The New Sudoku Players' Forum

[ArkieTech]

Code: Select all

 *-----------*
 |...|...|5..|
 |.71|.82|9..|
 |..5|.19|...|
 |---+---+---|
 |8..|.4.|..6|
 |.9.|253|.8.|
 |1..|.7.|..2|
 |---+---+---|
 |...|79.|2..|
 |..7|13.|64.|
 |..9|...|...|
 *-----------*


Play/Print this puzzle online

[SpAce]

Code: Select all

.----------------------.--------------.--------------------.
| 9     A2348    2(3)8 | a34  6  7    |  5   a123   a1348  |
| 346    7       1     |  5   8  2    |  9    36    b34    |
| 2346   23468   5     | b34  1  9    | c48   2367   3478  |
:----------------------+--------------+--------------------:
| 8     B23     C2(3)  |  9   4  1    |  7    5      6     |
| 7      9       6     |  2   5  3    |  14   8      14    |
| 1      5       4     |  68  7  68   |  3    9      2     |
:----------------------+--------------+--------------------:
| 3456   13468   8-3   |  7   9  4568 |  2   d1(3)   1358  |
| 25     28      7     |  1   3  58   |  6    4      9     |
| 3456   13468   9     |  68  2  4568 | c18   137    13578 |
'----------------------'--------------'--------------------'

Kraken Row (3)r1c23489

(3)r1c2 - r4c2 = (3)r4c3
||
(3)r1c3
||
(3)r1c489 - (3=4)r3c4|r2c9 - (4=81)r39c7 - (1=3)r7c8

=> -3 r7c3; stte

[Cenoman]

Code: Select all

 +-----------------------+-------------------+----------------------+
 |  9      2348    238   |  34   6    7      |  5    123    1348    |
 | b346    7       1     |  5    8    2      |  9    36   dc34      |
 | b2346   23468   5     |dc34   1    9      | e48   2367   3478    |
 +-----------------------+-------------------+----------------------+
 |  8      23      23    |  9    4    1      |  7    5      6       |
 |  7      9       6     |  2    5    3      |  14   8      14      |
 |  1      5       4     |  68   7    68     |  3    9      2       |
 +-----------------------+-------------------+----------------------+
 | a3456   13468   8-3   |  7    9    4568   |  2   f13     1358    |
 |  25     28      7     |  1    3    58     |  6    4      9       |
 | a3456   13468   9     |  68   2    4568   | e18   137    13578   |
 +-----------------------+-------------------+----------------------+

(3)r79c1 = r23c1 - (3)r2c9&r3c4 = (4)r2c9|r3c4 - (48=1)r39c7 - (1=3)r7c8 => -3 r7c3; ste
My first idea, very close to SpAce's, so

Code: Select all

 +-----------------------+-------------------+----------------------+
 |  9      2348    238   |  34   6    7      |  5   g23-1   1348    |
 |  346    7       1     |  5    8    2      |  9    36     34      |
 | e2346   23468   5     |  34   1    9      |  48  f2367   3478    |
 +-----------------------+-------------------+----------------------+
 |  8      23      23    |  9    4    1      |  7    5      6       |
 |  7      9       6     |  2    5    3      |  14   8      14      |
 |  1      5       4     |  68   7    68     |  3    9      2       |
 +-----------------------+-------------------+----------------------+
 |  3456   13468  b38    |  7    9    4568   |  2   a13     1358    |
 | d25    c28      7     |  1    3    58     |  6    4      9       |
 |  3456   13468   9     |  68   2    4568   |  18   137    13578   |
 +-----------------------+-------------------+----------------------+

(1=3)r7c8 - (3=8)r7c3 - (8=2)r8c2 - r8c1 = r3c1 - r3c8 = (2)r1c8 => -1 r1c8; ste

[SpAce]

(3)r79c1 = r23c1 - (3)r2c9&r3c4 = (4)r2c9|r3c4 - (48=1)r39c7 - (1=3)r7c8 => -3 r7c3; ste

Nice! I thought there was probably a way to write it as an AIC but didn't bother to look around. Glad you showed how.

