RSW wrote:- Code: Select all
+--------------+--------------+---------------+
| 789 6 789 | 1 5 2 | 3 89 4 |
| 1 49 489 | 6 3 7 | 259 289 25 |
| 2 3 5 | 4 89 89 | 7 1 6 |
+--------------+--------------+---------------+
| 47 8 247 | 257 147 156 | 1246 3 9 |
| 49 5 2349 | 23 1489 16 | 1246 7 128 |
| 6 1 23479 | 237 4789 89 | 24 5 28 |
+--------------+--------------+---------------+
|b39 a29 6 | 57 17 15 | 8 4 23 |
| 458 47 48 | 9 2 3 | 15 6 157 |
|c35 *7-2 1 | 8 6 4 |d259 d29 2357 |
+--------------+--------------+---------------+
(2=9)r7c2 - (9=3)r7c1 - (3=5)r9c1 - (5=29)r9c78 => -2r9c2; stte
I assume this is similar to Denis Berthier's 2nd biv chain, but I don't understand that terminology.
It is close but not the same chain.
biv-chain[3]: r7n2{c2 c9} - b9n3{r7c9 r9c9} - r9n7{c9 c2} ==> r9c2≠2
has length 3; your chain has equivalent length 4 (in AIC notation, in case there are no inner Subsets, just count the number of = signs; in this case the end is a Subset[2]; so it counts for 2)
my biv-chain[3] uses 3 bivalue (i.e. in the present case bilocal) cells: r7n2, b9n3 and r9n7. I'm sure you can easily translate this into AIC notation.
If you want a better comparison, there is Leren's chain:
(3=2) r7c9 - r7c2 = (2-7) r9c2 = (7) r9c9 => - 3 r9c9; stte
In my notation, it's exactly:
r7c9{n3 n2} — c2n2{r7 r9} — r9n7{c2 c9} => r9c9≠3
which is close to my third chain; the only difference is, mine uses conjugacy in block b7 instead of column c2 (the two are valid).
In the nrc notation, super-symmetry is fully taken into account and there's no difference between bivalue and conjugacy; bivalue pairs/conjugate pairs always appear inside curly brackets. But, when needed, the type of CSP-Variable (or 2D-cell) before the brackets allows to know precisely what is meant.