bo$22b2o20b2o$12b2o8b2o20b2o$11bo8bo2bo4bo$2o8bo5bo4bobo2bo2bo3bo$
2o9bo4bo4b2o3bo7bo10b2o$12b3o10bo4b2o2bo10b2o$25b2o5b2o$b2o22b2o3b3o$b
2o$25b2o3b3o$24bobo5b2o$24bo5b2o2bo10b2o$24b2o8bo10b2o$29bo3bo$28bo$
44b2o$44b2o4$45b2o$35b2o7b4o$33b2o2bo6bobobo$33b6o4b3ob2o$33b4o5bob2o$
42bob2o2$42bob2o$33b4o5bob2o$33b6o4bo2bo$33b2o2bo6b2o$16b2o17b2o7b2o$
15bobo6bo5bo$15bo9bo5bo$14b2o7b3o3b3o!

dvgrn wrote: April 7th, 2025, 7:37 am [...]
Articles with subject-of-article guns in infobox, no gliders/output shown, in minimum phase: 20
Articles with subject-of-article guns with no infobox, no gliders/output shown, in minimum phase: 2

B-52 bomber, bi-gun, Gosper glider gun, hotcrystal0 reaction (has Book theme, KILLGIDERS), new gun 2, p4-assisted period-28 glider gun (number of cells listed as 157, should be 159 I think), p80 gliderless HWSS gun, period-117 glider gun, period-16 glider gun, period-184 glider gun, period-201 glider gun, period-246 glider gun, period-27 glider gun (c=323, but 316-cell phase shown, may not be minimal), period-36 glider gun, period-44 glider gun (c=140, but 115-cell phase is shown, may not be minimal), period-44 MWSS gun, period-50 glider gun, p67 Herschel loop (c=294, correct for the no-output loop shown)[/wiki], period-690 glider gun, period-856 glider gun (c=275 matches phase shown), Simkin glider gun, vacuum (no infobox)
[...]

In general, my preference is to show the canonical form (per readme.txt), unless there is an obvious reason why that would be impossible.

Even then, whenever there is an obvious way to extend the notion of canonical form of a gun, I would prefer that. For example, I would use the following form of period-44 MWSS gun because it follows as many of the same guidelines as possible (width >= height, the last generation before an output spaceship leaves the bounding box):

Code: Select all

x = 50, y = 40, rule = B3/S23
7bo$2b2o3bobo4bo$2bo2b2o2bobob2obo2b3o2b2o$4b2o7b2obobob2o2b2o$3b3o7bo
7b3o$23bo$23bo2$9b2o$9b2o4bo$15bo$14b3o12b2o$29bo$12bo5bo8bobo$10bo9bo
6b2o$2o7b2o4bo4b2o$2o6b2o4b3o4b2o8b2o$9bo2b3ob3o2bo9b2o$10b2obo3bob2o
25bo$12b2o3b2o26bobo$12b2o3b2o13bo12bobo$10b2obo3bob2o11b2o10b2ob2o$9b
o2b3ob3o2bo9bo2bo13bo$2o6b2o4b3o4b2o11bo9b2obo$2o7b2o4bo4b2o9b2o2b2o7b
2obobo$10bo9bo10b2o4bo10b2o$12bo5bo17b3o$37bo$14b3o18b2o$15bo18bo$9b2o
4bo$9b2o18bobo$30b2o$23bo6bo$23bo7b3o6bobo$3b3o7bo7b3o9bo5bo$4b2o7b2ob
obob2o2b2o13bo3bo$2bo2b2o2bobob2obo2b3o2b2o13bo$2b2o3bobo4bo24bo2bo$7b
o31b3o!

Posted: **April 7th, 2025, 9:55 am**

confocaloid wrote: April 7th, 2025, 8:05 am Looks like at least some of these links are listed in the wrong category, so the counts are incorrect.

Thanks for reviewing the minpops -- we can maybe use those corrections to adjust the "c" parameter, if and when someone goes through and cleans up all of the inconsistencies.

My previous post was getting long already, so there were some subtleties I didn't get into. When the displayed version uses STARTFROM, it's not entirely clear which category it should go into. I noted cases like that when I saw them, but didn't notice for the hotcrystal0 reaction script commands.

I'll just leave my imperfect counts as they are, rather than trying to go back and patch them up. I might make more mistakes in attempting any corrections -- which would be an unnecessary distraction, since the corrections don't change the overall picture.

