Catagolue Suggestions Thread

[GUYTU6J]

How about a downstream synthes[i/e]s report for each GoL object? For example, a button on the "glider synthesis" section of the object that returns a list of apgcodes, glider count increments in next step, and whether or not this step is in the optimal route, of immediate synthetic descendants. This will be useful to understand how many benefits a reduction have.

[confocaloid]

A suggestion to completely disable animated previews of objects in apgcode-prefix listings, so that only static images of phases are shown in lists (regardless of period/size of objects).

For example, these pages significantly lag my browser: b2n3s23-q/C1/xp30, b2n3s23-q/C1/xp6. Sometimes a long listing of animated previews causes the page to freeze, so that I have to close and reopen the browser tab.

It seems also that the logic behind animated previews is not always consistent. The following two pages behave differently with respect to animations: xq45 displays static image of a phase; xq57 displays animation.

[silversmith]

Would it be possible to implement INT generations support for the RLE identification box on Catagolue? I have several patterns which I would like to find some sample soups for, and RLE identification would be very useful.

[benetnasch85]

Is there a way to list all of the lifelike censuses with a particular symmetry, such as C1? If not, can one be provided?

I'm looking for a list of rules with the number of objects for each, which can be sorted by either rulestring or number of objects.

Alternative: is there a way to show an extended version of the "Most Popular" list on the Census page? If not, can that be provided?

[muzik]

For the syntheses page, would it be possible to italicise object groups whose contents consist solely of pseudo-objects, such as xp354?

[DroneBetter]

A suggestion to completely disable animated previews of periodic objects in apgcode-prefix listings, so that only static images of phases are shown in lists (regardless of period/size of objects).

I have encountered this problem myself, but I think a better solution would be to also be able to specify the number of patterns you would like to show, or (being that each cell is a html element) the number of cells to display, such that pages of the list can be of variable length, for people who would like only to review all objects in a list by seeing them in motion. Also, to allow sorting by different characteristics like population and volatility would be nice.
(also, add all-soups censuses to other rules and bring back the gap/skew symmetries into the official ones so that we can have the skewed pre-pulsar shuttle in all-soups :-)

[muzik]

Would it be at all possible to provide a copyable RLE for infinite growth patterns since some of them also have syntheses? These would obviously have to be provided manually.

[AlbertArmStain]

There NEEDS to be a feature for Catagolue where all C1 soups of the certain object are in a computer generated RLE. It’d be extremely useful for syntheses, mainly oscillator syntheses but it could be useful for still life syntheses as well. I’m practically losing my mind over this because I’m looking for optimal one glider seeds, and looking through more than 50 soups isn’t fun.

[dvgrn]

There NEEDS to be a feature for Catagolue where all C1 soups of the certain object are in a computer generated RLE. It’d be extremely useful for syntheses, mainly oscillator syntheses but it could be useful for still life syntheses as well. I’m practically losing my mind over this because I’m looking for optimal one glider seeds, and looking through more than 50 soups isn’t fun.

Just for the record, there is a lifelib script that does exactly this -- find-preds.py. The LifeWiki tutorial article links to this version of the script.

Fair warning: I have personally found that installing lifelib on Windows, though theoretically easy if nothing goes wrong, is not a task for the faint of heart if anything does go wrong. I still haven't had enough time to successfully untangle the mess I made the last time I tried this.

So the above unfortunately may not be very helpful for someone who "doesn't know how to use scripts". The first step would be to make a firm decision to learn how to use scripts.

[muzik]

Would it be possible to track glider syntheses in non-B3/S23 rules such as tlife and DryLife?

[confocaloid]

Would it be possible to track glider syntheses in non-B3/S23 rules such as tlife and DryLife?

If you allow rules where the most common spaceship (= probably the one to be used for synthesis) is not the glider (or at least there are other common spaceships usable for syntheses), it seems nontrivial to have an uniform way of recording syntheses that works for different rules. As soon as the glider used for syntheses can be something like (35,10)c/145, documenting 2G collisions becomes somewhat of a headache.

