Lol yeah. It's very broken.
Bro says he's not very familiar with scratch.
Haha. I use C# and Lua more often, scratch is when i'm bored.
Just don't scratch yourself with that sharp C...
= n). This is 6 for the first turn and 5 for all subsequent turns unless the previous two turns were parallel, in which case it’s 4. You just need to generate a random number from 1 to n (or 0 to n-1 if you swing that way), then skip over the previous turn or axis (I’m assuming the face assignments group parallel faces together, for example U D F B R L). You could use conditionals:IF n = 6 THEN t2 = r
IF n = 5 THEN IF r >= t1 then t2 = r + 1
IF n = 4 THEN IF Axis(r) >= Axis(t1) THEN t2 = r + 2SELECT CASE n
CASE 5: t2 = r - (r >= t1)
CASE 4: t2 = r - 2 * (Axis(r) >= Axis(t1))
CASE 6: t2 = r
END SELECT
I mean cool. But haven’t you put in a little too much effort into a post trying to improve a scratch timer that deletes your solves everytime you run it (adding cloud variables would just be a waste, but if the owner wants that I don’t mind helping)I can't help you with Scratch because my programming never progressed beyond QuickBASIC 4.5 (a structured version that allows you to avoid line numbers and spaghetti code). I have written my own scramble generator, though, so I can discuss methodology, I just can’t give examples in contemporary code. I also give more attention than probably necessary to minimizing runtime, but I started with a $100 Timex/Sinclair, and old habits die hard (you could speed that thing up by moving the most frequently called routines closer to the beginning of the program).
To generate a random scramble without cancellations, you need to avoid turning a face twice in a row and after turning the parallel face, so no UU or UDU. I will use t2 for the current turn (the one you are generating now), t1 for the previous one, and t0 for the one before that.
An easy but inefficient method would be to re-randomize if t2 = t1 for the first case, or if Parallel(t0,t1) AND Parallel(t1,t2) for the second, where Parallel is True if the two turns are parallel. However, generating a pseudorandom number is slower than a few conditionals and integer operators, so let’s only do it once per turn. This means you need to determine the number of allowable faces for each turn
r:= your randomly generated number; n, t0, t1, t2 as before. Axis(x):= (x+1)\2 (or x\2 if you index from 0) where \ denotes integer division.
Code:IF n = 6 THEN t2 = r IF n = 5 THEN IF r >= t1 then t2 = r + 1 IF n = 4 THEN IF Axis(r) >= Axis(t1) THEN t2 = r + 2
Example: t1=3 (F in my assignment above). If n=5, an r of 1 or 2 is unchanged, while 3-5 --> 4-6, so the possible outcomes are U D B R L. If n=4, Axis(t1)=2 and r=1,2 is again unchanged, and 3,4 --> 5,6, so the outcomes are U D R L.
QuickBASIC allows Boolean operations (False:=0, True:=-1) and has a CASE statement that allows you to specify code for different values of a variable without repetitive conditionals, so my version looks like this (the order of the CASEs is because it drops out of the SELECT block after a True case, do this minimizes the number of CASEs evaluated):
Code:SELECT CASE n CASE 5: t2 = r - (r >= t1) CASE 4: t2 = r - 2 * (Axis(r) >= Axis(t1)) CASE 6: t2 = r END SELECT
That’s all you need for odd cubes. Even cubes have one more restriction – you can’t make two parallel turns if both are the maximum width. For example, on a 4x4, ud (or UwDw) has the same effect as u2/Uw2 and ud’/UwDw’ would just be y. On a 6x6, only three-wide turns would be restricted. You need a conditional like IF (c MOD 2 = 0) AND (n = 5) THEN IF Tmax(t1) = Tmax(t2) THEN n = 4 where c:= size of the cube, MOD is the remainder of integer division and Tmax(x):= (width(x)=c\2).
Note: On a 2x2, the above would reduce to n=6 for the first turn and n=4 for all others, if you generate 6-face scrambles for 2x2. They are usually scrambled using only UFR, so you would use n=3/n=2, and wouldn’t need to worry about parallel axes.
Thank you! I will try implement all your suggestions.I can't help you with Scratch because my programming never progressed beyond QuickBASIC 4.5 (a structured version that allows you to avoid line numbers and spaghetti code). I have written my own scramble generator, though, so I can discuss methodology, I just can’t give examples in contemporary code. I also give more attention than probably necessary to minimizing runtime, but I started with a $100 Timex/Sinclair, and old habits die hard (you could speed that thing up by moving the most frequently called routines closer to the beginning of the program).
To generate a random scramble without cancellations, you need to avoid turning a face twice in a row and after turning the parallel face, so no UU or UDU. I will use t2 for the current turn (the one you are generating now), t1 for the previous one, and t0 for the one before that.
An easy but inefficient method would be to re-randomize if t2 = t1 for the first case, or if Parallel(t0,t1) AND Parallel(t1,t2) for the second, where Parallel is True if the two turns are parallel. However, generating a pseudorandom number is slower than a few conditionals and integer operators, so let’s only do it once per turn. This means you need to determine the number of allowable faces for each turn
r:= your randomly generated number; n, t0, t1, t2 as before. Axis(x):= (x+1)\2 (or x\2 if you index from 0) where \ denotes integer division.
Code:IF n = 6 THEN t2 = r IF n = 5 THEN IF r >= t1 then t2 = r + 1 IF n = 4 THEN IF Axis(r) >= Axis(t1) THEN t2 = r + 2
Example: t1=3 (F in my assignment above). If n=5, an r of 1 or 2 is unchanged, while 3-5 --> 4-6, so the possible outcomes are U D B R L. If n=4, Axis(t1)=2 and r=1,2 is again unchanged, and 3,4 --> 5,6, so the outcomes are U D R L.
QuickBASIC allows Boolean operations (False:=0, True:=-1) and has a CASE statement that allows you to specify code for different values of a variable without repetitive conditionals, so my version looks like this (the order of the CASEs is because it drops out of the SELECT block after a True case, do this minimizes the number of CASEs evaluated):
Code:SELECT CASE n CASE 5: t2 = r - (r >= t1) CASE 4: t2 = r - 2 * (Axis(r) >= Axis(t1)) CASE 6: t2 = r END SELECT
That’s all you need for odd cubes. Even cubes have one more restriction – you can’t make two parallel turns if both are the maximum width. For example, on a 4x4, ud (or UwDw) has the same effect as u2/Uw2 and ud’/UwDw’ would just be y. On a 6x6, only three-wide turns would be restricted. You need a conditional like IF (c MOD 2 = 0) AND (n = 5) THEN IF Tmax(t1) = Tmax(t2) THEN n = 4 where c:= size of the cube, MOD is the remainder of integer division and Tmax(x):= (width(x)=c\2).
Note: On a 2x2, the above would reduce to n=6 for the first turn and n=4 for all others, if you generate 6-face scrambles for 2x2. They are usually scrambled using only UFR, so you would use n=3/n=2, and wouldn’t need to worry about parallel axes.
Yes, my problem was that i used a list with all the notation, and i used a variable that chose one of them as long as it wasnt the same as the last one, except i detected it incorrectly. When i have the time I will improve it
thanks guys!
Ok so i made literally everything better. 2x2 scramble function, scramble is fixed, next scramble option, inspection no longer breaks, next step is to add codes to save your times and also add sessions.