Something about my favorite topic, i.e. notation. We wrote this part differently:

You: (3)r2c9&r3c4 = (4)r2c9|r3c4
Me: (3=4)r3c4|r2c9

Now, I know that yours is the by-the-book way to write it and definitely correct. Mine is not fully correct, and some time ago I might have complained if someone wrote it that way. Then I saw Steve's solution here which used a similar shortcut. It looked a bit weird at first, but the more I thought about it, the more acceptable it started to feel, especially since it's more readable than the fully correct alternative. So, should we accept it?

The only problem I see is that there's only one operator | or & while the full form has both (one for the weak link and one for the strong link), so it does require some interpretation. The interpretation that works for both Steve's and mine is that the first operator (in both cases the weak link one) is skipped and the one shown applies to the final result read from left to right. If that's defined as a rule, then there shouldn't be much ambiguity.

[SteveG48]

Code: Select all

 *----------------------------------------------------------------------*
 | 9      2348   238    |  34     6      7      |  5      123    1348   |
 |b346    7      1      |  5      8      2      |  9      36   cd34     |
 |b2346   23468  5      |cd34     1      9      |cd48     2367   3478   |
 *----------------------+-----------------------+-----------------------|
 | 8      23     23     |  9      4      1      |  7      5      6      |
 | 7      9      6      |  2      5      3      |  14     8      14     |
 | 1      5      4      |  68     7      68     |  3      9      2      |
 *----------------------+-----------------------+-----------------------|
 |a3456   1468-3 8-3    |  7      9      4568   |  2     e13     1358   |
 | 25     28     7      |  1      3      58     |  6      4      9      |
 |a3456   13468  9      |  68     2      4568   | e18     137    13578  |
 *----------------------------------------------------------------------*


Looks like mine is essentially the same as SpAce's:

3r79c1 = 3r23c1 - 3r2c9&r3c4 = (48)r2c9,r3c47 - (8=13)b9p27 => -3 r7c23 ; stte

And exactly the same as Cenoman's first.

[Ngisa]

Code: Select all

+----------------------+-----------------+-----------------------+
| 9       2348     238 | 34    6    7    | 5     d123     c1348  |
| 346     7        1   | 5     8    2    | 9      36       34    |
|f2346    23468    5   | 34    1    9    | 48    e2367     3478  |
+----------------------+-----------------+-----------------------+
| 8       23       23  | 9     4    1    | 7      5        6     |
| 7       9        6   | 2     5    3    | a14    8       b14    |
| 1       5        4   | 68    7    68   | 3      9        2     |
+----------------------+-----------------+-----------------------+
| 3456    13468   i38  | 7     9    4568 | 2     i13       1358  |
|g25     h28       7   | 1     3    58   | 6      4        9     |
| 3456    13468    9   | 68    2    4568 | 8-1    137      13578 |
+----------------------+-----------------+-----------------------+


(1)r5c7 = r5c9 - r1c9 = (1-2)r1c8 = r3c8 - r3c1 = r8c1 - (2=8)r8c2 - (8=31)r7c38 => - 1r9c7; stte

Clement

[Cenoman]

Something about my favorite topic, i.e. notation. We wrote this part differently:

You: (3)r2c9&r3c4 = (4)r2c9|r3c4
Me: (3=4)r3c4|r2c9

[...]Then I saw Steve's solution here which used a similar shortcut

[...]the first operator (in both cases the weak link one) is skipped and the one shown applies to the final result read from left to right. If that's defined as a rule, then there shouldn't be much ambiguity.

1) Notation is not part of my favorite topics...
2) As you found some similar use in Steve's solutions, I was expecting some feedback from Steve before giving any statement.
3) My position: the shortcut has two drawbacks to me. First, it is a false representation of the logic, and second, it is confusing for newcomers that would try to follow the logic of a solution for their own learning.
The readability benefit seems not to balance equally these drawbacks.
Note that the AIC such as I have written it is fully bidirectional.
If you try to reverse (3)r79c1 = r23c1 - (3=4)r2c9|r3c4 - (48=1)r39c7 - (1=3)r7c8, you should write
(3=1)r7c8 - (1=84)r39c7 - (4=3)r2c9&r3c4 - r23c1 = (3)r79c1

The word "rule" is far too strong. At most, a convention. Now, if it is a commonly accepted shortcut here (on this forum) I accept to read it without reacting to the logical inconsistency. I use split nodes so rarely that I consider it useless in my own writing.