What could be a reasonable solution for a rule with two approximately equally common gliders? Should it be possible to record syntheses that use both, or instances of one glider would need to be constructed using the other one?

Code: Select all

#C 2G carrier
#C Rule and spaceship from here: https://conwaylife.com/forums/viewtopic.php?p=39403#p39403
x = 38, y = 14, rule = B3-q5nq/S2-ck3-n4ir5y
2o$b2o$o9$36bo$35b2o$35bobo!

Code: Select all

#C 2G loaf
#C Rule and spaceship from here: https://conwaylife.com/forums/viewtopic.php?p=39406#p39406
x = 47, y = 9, rule = B35ijn6i/S23-cnq4jy5n
44b3o$44bo$45bo4$3o$2bo$bo!

[muzik]

I was thinking that rules would just be granted synthesis-recording privileges on a case-by-case basis. Only rules that support the glider would be granted this, since I'm sure most of the technology used for recording syntheses is optimised for them.

[muzik]

Since conwaylife.com links to Catagolue in the top bar, would it be possible for Catagolue to also link to conwaylife.com to permit easier navigation between the two sites?

[muzik]

Could some way be added for soups on Catagolue object pages to be played from within the browser, rather than having to copy the RLE into another program (but of course leaving the RLE available so that can still be an option)?

[muzik]

A minor question: could the rules page be presented as a table instead of a list? The variable length of rulespace names makes the counts displayed next to them slightly irritating to read.

(There's also the comparatively minor, although differently, issue that "von Neumann" isn't capitalized.)

And a question: why are isotropic non-totalistic B0 rules given their own category, whereas outer-totalistic B0 rules aren't? I assume it's related to support for them being added later, perhaps via the help of CAViewer if I remember right, but I don't know why or if that should have any bearing on how the site categorises such rules.

[calcyman]

Congratulations on the quadruple-post.

(There's also the comparatively minor, although differently, issue that "von Neumann" isn't capitalized.)

Fixed: https://gitlab.com/hatsya/open-source/c ... dc60806719

And a question: why are isotropic non-totalistic B0 rules given their own category, whereas outer-totalistic B0 rules aren't? I assume it's related to support for them being added later, perhaps via the help of CAViewer if I remember right, but I don't know why or if that should have any bearing on how the site categorises such rules.

Yes, they're handled differently by lifelib, and Catagolue's rule families are copied from lifelib's rule families as part of update-repo.sh.

[confocaloid]

Currently the "Isotropic non-totalistic" family has over 14000 rules, and the /rules/isotropic page is very long. Would it be possible to split that page further into separate subpages that can be browsed one at a time, according to the number of toggled isotropic conditions away from Life? (In other words, a page for 1TFL rules, a page for 2TFL rules, a page for 3TFL rules, and so on. For the URL, one could probably use "/rules/isotropic/1tfl", "rules/isotropic/2tfl", etc.)
Yes, this subdivision is arbitrary and not perfect, but this would work for all isotropic rules, and it is the simplest subdivision that I can think of.

To give some data, here is the number of rulestrings with an existing C1 census (derived from abclist), for each number of toggled isotropic conditions away from Life:

Code: Select all

68 1TFL rulestrings
1862 2TFL rulestrings
509 3TFL rulestrings
495 4TFL rulestrings
428 5TFL rulestrings
351 6TFL rulestrings
316 7TFL rulestrings
346 8TFL rulestrings
338 9TFL rulestrings
341 10TFL rulestrings
303 11TFL rulestrings
308 12TFL rulestrings
283 13TFL rulestrings
284 14TFL rulestrings
275 15TFL rulestrings
269 16TFL rulestrings
290 17TFL rulestrings
254 18TFL rulestrings
291 19TFL rulestrings
289 20TFL rulestrings
350 21TFL rulestrings
338 22TFL rulestrings
306 23TFL rulestrings
321 24TFL rulestrings
341 25TFL rulestrings
372 26TFL rulestrings
403 27TFL rulestrings
425 28TFL rulestrings
387 29TFL rulestrings
414 30TFL rulestrings
389 31TFL rulestrings
398 32TFL rulestrings
354 33TFL rulestrings
365 34TFL rulestrings
363 35TFL rulestrings
361 36TFL rulestrings
325 37TFL rulestrings
308 38TFL rulestrings
307 39TFL rulestrings
297 40TFL rulestrings
301 41TFL rulestrings
261 42TFL rulestrings
249 43TFL rulestrings
254 44TFL rulestrings
277 45TFL rulestrings
222 46TFL rulestrings
215 47TFL rulestrings
217 48TFL rulestrings
180 49TFL rulestrings
181 50TFL rulestrings
177 51TFL rulestrings
156 52TFL rulestrings
151 53TFL rulestrings
147 54TFL rulestrings
138 55TFL rulestrings
115 56TFL rulestrings
113 57TFL rulestrings
89 58TFL rulestrings
80 59TFL rulestrings
49 60TFL rulestrings
35 61TFL rulestrings
27 62TFL rulestrings
27 63TFL rulestrings
21 64TFL rulestrings
21 65TFL rulestrings
12 66TFL rulestrings
22 67TFL rulestrings
12 68TFL rulestrings
17 69TFL rulestrings
15 70TFL rulestrings
19 71TFL rulestrings
19 72TFL rulestrings
16 73TFL rulestrings
12 74TFL rulestrings
14 75TFL rulestrings
11 76TFL rulestrings
14 77TFL rulestrings
6 78TFL rulestrings
5 79TFL rulestrings
8 80TFL rulestrings
4 81TFL rulestrings
1 82TFL rulestrings
1 83TFL rulestrings
1 84TFL rulestrings
2 86TFL rulestrings
3 87TFL rulestrings
1 88TFL rulestrings
1 89TFL rulestrings
2 90TFL rulestrings
1 91TFL rulestrings
2 94TFL rulestrings
2 95TFL rulestrings
1 96TFL rulestrings

I also love the "rule properties" page idea — maybe that should be where the links at the /rules subpages take you first, and then there's a link from there to the rule census, LifeWiki page (if it exists), and even maybe the forum thread, plus of course the list of spaceships. I notice there don't seem to be any pages in the /rules directory that start with lowercase letters, so doing something like https://catagolue.hatsya.com/rules/b3s23 might be perfectly fine... There might still be "rules" people have submitted to Catagolue as a joke that do start with capital letters, so I don't know.

I support the idea of having "rule properties" associated with rules.
E.g. definitely it would be very nice to have links to forum threads for all rules that have dedicated forum threads, maintained in some maintainable way. Probably also links to forum posts regardless of whether there is a thread. The forum search does not always find posts on a specific rule.

(There is an existing answer (which I did not test myself), but that seems to rely on a scraper script that takes some amount of server resources, and depends on someone regularly running the script and uploading the results somewhere, so that other people can access the cached search results instead of re-running the same search again.

I have an incomplete table of isotropic rules linked from my signature, but (apart from incompleteness) it suffers from the same disadvantage, i.e. depends on someone updating it in a semiautomated-but-not-completely-automated way.)

Here's one extremely straightforward solution to the problem:

• By definition, an interesting rule is any rule with a nonempty set of associated properties that someone would like to record (excluding any properties that are already automatically available for every rule on Catagolue).
• For every interesting rule, create a separate dedicated wiki page in the OCA namespace. By convention, the wiki pagename can be formatted like OCA:B35/S236. The rulestring-based pagename can be made into a redirect when there is already a commonly accepted name.
• Have Catagolue rule page link to the wiki page, in the same way as /census/b38s23 currently links to OCA:Pedestrian Life.
• Put all associated properties on the wiki page.