Please don't worry about completing a similar catalogue for all of the articles that have multiple versions of guns shown. There are a lot of those, and I'm pretty sure that gun-collection canonical form shows up a lot in the galleries of increasingly smaller forms. If you count up all of those, I think that gun-collection canonical form will easily "win" on the count metric.

Conversely, "no output shown at T=0" has a clear plurality among patterns displayed in infoboxes, whether or not we include the first patterns shown in articles with no infobox. The 37-or-so "no output shown at T=0" guns either show the true minpop or else, usually, some relatively low-population phase. Maybe people were sometimes choosing some particularly recognizable phase of the gun's engine?

Anyway, it was the "no output shown at T=0" infobox convention that WhiteHawk was following in making the edit to the period-14 glider gun article.

Next steps
Again, the point here is not whether "37 > 24" or "24 > 22" or whatever is the right equation. There isn't a "right equation" to decide this question.

The point is just that both the "gun-collection canonical" convention and the "no output at T=0" convention are clearly well-established in LifeWiki articles. So in this discussion we're checking in with the community to see how many people strongly prefer one convention or the other. If nobody cares about keeping the displayed minpop phases, that's easy -- we can switch everything over, either gradually or in a deliberate sweep.

However, WhiteHawk made a change recently in the direction of the "no output at T=0" convention. So it seems like a good idea to spend some time asking for comments from more people, before making any more changes in either direction.

Posted: **April 7th, 2025, 10:03 am**

dvgrn wrote: April 7th, 2025, 9:55 am [...] both the "gun-collection canonical" convention and the "no output at T=0" convention [...]

I don't see this as a meaningful comparison.

In the former case, the convention was actually written down (readme.txt), it was followed, it is reasonably well-known, and it can in fact be followed to determine a specific generation and orientation of the gun.

In the latter case, there is simply not enough information provided. The phrase "No output at T=0" doesn't tell what exactly should be the initial generation and orientation. (And I don't remember any place where that was previously written down, either.)

Posted: **April 7th, 2025, 10:10 am**

confocaloid wrote: April 7th, 2025, 10:03 am
dvgrn wrote: April 7th, 2025, 9:55 am [...] both the "gun-collection canonical" convention and the "no output at T=0" convention [...]
I don't see this as a meaningful comparison.

That's okay. Other people will understand the comparison just fine. It's simply the two obvious options for a LifeWiki infobox standard, based on the choices that multiple editors have made in setting up gun patterns.

confocaloid wrote: April 7th, 2025, 10:03 am In the former case, the convention was actually written down (readme.txt), it was followed, it is reasonably well-known, and it can in fact be followed to determine a specific generation and orientation of the gun.

In the latter case, there is simply not enough information provided. The phrase "No output at T=0" doesn't tell what exactly should be the initial generation and orientation.

Neither of these conventions were ever written down on the LifeWiki. We're aiming to change that, if possible -- but, ideally, not until we hear from more people than just you and me.

There are plenty of cases left where a minpop phase was shown -- that's a fine convention, since we have to find that phase anyway to document the "c" parameter correctly. Length >= height is pretty universally followed already. Never fear, we can easily write up all the details for whichever convention we decide on -- and then there won't be any argument about the reason for corrective edits, when a new gun page is created that doesn't quite meet the official standard. (That's if we can actually decide on a standard, of course.)

Posted: **April 7th, 2025, 10:18 am**

My point is that the phrase "no output at T=0" fails to make a convention (actual or potential). It doesn't tell enough to be a convention.

dvgrn wrote: April 7th, 2025, 10:10 am [...] Never fear, we can easily write up all the details for whichever convention we decide on -- [...]

I think that misses the problem.
I would expect some level of understanding and agreement on just what the possibilities are (along with all the details), some understanding of what the situation is and what exactly can be done, before people can decide what they think should be done.

This is essentially the same situation as in an earlier discussion quoted below.

dvgrn wrote: February 25th, 2024, 10:48 pm
Haycat2009 wrote: February 25th, 2024, 9:30 pm Obviously I will put in relevant links and correct it. Do not forget that this is not as-is, but a rough guide. I will refine it if the answer is yes.
Heh, as you've discovered, the answer might not be "yes" until there's a final form to look at and say "yes" to.

Posted: **April 7th, 2025, 11:20 am**

confocaloid wrote: April 7th, 2025, 10:18 am My point is that the phrase "no output at T=0" fails to make a convention (actual or potential). It doesn't tell enough to be a convention.

Of course not! But I've said lots of other words recently on this subject besides just "no output at T=0".

To spell it out more formally, the "no output at T=0" convention can easily be expanded to

"no output at T=0, minimum-population phase, width >= height, output leaves the bounding box moving to the right (first choice) or down (second choice)".

