Page 3 of 3
Re: Catagolue Suggestions Thread
Posted: July 14th, 2023, 3:03 pm
by wirehead
I have an idea that would prevent
this mess especially for multistate rules. It could be implemented as a sub-symmetry. I suggest adding the suffix "_culled" to it, so it would be called C1_culled.
When apgsearch is in a culled symmetry, after it identifies a blob of cells as an oscillator it would start removing stator cells at random and stop when it can't remove any cell without breaking the oscillator. The result, should hopefully be the minimum form, without all of the random junk around it as in the example above (they're all the same xp2 just with different random crap around them, as Yoel already noted in the Sticky thread.)
Re: Catagolue Suggestions Thread
Posted: November 3rd, 2023, 1:00 am
by confocaloid
benetnasch85 wrote: October 5th, 2022, 12:06 am
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?
I created a userpage that lists some more Life-like rules than the "Most popular" list:
User:Confocal/Life-like rules
It is not auto-updated, but maybe it will be useful.
muzik wrote: May 25th, 2023, 10:46 am
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?
What about adding only a single link to the object page for the previous stage? (So
Gabriel's p138 plus eight blocks could give a link to
Gabriel's p138 plus four blocks, which in turn would link to
Gabriel's p138.)
Alternatively, add links for up to N previous stages, for some small N.
Re: Catagolue Suggestions Thread
Posted: November 14th, 2023, 8:17 pm
by Para
Does there exist an "all-soups" census (that, obviously enough, tabulates all soups) in a given rule?
Re: Catagolue Suggestions Thread
Posted: November 15th, 2023, 4:42 pm
by dvgrn
Para wrote: November 14th, 2023, 8:17 pm
Does there exist an "all-soups" census (that, obviously enough, tabulates all soups) in a given rule?
Not quite clear on the question -- you don't specify a bounding box, and without that, there isn't any such thing as "all soups".
You might be looking for
this existing apgsearch hack, though. That works up to somewhere around 5x6 soups -- or it used to: now that I look again, it's old enough that it would be the Python2 version of apgsearch (which has never been ported to Python3). Older versions of Golly and Python would be needed to run it.
EDIT: Oh, right -- more recently, exhaustive enumerations of small soups have been piped in to Catagolue's stdin symmetry, producing census results like
this and
this. Depending on whether you're looking for a simple census or something trickier like longest-lasting diehard or methuselah, those tabulations might or might not have what you need.
Re: Catagolue Suggestions Thread
Posted: November 15th, 2023, 4:45 pm
by hotdogPi
dvgrn wrote: November 15th, 2023, 4:42 pm
Para wrote: November 14th, 2023, 8:17 pm
Does there exist an "all-soups" census (that, obviously enough, tabulates all soups) in a given rule?
Not quite clear on the question -- you don't specify a bounding box, and without that, there isn't any such thing as "all soups".
You might be looking for
this existing apgsearch hack, though. That works up to somewhere around 5x6 soups -- or it used to: now that I look again, it's old enough that it would be the Python2 version of apgsearch (which has never been ported to Python3). Older versions of Golly and Python would be needed to run it.
There's literally something called "all-soups" in existence:
http://catagolue.hatsya.com/census/b3s23/all-soups/. Para is asking whether this could exist for non-Life rules.
Re: Catagolue Suggestions Thread
Posted: February 15th, 2024, 1:09 pm
by confocaloid
- Maybe it would be possible to add a functionality to identify a known rotor, similar to https://catagolue.hatsya.com/object but for oscillator rotors?
- Can testing/unofficial censuses be excluded from the "Most popular" list? That would allow more censuses in official (pseudo-)symmetries to be shown instead.
Re: Catagolue Suggestions Thread
Posted: August 20th, 2024, 11:43 pm
by confocaloid
Would it be feasible to adjust Catagolue to respect the distinctions between pseudo objects (example:
xs8_rr), quasi objects (example:
xs8_330oo), and constellations (example:
xs8_33w66)?
