Tawal wrote: February 10th, 2024, 9:47 am
It seems the script only searches for
2obj (or only use 2obj database).
It also seems that there is a "
3obj" database ...
Well, how to use this data ?
There are currently three octohash databases, using two slightly different hash methods unfortunately. We've got
the original octohash database -- all arrangements of up to two common small constructible objects inside a 12x12 bounding box
the octo3obj database -- all arrangements of three [blinker, block, beehive, half blockade, bi-block, teardrop]s inside 11x11 (all of these are smallish and can be created by colliding two gliders)
the octo3g three-glider collision database -- an octo-fingerprinted version of 2718281828's enumeration of 3G collisions
The original octohash database doesn't correctly de-duplicate results that are just different phases of the same p2 constellation.
The octo3g database suffers from the same problem as 2718281828's original enumeration: it's not quite a complete enumeration, and it also contains quite a few duplicates in different phases that nobody has taken the time to work through and remove.
These were all intended just as temporary experiments, to make the data searchable while someone figured out how to do the job properly. Using just those simple Python scripts to retrieve results from those databases is... at least a couple of orders of magnitude slower than the optimal solution.
I think we could get pretty close to optimal with a properly indexed set of octo-fingerprints stored on a server somewhere,
queryable nearly instantaneously from a browser page.