That works for non-glider outputs and is a little more specific about square bounding boxes. Could also add something a little more subjective for cases where there isn't a unique choice, like "if multiple phases share the same minpop, choose the phase with the most recognizable subpatterns if applicable."

Or to make a really definite "canonical minpop" convention with no space for judgment calls, we could say "choose the minpop phase that is the smallest number of generations before the output is generated".

Posted: **April 7th, 2025, 11:38 am**

But in that case your earlier attempts to find some justification (by collecting links to wiki pages where people presumably followed such a "convention") fail to be relevant.

More explicitly:

It might be that there are many pages where the gun in the infobox doesn't show an output glider at T = 0.
But that doesn't make a coherent useful convention. It is just an arbitrary observation, it doesn't tell enough to count as an option when discussing what could be a consistent convention for LifeWiki.
(There are also many pages where the gun in the infobox doesn't contain any scorpions or paperclips at T = 0. Neither of that would make sense as a convention, either.)
It might be that one potentially applicable convention is "no output gliders at T = 0, a minimum-population phase, width >= height, output glider leaves the bounding box moving to the right (first choice) or down (second choice)" (or some variant/clarification of that).
But that is no longer supported by your links to wiki pages. Instead, you would need to link to pages where people followed that particular version of whatever "potential convention" is considered. Just because some gun happened to lack an output glider at T = 0, doesn't mean someone had in mind the exact version of a "potential convention" you stated.

I think that means the link-collecting activity is likely to miss the main point (if there was any) of this discussion. It misleads more than it helps.

I would definitely prefer some more brainstorming/actual discussion involving other people, regarding

what the current situation regarding glider guns is,
what can be done? What are the plausible interesting possibilities?
and only when the above is understood, "what should be done".

dvgrn wrote: April 7th, 2025, 11:20 am [...] To spell it out more formally, the "no output at T=0" convention can easily be expanded to

"no output at T=0, minimum-population phase, width >= height, output leaves the bounding box moving to the right (first choice) or down (second choice)".

That works for non-glider outputs and is a little more specific about square bounding boxes. Could also add something a little more subjective for cases where there isn't a unique choice, like "if multiple phases share the same minpop, choose the phase with the most recognizable subpatterns if applicable."

Or to make a really definite "canonical minpop" convention with no space for judgment calls, we could say "choose the minpop phase that is the smallest number of generations before the output is generated".

Posted: **April 7th, 2025, 11:47 am**

confocaloid wrote: April 7th, 2025, 11:38 am I would definitely prefer some more brainstorming/actual discussion involving other people...

That's a wonderful idea. @Other people, please discuss.

Posted: **April 7th, 2025, 12:50 pm**

dvgrn wrote: April 7th, 2025, 11:47 am @Other people, please discuss.

I prefer solution of enriching gun patterns with more metadata.
In the simplest case, with LifeHistory.
Even with gliders removed, their path/lane will be usually visible.
Besides LifeHistory, something like GliderHistory can be invented.
Another option is to add marking not related to CA - arrows for example.
However, I doubt my ideas will be popular.
Regarding two main solutions presented in this topic, I can't say which is better.

Posted: **April 7th, 2025, 1:10 pm**

Many people, especially people who read the page but aren't active in the community, and even some who are in the community, don't know how to convert LifeHistory to two-state without making the history cells turn on.

Catagolue puts its glider gun pages in LifeHistory and I hate it.

As for Catagolue-defined canonical phase, it's sometimes annoying when it's a high-period gun and it requires almost a full cycle to see the LCM interaction and get one glider out, since a glider was just created a few generations ago if there's a glider at the edge of the bounding box.

Posted: **April 7th, 2025, 1:22 pm**

hotdogPi wrote: April 7th, 2025, 1:10 pmMany people, especially people who read the page but aren't active in the community, and even some who are in the community, don't know how to convert LifeHistory to two-state without making the history cells turn on.

This is trivial programming task and not-so-hard UX task.
Making button for single click conversion and writing help page explaining how to press this button should not take much effort.

There is a risk of having too much overlaid data, which would be really hard to "parse", but that's a problem of far future.
Few arrows, few numbers, few extra states - it should be ok.

Posted: **April 7th, 2025, 1:36 pm**

I would certainly prefer to stick to two-state RLE by default, to keep things simple. If there is/can be some additional programming/automated processing, then it would be better to autoconvert from two-state RLE to "more rich" presentations if/when requested (rather than in the other direction), autogenerating any extra markings in some understood consistent way. In the context of LifeWiki, it may or may not be possible or reasonable to add automated processing, but two-state patterns work anyway.