If that is done, then in every "synthesis-costs" tabulation, all the strict objects would come first (as they currently do), then followed by all the pseudo objects, then followed by all the quasi objects, and only then followed by arbitrary constellations at the end. In other words, the objects would be grouped by "interestingness".
A desirable side effect would be that if some tabulation becomes too large, then one would begin losing some (relatively uninteresting) arbitrary constellations, but one would never lose any quasi object before losing all arbitrary constellations, and one would never lose any pseudo object before losing all quasi objects.
Another desirable side effect would be that it would be easier to look through all pseudo objects in a certain tabulation, as those would be tabulated immediately after all the strict objects.
calcyman wrote: April 21st, 2019, 12:20 pm
Goldtiger997 wrote:Posting here to announce that all syntheses of 15-bit still-lifes have been reduced to below 14 gliders. This result can be seen
here on catagolue. However, in it's current state the list is slightly confusing since it also includes some pseudo still-lifes. Looking at the list, you can see that all of the still-lifes marked as costing 14 or more gliders are actually pseudo still-lifes.
The Catagolue update process now runs all of the new Shinjuku objects through apgsearch, thereby determining which ones are true and which are pseudo. By a dodgy hack (internally adding 10^17 to the glider count for true objects), all true objects are displayed above all pseudo-objects in Catagolue's synthesis-costs tabulations. For example:
https://gol.hatsya.co.uk/census/b3s23/s ... ffset=3200
Line 3286 contains the cheapest true xs16 (the 3-glider bipond), and line 3287 contains the most expensive pseudo xs16 (the 27-glider block on snake-tie-snake). This is perfectly consistent with the OEIS count of 3286 strict still lifes of 16 bits.
[...]
EDIT: Aha, this trick of pushing true still-lifes to the top has the fortunate side-effect that if a tabulation exceeds 900 KB compressed, it will only lose pseudo still-lifes (because the ones towards the bottom drop off). That's good if we ever attempt a '20 in 20' project.
confocaloid wrote: August 20th, 2024, 5:38 pm
May13 wrote: August 19th, 2024, 8:01 pm
[...] How about synthesising 21-bit pseudo still lifes? There are a total of 273188 strict and 179011 pseudo 21-bit still lifes, which is less than 672172, so it's definitely doable.
As far as I know, Catagolue doesn't "understand" the difference between a pseudo still life, a quasi still life, and an arbitrary constellation.
For example, from the total number of 8-bit strict still lives and 8-bit pseudo still lives, it would be reasonable to expect to see the biblock (
xs8_rr) on the first page of the tabulation
census/b3s23/synthesis-costs/xs8. However, the biblock is not present on the first page; instead, that page lists all nine 8-bit strict still lives, followed by many stable 8-bit constellations.
This might become a problem with higher bit-counts (e.g.
xs21), if/when the synthesis-costs tabulation includes so many (relatively uninteresting) constellations that some (relatively more interesting) pseudo still lives and/or quasi still lives are lost.
Re: Catagolue Suggestions Thread
Posted: March 28th, 2025, 7:49 pm
by LuveelVoom
Could yl and zz (generally, all patterns that display a ??) open a sample soup in the LV applet when you go to their page? It's very tedious right now to have to copy the sample soup and paste it in a viewer tool.
Re: Catagolue Suggestions Thread
Posted: March 28th, 2025, 8:19 pm
by confocaloid
LuveelVoom wrote: March 28th, 2025, 7:49 pm
Could yl and zz (generally, all patterns that display a ??) open a sample soup in the LV applet when you go to their page? It's very tedious right now to have to copy the sample soup and paste it in a viewer tool.
In general, the choice of a sample soup would be arbitrary, and there can be differences between growths tabulated as the same entry.
If some "yl..." growth is very common, there can be no sample soup links at all (soups for common entries aren't stored).
Unlike apgcodes beginning with "x", for yl/zz it is impossible to unambiguously recover the pattern from the descriptor.
It could be possible to allow choosing any sample soup (when the soups are available) and viewing the chosen sample soup directly on the page, in some way. In that case it could work for xs/xp/xq objects too.