A factor of four is quite an improvement -- congratulations!
calcyman wrote: October 20th, 2019, 6:39 am
danny wrote: October 20th, 2019, 3:27 am
I don't know how to submit to the catagolue box and I have to go to bed, I would politely request someone submit it and direct me to the necessary programs.
It literally just requires copying this RLE, pasting it in the box at the bottom of
https://catagolue.appspot.com/syntheses and pressing Post.
The idea of the box was to allow people to submit syntheses
without needing to mess around with the internals of Shinjuku.
The fabulously nonmagical box also accepts incremental syntheses. According to
this discussion on Discord it might be pretty simple to get Catagolue to prefer a multi-stage synthesis to a single-stage one --
Exa wrote:that second one is mostly a matter of setting "prefer_longest=True" in wherever the shinjuku dijkstra() call is
it's already a parameter that exists
-- but for right now whichever one gets in first will get stuck in the database. So I've gone ahead and submitted the following in hopes it will get in ahead of the other one, since continuous syntheses kind of obfuscate the stage information:
Code: Select all
x = 109, y = 25, rule = B3/S23
15bo57bo21bo$14bobo57b2o2bobo13bobo$14b2o57b2o4b2o13b2o$79bo2$bo78b2o$
o80b2o$3o77bo$100bo$b3o95b2o$3bo81b2o12bobo4b3o$2bo81bobo19bo$85bo21bo
$5bo$4b2o$4bobo86b2o$92b2o$94bo5$94b2o$94bobo$94bo!
Those are the individual stages that Catagolue doesn't already know about yet, with at least 20 blank cells between the groups. The system doesn't mind duplication, it just discards what it doesn't need. But some of the stages were a little too close to each other. Now it's just a matter of waiting up-to-eight hours for the next automatic update.