Posted: **April 7th, 2025, 2:48 pm**

Vort wrote: April 7th, 2025, 12:50 pm I prefer solution of enriching gun patterns with more metadata.
In the simplest case, with LifeHistory.
Even with gliders removed, their path/lane will be usually visible.

Interesting point.

It came up in my survey that an occasional gun had been set up with the LifeViewer Book theme, and usually also a STARTFROM value (or just AUTOSTART). This creates a somewhat "LifeHistory-like" multicolor view, showing the shape of the reaction envelope and the trail left by the output -- but RLE copied out of those views remains plain simple two-state.

Just as an idea, it would theoretically be possible to make that the standard across the board for guns -- e.g., we could start with minpop RLE, and run the pattern to the point where a glider is present at the edge of the gun's bounding box... and then do one more full cycle to show where the gliders are going.

However, it's a separate open question whether we want to move away from the current "LifeWiki theme" of simple black-and-white pattern displays in LifeWiki articles. I think this is the most recent thing that anybody has said about Book theme:

confocaloid wrote: March 29th, 2025, 5:12 am It should be an improvement to get rid of "THEME Book" which (in LifeWiki context) was mostly a failed experiment. Patterns usually look better when they are plain two-state black/white.

I tend to agree with confocaloid on that point. Two-state is what CGOL actually is, and so it makes some sense to follow the "keep it simple" rule and display in two states. Otherwise, with LifeHistory or "GliderHistory" or Book theme or whatever, the colors aren't really part of the pattern. Per Vort's suggestion, maybe we could write up a standard explanation of the colors of Book theme (for example) and figure out how to set up a template so that an obvious link to the explanation is always available.

Still, LifeViewer makes it pretty darn simple to run two-state gun patterns, to easily find out where the gliders are going.

confocaloid wrote: April 7th, 2025, 1:36 pm... it would be better to autoconvert from two-state RLE to "more rich" presentations if/when requested (rather than in the other direction).

I was thinking somewhat along the same lines -- e.g., at a minimum, a user-configurable default theme for LifeViewer, with plain black-and-white being the "default default".

That might work best in conjunction with a consistent STARTFROM setting for glider guns, so that if you switch to Book theme you can see the colorization that you want to see. That could also make it possible to retrieve either the minpop gun (at T=0, by hitting Reset) or the gun-collection canonical version (at T=whatever, by just copying the pattern with Ctrl+C).

Posted: **April 7th, 2025, 4:27 pm**

Comparing between (simplified, missing details) alternatives
(A) "last phase before an output glider leaves the bounding box"
(B) "minimum-population phase"

- the former (A) is not hard to find "semi-manually" in practice (using shortcuts Alt+H and Alt+J in Golly for conversions), and also easier to understand in the sense that the potential ambiguity is predictable. (In the worst case, you would have a glider gun with square-shaped bounding box, so that width=height and there is ambiguity in whether to flip the gun across the diagonal; and at the same time the glider gun would be pseudo period, so that the output stream period is n but the mechanism works at period (kn) for some integer k > 1. Even then, it is easy to find all possible choices given any one of them; evolve n ticks at a time, flip across the diagonal, repeat.)

- the latter (B) is significantly harder to find "semi-manually" (with a comparable approach, one would have to actually check every phase, carefully detecting and deleting output gliders as soon as they can be proven to escape without any future interaction, and checking the populations), and also harder to understand, in the sense that knowing just one of minimum-population phases does not automatically tell you what are other minimum-population phases, nor how many are there minimum-population phases.

Even though neither of two choices is necessarily unique without further clarifications, the old traditional choice of "last phase before an escaped glider leaves the bounding box" makes it easy to predict any possible non-uniqueness. (Say, if a gun is pseudo-period merely due to the period doubled by a boat-bit, then it would be natural to choose the canonical form where the boat is absent, as it has lower population.)

Verification is also easier for (A): when a glider gun is already given in the canonical form, it is easy to verify that.
For (B), you would basically need to redo most of the work to verify correctness.

Posted: **April 7th, 2025, 11:13 pm**

confocaloid wrote: April 7th, 2025, 1:36 pmI would certainly prefer to stick to two-state RLE by default, to keep things simple. ...

In my opinion, this discussion shows that such simplicity is not enough in this case.

dvgrn wrote: April 7th, 2025, 2:48 pm ... Just as an idea, it would theoretically be possible to make that the standard across the board for guns -- e.g., we could start with minpop RLE, and run the pattern to the point where a glider is present at the edge of the gun's bounding box... and then do one more full cycle to show where the gliders are going. ...

