The latest version of the OSS project blueroux includes a method editor to let you define any solution method you like and get breakdowns of solves in your method of choice. This needs a lot of work to fix up and it'd be great to get some users of other methods testing them and defining them.
The method editor is also useful for defining more and less complex variations of the same method - e.g. Roux with mismatched blocks or centers should be possible.
The way it works is that you first define a method in the method editor. For example, here is what Roux looks like:
In this screenshot I have focused the 'CMLL' step, which shows the four top corners solved as well as the first two blocks. Importantly, 'U' is named as a 'free face' meaning any cube within U, U', or U2 of this configuration will be considered solved. Another feature that can be seen here is that a solution method is a DAG of steps; after 'fb' in this definition, the user may go to either 'ss_back' or 'ss_front' and ss_back then leads to 'sb', 'cmll', and 'solved'. In this example 'ss_front' also leads to 'sb'. Arbitrarily complex dags can be defined in this manner.
A solve breakdown against this method definition looks like this:
In this example, I've clicked on the fb->ss_back step, so the playback shows the dimmed completed blue/yellow first block, and bright cubies for the second square. This also exposes a bug in the generic reconstruction, because the L/L2 moves should be written as r/R2, but in the current solution editor there is no way to specify the frozen face (coming soon: the fb definition will say the L face is frozen after fb is complete, and thereafter all L moves will get rewritten into r moves instead, creating the same answer as the computer's answer shown in green above).
The masking for the cube and the computer analysis are meant to be generic for any solution method but much work needs to be done to define and test alternate solution methods like Fridrich, ZZ, etc. Here is an example of me being incompetent, attempting a Fridrich solve:
In this example we can see that at least the stickering and stage breakdowns work, but no computer analysis is appearing in green for the stage yet and my solution is hilariously bad since I don't remember the method at all.
So what's needed? Users who want to solve in other methods (or in Roux for that matter) or who want to define their own solution method, or who want to help on the codebase.
Please PM me if interested.
Thanks
The method editor is also useful for defining more and less complex variations of the same method - e.g. Roux with mismatched blocks or centers should be possible.
The way it works is that you first define a method in the method editor. For example, here is what Roux looks like:
In this screenshot I have focused the 'CMLL' step, which shows the four top corners solved as well as the first two blocks. Importantly, 'U' is named as a 'free face' meaning any cube within U, U', or U2 of this configuration will be considered solved. Another feature that can be seen here is that a solution method is a DAG of steps; after 'fb' in this definition, the user may go to either 'ss_back' or 'ss_front' and ss_back then leads to 'sb', 'cmll', and 'solved'. In this example 'ss_front' also leads to 'sb'. Arbitrarily complex dags can be defined in this manner.
A solve breakdown against this method definition looks like this:
In this example, I've clicked on the fb->ss_back step, so the playback shows the dimmed completed blue/yellow first block, and bright cubies for the second square. This also exposes a bug in the generic reconstruction, because the L/L2 moves should be written as r/R2, but in the current solution editor there is no way to specify the frozen face (coming soon: the fb definition will say the L face is frozen after fb is complete, and thereafter all L moves will get rewritten into r moves instead, creating the same answer as the computer's answer shown in green above).
The masking for the cube and the computer analysis are meant to be generic for any solution method but much work needs to be done to define and test alternate solution methods like Fridrich, ZZ, etc. Here is an example of me being incompetent, attempting a Fridrich solve:
In this example we can see that at least the stickering and stage breakdowns work, but no computer analysis is appearing in green for the stage yet and my solution is hilariously bad since I don't remember the method at all.
So what's needed? Users who want to solve in other methods (or in Roux for that matter) or who want to define their own solution method, or who want to help on the codebase.
Please PM me if interested.
Thanks
