Thread for 90 degree paired slow salvo syntheses of spartan and non-spartan components

[Wyirm]

I'm currently looking into 90 degree paired slow salvo synthesis for a universal constructor idea (see this thread)https://www.conwaylife.com/forums/viewt ... f=7&t=5450
Before this method can really take off, we need to know more about how 90 degree slow salvo synthesis.
This thread is reserved for 2 glider, 2 gider and object, and 1 glider and object reactions. Currently 90 degree paired slow salvo synthesis hasn't been looked into nearly as much as slow salvo, and I intend to pioneer this exploration.

[GUYTU6J]

To moderators: I admit this is not constructive, but I suggest merging this spin-off to the main thread so as to avoid clutter. We can develop both circuitry and syntheses in the same place, at the same time. And as the circuitry is limiting the pool of usable recipes, it makes more sense to investigate them together.
…
To begin with, a list of 90-degree collisions is shown in 2-glider collision#Classification of the results.

[dvgrn]

To moderators: I admit this is not constructive, but I suggest merging this spin-off to the main thread so as to avoid clutter...

This is a reasonable suggestion, but I'm not going to move anything for the moment (though someone else can if they want to). This is a potentially important topic, and I'm perfectly happy for it to have its own thread.

The main problem is that people are likely to start posting single recipes on this thread, and that has the potential to get so unmanageable so quickly that I'd rather have that mess on a separate thread, rather than having to page through dozens of singleton recipe posts in what could otherwise be a readable circuitry-development thread.

To make progress on paired slow salvo syntheses, honestly, a search script should be written sooner rather than later. Why bother inventing hundreds of basic recipes manually when a script can produce them all in a standard format usable by a compiler, and can pretty much guarantee that they're the most efficient ones available?

Prior Art
As a reminder, there is a good existing starting point for a compiler script -- calcyman wrote synthesise-pattern2.py way back in 2010 (message #2681!) A number of useful slow-90-degree-glider-pair recipes are embedded in it. I think at one point somebody made an RLE stamp collection showing all of the synthesise-pattern2 recipes, but offhand I can only find this random sampling.

A good first step might be for someone to update that script to Python3 (if necessary, I haven't checked) and figure out whether its internal representation of the various recipes is good enough to use, or if it should be translated into something more modern. These days I would tend to use a custom rule and RLE to record the neighbor-count envelopes, for example. And since these recipes can all be looked at as converters from Object A to Object B using two gliders, it might be really good to use Shinjuku format to record them all.

Really it probably makes sense to port the script to something based on lifelib, partly to allow for much faster searching for new recipes when we get to that point.