There are two classes of "enriching" solutions:
1. Add more data manually, then strip extra data when needed;
2. Start with minimum of data and then add extra data automatically.

Option 2 is better (less chances to make mistakes), but requires more work.

dvgrn wrote: April 7th, 2025, 2:48 pm ... I think this is the most recent thing that anybody has said about Book theme:

confocaloid wrote: March 29th, 2025, 5:12 am It should be an improvement to get rid of "THEME Book" which (in LifeWiki context) was mostly a failed experiment. Patterns usually look better when they are plain two-state black/white.
I tend to agree with confocaloid on that point. Two-state is what CGOL actually is, and so it makes some sense to follow the "keep it simple" rule and display in two states. Otherwise, with LifeHistory or "GliderHistory" or Book theme or whatever, the colors aren't really part of the pattern. Per Vort's suggestion, maybe we could write up a standard explanation of the colors of Book theme (for example) and figure out how to set up a template so that an obvious link to the explanation is always available.

I think THEME Book just has bad choice of colors.
At least, they distract viewer from seing actual pattern.
Switching from rainbowy style to shades of the same color and making them less protruding may fix problems of this style.

---

One more solution to the problem of this topic:
It should be possible to add tabs like "Phase 0" / "Phase 1" / "Phase 2".
There will still be a need to decide which phase should be shown by default, but for people who want to see different phase it will be 1 click away.

Posted: **April 8th, 2025, 4:12 am**

Vort wrote: April 7th, 2025, 11:13 pm [...] There are two classes of "enriching" solutions:
1. Add more data manually, then strip extra data when needed;
2. Start with minimum of data and then add extra data automatically.

Option 2 is better (less chances to make mistakes), but requires more work. [...]

Option 2 would require more design/programming/testing work from those people who would implement it. It would be "more work" in the sense convincingly explained years ago in a discussion of another topic: (i.e., "Easier for whom?")
viewtopic.php?p=3195#p3195
viewtopic.php?p=3200#p3200
However, Option 2 would require that work only once (as opposed to every interested reader having to repeat the same conversions or calculations, or every related editing task requiring repeated conversions or calculations), and it would guarantee consistency.

In practice, Option 1 (with editors manually adding extra data) is guaranteed to lead to duplicated efforts and visible failures of consistency.

Hence I think Option 2 would be a better choice ("Store less; compute more").
Assuming that people who are supposed to design and implement it can agree on some reasonably good working approach.
For comparison: the discussions I linked to happened in 2010. Today in 2025, there *is* reasonably good support for CA defined using weighted neighbourhoods, both in Golly and in LifeViewer. As an example, here is a Bosco-based p498 glider gun (shown in canonical form):

Code: Select all

#C A period-498 glider gun made using several interacting Boscos.
#C Source: https://conwaylife.com/forums/viewtopic.php?p=197882#p197882
#C [[ MAXGRIDSIZE 9 GRID THEME Catagolue ]]
x = 193, y = 193, rule = R5,C0,S226,228-229,257-260,265,273,339-341,344,346-350,352-354,369-393,481-505,593-617,705-729,817-841,931-953,B34-45,146-157,232-236,241,258-269,339-350,352-354,370-381,482-493,594-595,597-605,712-717,820,823,936,NW01010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010171717101010101010101017100710101010101010101717171010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101
191b2o$191b2o44$143b3o$140b7o$139b5o2b2o$139b4o4b2o$138b5o5bo$138b4o4b
4o$139b4o3b3o$140b9o$141b7o$141b7o$142b5o23$86b2o$86b2o17$153b2o$152b
5o$151b6o$150b8o$148b10o$148b4o3b4o$148b4o5b2o$148b4o5b2o$148b5o4bo$
149b6obo$151b5o33$47b5o$46b6o$46b7o$45b9o$45b3o3b4o$45b2o4b5o$45b2o4b
5o$45bo4b5o$46bo3b5o$47b6o$48b3o49b6o$100b6o$99b8o$98b9o$97b5o2b4o$97b
4o4b3o$97b5o4b2o$98b4o4bo$99b4o2bo$100b5o32$181bo$2o180bo$2o178b3o!

confocaloid wrote: November 13th, 2024, 3:28 pm There may be bugs in CGoL: [...]
This cellular automaton supports the p166 Bosco and bugs, while also supporting several small objects from Conway's Life. [...]

ConwayLife.com

Standardizing guns on LifeWiki

Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki

Re: Standardizing guns on LifeWiki