Just keep in mind the audience that you aim to be read by. The "happy few" already well informed and able to interpret rather unusual notations, or anybody entering there by chance and happy to find every day an almost easy puzzle ?

[SteveG48]

Something about my favorite topic, i.e. notation. We wrote this part differently:

You: (3)r2c9&r3c4 = (4)r2c9|r3c4
Me: (3=4)r3c4|r2c9

Now, I know that yours is the by-the-book way to write it and definitely correct. Mine is not fully correct, and some time ago I might have complained if someone wrote it that way. Then I saw Steve's solution here which used a similar shortcut. It looked a bit weird at first, but the more I thought about it, the more acceptable it started to feel, especially since it's more readable than the fully correct alternative. So, should we accept it?

We should not. I've reviewed what I wrote in the example, and it's wrong. Readability does not excuse error. I agree with Cenoman's response.

[SpAce]

So, should we accept it?

We should not. I've reviewed what I wrote in the example, and it's wrong. Readability does not excuse error. I agree with Cenoman's response.

Well, I guess that settles it. I was just following your lead I was a bit surprised you wrote it that way in the first place, knowing that you're a stickler for logical correctness. I was about to comment it then already, but then thought that it had some redeeming qualities and I might be tempted to use it myself (which happened sooner than I thought). I'm a stickler for logical correctness too, but I guess I'm also a pragmatist and more willing to accept certain shortcuts if I think they actually communicate the idea better and there's little risk of confusion. I still don't think it would be a huge mistake to use this particular shortcut, but I guess it's better to avoid it.

3) My position: the shortcut has two drawbacks to me. First, it is a false representation of the logic,

I fully agree with that one -- unless there's an agreed convention to interpret it correctly.

and second, it is confusing for newcomers that would try to follow the logic of a solution for their own learning.

Not so sure about that. I'm actually pretty sure most newcomers would find the shorter form easier to understand without ever realizing it's logically incorrect. I bet (from my own experience) that many of our logically correct notations make little sense to newcomers as they're not exactly intuitive. To fully understand the rules of writing and interpreting fully correct Eureka pretty much requires understanding how it translates into logical ORs, NANDs, and ANDs (and of course understanding how those work too). That part took me a while even with generous help from some veterans, so I think it's a pretty tall order in general. We've all seen examples from experienced solvers who still get confused about when to use | and when &, and it's no wonder because most people's intuition probably produces incorrect results. I know mine did.

Note that the AIC such as I have written it is fully bidirectional.

Yes, that would have been my main argument against the shortcut if I'd followed through about commenting Steve's use of it. I just thought about it some more, and my conclusion was that it was more or less irrelevant from any practical point of view. If the interpretation I suggested were accepted, then reading from right to left would work without flipping the operator, since it still applies to the same link (it just comes first, and not second, when read from right to left). Of course if you actually write the chain in its reversed form, then you have to flip it as you did:

If you try to reverse (3)r79c1 = r23c1 - (3=4)r2c9|r3c4 - (48=1)r39c7 - (1=3)r7c8, you should write
(3=1)r7c8 - (1=84)r39c7 - (4=3)r2c9&r3c4 - r23c1 = (3)r79c1

Not that hard, but surely an added level of complexity and easy to forget.

Just keep in mind the audience that you aim to be read by. The "happy few" already well informed and able to interpret rather unusual notations, or anybody entering there by chance and happy to find every day an almost easy puzzle ?

As I recently said in a similar discussion, I'm not exactly worried about imaginary newcomers. I'll change my mind if I actually see some that are interested in learning even the basics of Eureka. No sign of those, so far. If someone really wants to learn, all they have to do is observe and ask questions. That's how I learned, and it's pretty much the only way too, because I don't think the complex parts of Eureka (like split-nodes) and its extensions (like memory chains and other nets) are actually explained anywhere in detail, and some of the existing documentation is outdated anyway.

Thanks to both of you for your feedback!