The New Sudoku Players' Forum

[denis_berthier]

in the chains the algorithm builds a link is added to the chain on the consequences of all the precedings links not only the last one: in R1C2 the 5 has been eliminated by the link R2C3, it is no longer in the cell when R1C8 is reached: is this remembering?

Yes, while (pure) AIC-links are independent from preceding ones.
Your's is similar to (some of) Denis' SudoRules chains. Finding such chains manually is harder, because there are many things to remember, when you follow one (you don't know in forward, which are useful).

It seems P.O. is trying to re-invent the wheel.
There is strictly nothing to remember - no more than for the most elementary bivalue-chain (= basic AIC). Once a partial-whip has been obtained, no z- or t- candidate used to prove it needs to be remembered for extending it.

[P.O.]

hi Denis. i didn't say there is remembering, it was a question not a affirmation.

in fact i don't think there is any remembering, when a link is added to the chain the state of the grid change.

it is in this new state that the search continues.

in this example the link R1C8 on 4 eliminates the 4 in R1C2 and set the 1.

it was to give a explanation that the link R2C3 was cited.

and i think that such clarification may be needed.

[denis_berthier]

hi Denis. i didn't say there is remembering, it was a question not a affirmation.
in fact i don't think there is any remembering, when a link is added to the chain the state of the grid change.
it is in this new state that the search continues.
in this example the link R1C8 on 4 eliminates the 4 in R1C2 and set the 1.
it was to give a explanation that the link R2C3 was cited.
and i think that such clarification may be needed.

Hi P.O.
In this case, what you're describing is T&E and your own way of keeping track of the intermediate inferences.

[P.O.]

Denis, i wont argue, you know better.
To say that the state of the grid is changing is to say that the eliminations done by the preceding links are eliminations for the next ones, that certainly must be enforced in some way. Many links are added to the current one building as many chains that will eventually lead to a elimination: to know which ones they must be tried against the start of the chain. But they are not chosen at random, but in accordance with the rules of the game. R1C2 is a bivalue cell when R1C8 is reached and all the preceding links are obtained by bilocation.
So the BFS algorithm builds a tree randomly over the grid but each node is added following the rules.

[denis_berthier]

Denis, i wont argue, you know better.
To say that the state of the grid is changing is to say that the eliminations done by the preceding links are eliminations for the next ones, that certainly must be enforced in some way. Many links are added to the current one building as many chains that will eventually lead to a elimination: to know which ones they must be tried against the start of the chain. But they are not chosen at random, but in accordance with the rules of the game. R1C2 is a bivalue cell when R1C8 is reached and all the preceding links are obtained by bilocation.
So the BFS algorithm builds a tree randomly over the grid but each node is added following the rules.

Of course, in T&E or in BFS (or for that matter in DFS also), changes are made according to the rules.