The New Sudoku Players' Forum

[mith]

FWIW, this is what yzf's solver has for the image posted:

Grouped Almost Locked Set W-Wing: A=r4c3-{48}, B=r9c1-{48}, connect by 8b1 =>
r5c1<>4 r7c3<>4

It doesn't appear to have Grouped W-Wing implemented separately from its ALS W-Wing implementation.

[SpAce]

FWIW, this is what yzf's solver has for the image posted:

Grouped Almost Locked Set W-Wing: A=r4c3-{48}, B=r9c1-{48}, connect by 8b1 =>
r5c1<>4 r7c3<>4

Ok, thanks for the clarification. So my assumption was wrong!

It doesn't appear to have Grouped W-Wing implemented separately from its ALS W-Wing implementation.

In that case I've now seen the first mistake by yzfwsf. Too bad. I thought I could make an exception to my overarching "trust but verify" policy

That said, a naming error is not very significant, except that it can cause confusion -- like we just saw. It's also not an actual error if it's clearly stated that the same name covers all three cases: Grouped W-Wings, ALS-W-Wings, and Grouped ALS-W-Wings. I just don't see how such a combined naming scheme would make sense, as it should be easy to separate them into distinct names. Also, 'Grouped ALS-W-Wing' is a specific type itself (one that has both grouped and ALS nodes), so it shouldn't be used as a common name to them all. Perhaps something like 'Complex W-Wing' could be used for that purpose, if need be.

PS. I've always hated that Hodoku and (if I remember correctly) SudokuWiki list ALS nodes under "grouped", which is a slightly different but similarly confusing practice. Group and ALS nodes are two very different things, even though both use multiple cells. If one wants to have a common name for both, it shouldn't be one of the two subtypes (or one of the three in the YZF case).

It's interesting that there are many similar categorization errors in sudoku (Turbot Fish, ALS, MSLS, ...) where one of the subtypes is a synonym for the supertype. It makes accurate communication very difficult and verbose if you want to specify which one you actually mean. I guess those who named them weren't OO-programmers.