Consequences would include:

• For every interesting rule, exactly one needed change requires access to Catagolue (i.e. linking the wiki page). Everything else can be done on the wiki, in a community driven fashion.
• With a dedicated page for every interesting rule, there would be more space for editors. (For example, several editors can edit different wiki pages at the same time, without having technical edit conflicts due to interleaving edits).
• With a dedicated page for every interesting rule (with the above definition of "interesting rule"), some pages will be long and other pages will be short.

I do understand that not everyone is happy with having short pages on the wiki. But the advantage is simplicity. You would know exactly where to look for known information about a given rule, and where to add new information. It seems that this simplicity is always lost, whenever one attempts to draw the line between interesting rules that should have a page and interesting rules that should not have a page.

[GUYTU6J]

Currently the "Isotropic non-totalistic" family has over 14000 rules, and the /rules/isotropic page is very long. Would it be possible to split that page further into separate subpages that can be browsed one at a time, according to the number of toggled isotropic conditions away from Life? (In other words, a page for 1TFL rules, a page for 2TFL rules, a page for 3TFL rules, and so on. For the URL, one could probably use "/rules/isotropic/1tfl", "rules/isotropic/2tfl", etc.)
Yes, this subdivision is arbitrary and not perfect, but this would work for all isotropic rules, and it is the simplest subdivision that I can think of.
...

Splitting the page is a good idea, but the unnecessarily Life(B3/S23)-centric standard is not. A simpler division would be done according to the number of used isotropic transitions, i.e. those toggled away from B/S.*

*Although the rulestring for "nothing happens to any pattern at all" is in reality B/S012345678, it uses too many isotropic survival transitions.

If there needs to be a place to catalog n-transitions-away-from-Life rules (where 1≤n≤3 would be good enough), it would better be "/census/b3s23/proximity", a subpage of https://catagolue.hatsya.com/census/b3s23 linked from it. This may also be done to other outer-totalistic and isotropic non-totalistic rules such that we can check rules some transitions away from (say) MirrorLife.

[confocaloid]

Splitting the page is a good idea, but the unnecessarily Life(B3/S23)-centric standard is not. A simpler division would be done according to the number of used isotropic transitions, i.e. those toggled away from B/S.*

*Although the rulestring for "nothing happens to any pattern at all" is in reality B/S012345678, it uses too many isotropic survival transitions.

I did expect the "Life-centricity" objection. The reality is that Conway's Game of Life is the single most popular rule. These forums, the Catagolue and the LifeWiki are all centered around CGoL. If you attempt to make all rules equal, you risk losing all rules except B3/S23 (because obviously there are way too many and no reason to prefer some to others, apart from arbitrary reasons such as Life-centricity/popularity).

It is natural to try to tweak rules slightly and see what happens. Combine this with Life-centricity, and with the fact that many near-Life rules just happen to be interesting (while "toggled away from B/S" does not really tell much about interestingness of INT rules).

[pzq_alex]

I did expect the "Life-centricity" objection. The reality is that Conway's Game of Life is the single most popular rule. These forums, the Catagolue and the LifeWiki are all centered around CGoL. If you attempt to make all rules equal, you risk losing all rules except B3/S23 (because obviously there are way too many and no reason to prefer some to others, apart from arbitrary reasons such as Life-centricity/popularity).

It is natural to try to tweak rules slightly and see what happens. Combine this with Life-centricity, and with the fact that many near-Life rules just happen to be interesting (while "toggled away from B/S" does not really tell much about interestingness of INT rules).

I object for a different reason. It is natural to tweak rules, but the OCA community does not tweak Life itself all the time! If I am considering neighbors of a rule (say) 20tfl, then one-transition-away rules can end up in anywhere in {19,21}tfl, two-transitions-away rules can end up in {18,20,22}, and it gets linearly worse with higher distances. Consequently it's much more annoying to search and collect these rules, only because I'm researching a rule different from Life!

