Keyboard controls and piece filters in speedsolving

[Andrew Farkas]

Hello, hypercubers!

If you haven't seen the other thread about Hyperspeedcube, I recommend reading that first. My main motivation for making Hyperspeedcube was to see whether keyboard controls were viable for manipulating 4D puzzles, and that experiment was absolutely a success. They take some practice to get used to, but it's absolutely worth it if you're interested in speedsolving. Nevertheless, it's worth taking some time to discuss whether we want to allow these in speedsolves.

Keybindings

Keybindings are a total game-changer for speedsolving. First, a brief explanation of how keyboard-based puzzle controls work in Hyperspeedcube:

- Any number of keybinds can be bound to a key.

- A given keybinding can change which faces and/or layers are "gripped," and optionally twist the gripped faces by a fixed amount.

- Only one twist can be performed per keypress. (This is an arbitrary limit to prevent macros.)

The default keybinding set uses the WER/SDF/CV keys to grip the LUB/FIR/DB cells respectively, and the UOIJKL keys to execute 90-degree twists of the gripped cell. To perform a twist, hold one key from the first set and press another key from the second set. This is already quite fast, but there's also the option to bind a single key to certain moves. For example, the program also ships with two RKT presets which each bind many common RKT moves to single keys. Think of these presets as 4D hand positions, switching during the solve to make certain moves more or less ergonomic.

Unfortunately, it's pretty easy to take this to extremes and make a separate keybind set for every algorithm, where the letters QWERTYUIOP execute the moves of that algorithm in order. This is just a terrible reinvention of macros, which are generally not allowed in speedsolves. To prevent this, I propose that speedsolves making use of the keyboard be required to show the Keybinds reference on-screen for the duration of the solve, so that it's clear if they're using these "macros," as judged by community consensus or whoever moderates the speedsolving scene. This would also mostly answer the question "What keybinds do you use?" which would otherwise surely be asked in the video comments.

Piece filters

Hiding pieces is an essential feature for more complex puzzles, but should it be allowed in speedsolves of smaller puzzles such as 3^4? I believe so, for the following reasons.

- Searching for pieces is not an interesting aspect of speedsolving where there is lots of optimization to be done (in my opinion, as a speedsolver — others may disagree).

- Filtering pieces by color to immediately identify a specific piece takes more time, introducing a trade-off between manual and automated search.

What are your thoughts?

--

"Machines take me by surprise with great frequency." - Alan Turing

[Melinda Green]

On 9/2/2022 10:38 PM, Andrew Farkas wrote:

[...]

- Searching for pieces is not an interesting aspect of speedsolving where there is lots of optimization to be done (in my opinion, as a speedsolver — others may disagree).

I agree. I particularly like and recommend the piece-finding feature Roice added to Magic120Cell where you can hold a modifier key while clicking a piece to highlight the piece that belongs in its location. I can't imagine trying to solve any large puzzle without such a feature.

-Melinda

[Andrew Farkas]

[...] hold a modifier key while clicking a piece to highlight the piece that belongs in its location.

I'm not sure how to implement that feature in general for puzzles without fixed centers, but otherwise I think it would be an excellent addition.

[Matt Sheerin]

Hi Andrew,

I've been wanting to have keyboard controls for years. Not having them is why I didn't do many attempts at 2bld and why I would never have attempted 3bld. I was planning to eventually learn enough Java to do it myself, but it seems I now don't have to. Thank you so much!

I don't have huge amounts of free time and I'm still trying to de-rust at bld in general so it will take me quite a while to get to this, but I'm very excited.

As for how it works, I'm personally only interested in keys being able to do a single move at a time and no piece filtering, because that's how normal speedsolving works. I did use the piece filtering in my incomplete solve of the 120-cell, but a casual solve of a large puzzle feels very different to speedsolving. Finding pieces and tracking the ones you have found is part of the skill of being fast, though one I'm still bad at.

I do think speedrunning is a good comparison though, where the leaderboards often have different categories with the same goal but different rules for how you can get there. Maybe we could do the same if there's enough interest in different rules?

Matt 

--
You received this message because you are subscribed to the Google Groups "hypercubing" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hypercubing...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hypercubing/CALmdYqotm-K9o1Eap6qHY5qCp_3nRpqgw8WxrXjkgvfKY%2B9gHA%40mail.gmail.com.

[Blobinati Cuber]

I understand both points of view. There is something impressive about watching a "pure" speedsolve v.s. an optimized one with piece filtering and advanced keyboard controls. This also brings to mind the macro v.s. macroless solves. For a while now, I've wanted to have a speedrun.com page for hypercubing, as it would be like one of the best places to host the virtual hypercubes, with the different programs and puzzles. But unfortunately, they do not allow rubik's cube games. But maybe an exception can be made for hypercubes because they are different, and purely virtual (besides the physical puzzles). maybe it's worth a shot, I'll try it one day. Anyways, I think that different categories are perfectly valid, and I would participate in a few different ones myself.

- Rowan