Good question. I can't claim to have any kind of definitive answer, but ...get_Snacked wrote: September 19th, 2024, 1:26 pm why do we love this stuff? and why didn't we love this stuff before?
Glider syntheses have had big theoretical implications in CGoL since fairly early on. Maybe not so much back when Conway-and-graduate-students-and-I-think-Richard-Guy (heh, the "Conway group") were putting together their proof of CGoL's computational universality -- for that you just need logic circuitry, you don't need it to be glider-constructible. But certainly by the time construction universality was being discussed, it was obviously very important to know how to assemble the various basic building blocks of circuitry.
In 1981 Conway wrote: "If (ask Gosper) gliders can crash to form a pentadecathlon, then I can produce self-replicating machines, and it's undecidable whether any given machine is self-replicating or not." (page 20 of this PDF) ... Interesting that Conway didn't know that detail in 1981, since there had been pentadecathlon recipes around since 1973 (5 gliders) and 1974 (4 gliders). I guess that was a very early example of the Optimization Game for syntheses!
It looks like David Buckingham's seminal super-ambitious synthesis project, producing recipes for all still lifes up to 14 bits, didn't happen until sometime in the 1990s -- or at least didn't get finished until then. And then Mark Niemiec's database collecting and expanding on those results has been a source of wonder and inspiration and useful information ever since, supplemented by Catagolue's lists once those became available.
Point being, there seems to be pretty good evidence of people having lots of fun crashing gliders together, for over half a century now.
If indeed that is what's happening, maybe it's because glider syntheses are an area of CGoL that has a learning curve that doesn't look too scary. It's easy to dive in and start doing some good mad-scientist stuff, crashing gliders together and seeing what happens.
Glider-crashing experiments seem to have been a very popular CGoL pastime for a long time, even among people who don't get into the theory much. It's kind of the CA equivalent of setting up dominoes and/or watching them fall over (which is also something that I personally used to do a lot, by the way, back in the previous millennium).
-- Now, sure, most of us look at things like the spacefiller synthesis and kind of shake our heads in amazement. That synthesis has about a hundred different tricks in it that most of us wouldn't have any idea how to re-discover, if the recipe were somehow lost. But that's just to show that there is a continuum of players of the glider-collision game, right up to the occasional synthesis grandmasters.
It's still quite possible for a newcomer to find a still life on Catagolue that has never been synthesized before, spend a little time with a tutorial or two, and figure out a brand-new synthesis based on some serendipitous soup. That's a contribution right out on the frontier of CGoL research, a solution to a never-before-solved problem.
It makes sense to me that that would be more gratifying, and therefore more popular, than spending weeks or months trying to study up far enough to reach the frontier of research for something like reducing RCT15 to RCT14, or building a 2D memory for a computer-replicator, or finding an oblique spaceship smaller than Sir Robin, or (name your favorite problem here).