I'd think twice before giving some rule a special position. From a purely mathematical point of view, this breaks symmetry. Apart from history, this does not seem like a very well-supported choice. Are you suggesting a restructuring of Catagolue just to make your work of maintaining the 1tfl and 2tfl threads easier?

[confocaloid]

I object for a different reason. It is natural to tweak rules, but the OCA community does not tweak Life itself all the time! If I am considering neighbors of a rule (say) 20tfl, then one-transition-away rules can end up in anywhere in {19,21}tfl, two-transitions-away rules can end up in {18,20,22}, and it gets linearly worse with higher distances. Consequently it's much more annoying to search and collect these rules, only because I'm researching a rule different from Life!

If you are exploring rules near (in some sense of "near") to a 20TFL rule, then the actual exploration of these rules is likely be much more "fun" (read: likely to take much more time and other resources) than the relatively simple task of making a list of rules that are covered by the exploration.
The Catagolue does not explore the rules for you! And Catagolue will not answer the question "Which rulespace to explore next?" for you -- that's your decision.

I'd think twice before giving some rule a special position. From a purely mathematical point of view, this breaks symmetry. Apart from history, this does not seem like a very well-supported choice.

As I already said above:

These forums, the Catagolue and the LifeWiki are all centered around CGoL.

I believe Catagolue (as well as LifeWiki and these forums -- I don't know much about Discord) already gives Conway's Game of Life a special position.

The "About" page on Catagolue says: "Catagolue is an online database of objects in Conway's Game of Life and similar cellular automata."
The Statistics page presents Life-specific information.
Life has its own rule family.
Most of soup search is done in b3s23: census.

Personally, I don't see anything wrong with this. There are very few things that make rules really interesting -- popularity and history are two such things.

Are you suggesting a restructuring of Catagolue just to make your work of maintaining the 1tfl and 2tfl threads easier?

I'm afraid I don't understand this question at all.

I do not "maintain" these threads. Both the 1TFL thread and the 2TFL thread are maintained by the community. These threads will live as long as the community is interested in keeping them alive.
There is also another thread for 1TFL rules: Oscillator Discussion Thread (1-transition)
Neither of these threads is created by me, and neither is "maintained" by me.

Apart from the forum threads, I know about a wiki page User:Pzq Alex/Rules 1 transition from Life (current version: link) which collects 1TFL rules. Most contributions to that page (including the page creation) are from other people. As of now, my contributions there are mostly formatting -- the actual content was already there.

[confocaloid]

To reply to the second part of the post: while I do agree that having something like "proximity-to-given-rule" for all Catagolued rules would be nice-to-have, I do not expect that this will be implemented. It seems like a significant change for a relatively minor benefit (as I attempted to explain above, whenever you're exploring rules that are in some sense "near to" a specific rule, it's not too hard to make a list of these rules, compared to the difficulty of actually exploring them).

My suggestion to split /rules/isotropic is motivated by the observation that that page is long -- so I thought that some reasonable way to split it would be nice, and this would be something reasonable to suggest here. In this context, splitting /rules/isotropic by "distance-to-Life" seems at least as good as any other comparable split. I think splitting by "distance-to-Life" is actually better; for one thing, it nicely highlights the fact that there are currently very many rules on Catagolue that are two toggles away from Life (see the table in my post above). On the other hand, splitting rules by "distance-to-zero" (with "zero" interpreted as the "no birth and no survival" rule) does not seem to highlight any such interesting observations.

[muzik]

For objects whose syntheses are split into multiple stages, would it be possible for links to the object pages for each stage to be added?

[squareroot12621]

For objects whose syntheses are split into multiple stages, would it be possible for links to the object pages for each stage to be added?

xs690_yboow65la4oy04a4zydccw121o4kmw696wggzy38kcxc…40118kiczyd123wo4owra961hxcczyj9k8y069kb8ow66zyt11 joined the game

[galoomba]

Would it be possible to track glider syntheses in non-B3/S23 rules such as tlife and DryLife?

I second this.