The New Sudoku Players' Forum

[ArkieTech]

Code: Select all

 *-----------*
 |8..|5..|3..|
 |...|...|.64|
 |...|..4|285|
 |---+---+---|
 |6..|8..|5..|
 |5..|642|..9|
 |..1|..9|..8|
 |---+---+---|
 |413|7..|...|
 |72.|...|...|
 |..5|..3|..1|
 *-----------*


Play/Print this puzzle online

[SCLT]

Code: Select all

+----------------+-------------------+----------------+
|  8  e49  e249  |  5   f1269  16    |  3   e19   7   |
|  13  5    279  | g129  378   78    | d19   6    4   |
|  13  679  679  |  19   37    4     |  2    8    5   |
+----------------+-------------------+----------------+
|  6   349  49   |  8    17    17    |  5    234  23  |
|  5   378  78   |  6    4     2     | c17  b137  9   |
|  2   47   1    |  3    5     9     |  6   b47   8   |
+----------------+-------------------+----------------+
|  4   1    3    |  7    2689  568   |  89   259  26  |
|  7   2    68   |  149  1689  1568  |  489  359  36  |
|  9   68   5    |  4-2  268   3     |  478 a27   1   |
+----------------+-------------------+----------------+


A slightly long chain today, stte eliminations seem quite hard to come by in this one!

(2=7)r9c8 - c8r56 = (7-1)r5c7 = r2c7 - (1=492)r1c238 - r1c5 = r2c4 => -2 r9c4; stte

[SpAce]

Code: Select all

.-------------------.---------------------.----------------.
| 8   49+   a49[+2] |  5      169-2  16   | 3     19    7  |
| 13  5      79-2   | d1(2)9  378    78   | 19    6     4  |
| 13  679    679    |  19     37     4    | 2     8     5  |
:-------------------+---------------------+----------------:
| 6   49+3   49+    |  8      17     17   | 5    b23#4  23 |
| 5   378    78     |  6      4      2    | 17    137   9  |
| 2   47     1      |  3      5      9    | 6    b47    8  |
:-------------------+---------------------+----------------:
| 4   1      3      |  7      2689   568  | 89    259   26 |
| 7   2      68     |  149    1689   1568 | 489   359   36 |
| 9   68     5      | c24     268    3    | 478  b27    1  |
'-------------------'---------------------'----------------'

UR(49)r14c23 using mixed +internal/#external

(2)r1c3 == (47,2)r469c8 - r9c4 = (2)r2c4 => -2 r1c5,r2c3; stte

[SpAce]

A slightly long chain today, stte eliminations seem quite hard to come by in this one!

(2=7)r9c8 - c8r56 = (7-1)r5c7 = r2c7 - (1=492)r1c238 - r1c5 = r2c4 => -2 r9c4; stte

Nothing wrong with that, but if you think it's long... there's always the hydraulic press:

(27)r9c87 = (71)r52c7 - r1c8 = (16,2)b2p234 => -2 r9c4; stte

[SteveG48]

Code: Select all

 *--------------------------------------------------------------*
 | 8     49    249   |  5     1269  16    |   3     19    7     |
 |a13    5     279   |ac129   378   78    |  b9-1   6     4     |
 |b13    679   679   | b19    37    4     |   2     8     5     |
 *-------------------+--------------------+---------------------|
 | 6     349   49    |  8     17    17    |   5     234   23    |
 | 5     378  c78    |  6     4     2     |abf17    137   9     |
 | 2     47    1     |  3     5     9     |   6     47    8     |
 *-------------------+--------------------+---------------------|
 | 4     1     3     |  7     2689  568   |   89    259   26    |
 | 7     2    d68    |  149   1689  1568  |   489   359   36    |
 | 9    e68    5     |de24    268   3     |  f478   27    1     |
 *--------------------------------------------------------------*


1r2c14,r5c7 = 1r2c7&(19)r3c14&7r5c7 - (1|9|7=28)r2c4,r5c3 - (2|8)r9c4,r8c3 = (48)r9c24 - (4|8=71)r59c7 => -1 r2c7 ; stte

[SteveG48]

UR(49)r14c23 using mixed +internal/#external

(2)r1c3 == (47,2)r469c8 - r9c4 = (2)r2c4 => -2 r1c5,r2c3; stte

Oooh, that's nice!

[SpAce]

Oooh, that's nice!

Thanks, Steve! Btw, once again I first posted it without the comma -- like I've said, I'm not immune. One of these days you'll catch me!

I love yours too! I don't know how you come up with those. I might consider simplifying the beginning a bit:

(1)r2c1,r5c7 = (19)r3c14&(17)r25c7 - (9|1|7=28)r2c4,r5c3 ...

What do you think?

PS. A side question about Hodoku. Do you know how to use the color bar in the status line? I can only see one color of each pair but no way to select the other, which makes it useless. It'd be nice to be able to use colors without having to switch to the "Active Cell" view, but I can't figure it out.

[SteveG48]

I love yours too! I don't know how you come up with those. I might consider simplifying the beginning a bit:

(1)r2c1,r5c7 = (19)r3c14&(17)r25c7 - (9|1|7=28)r2c4,r5c3 ...

What do you think?

I like it!

PS. A side question about Hodoku. Do you know how to use the color bar in the status line? I can only see one color of each pair but no way to select the other, which makes it useless. It'd be nice to be able to use colors without having to switch to the "Active Cell" view, but I can't figure it out.

I think I know what you want to do. In coloring mode, you have two choices on the color bar. One is the choice of color pair, and the other is to have coloring effective on candidates or on whole cells. Make your choices and commence to coloring, with the shift key switching between the two colors in the pair. You can select a new color pair at any time without erasing what you've already done. Likewise, you can change between cell mode and candidate mode at any time. If you select white as your color, you are actually suspending coloring and going to normal mouse operation without affecting your existing coloring. Go back to coloring by selecting a new color. Erase all colors and return to normal operation by selecting R.

[SpAce]

with the shift key switching between the two colors in the pair

So it seems. Thanks! I swear I'd tried pressing "shift" (and "ctrl" too) but nothing happened before. I also tried to "RTFM" but found no hint. Now it works, and strangely enough I can easily find it mentioned in the documentation too. I'm seriously baffled!

Actually, I think I expected the palette itself to change in appearance when pressing "shift" (or something) but it doesn't. Only the color under the cursor tip does, but that's not visible if the cursor is near the palette, so I missed it. However, I thought I tried this way too, but I guess I didn't.

Kind of a bummer that it doesn't apparently let you lock the other color this way, i.e. you have to keep pressing "shift" to use it (or otherwise choose it from the main palette)? Still, this probably makes coloring quicker. So, thanks again! This was very helpful, just like your previous tip about the progress indicator (which actually led to discovering some other overlooked features too, including this one)

[SteveG48]

I'm glad that I could help.

One other thing to consider is that you can change default colors in the setup menus. The default pairs are darker and lighter shades of the same color. I keep that mostly, but I find it convenient to make one pair two entirely different colors.

[SpAce]

One other thing to consider is that you can change default colors in the setup menus. The default pairs are darker and lighter shades of the same color. I keep that mostly, but I find it convenient to make one pair two entirely different colors.

Yeah, changing the default colors (and saving the config) was the first thing I did when I started using Hodoku coloring. I think the two rightmost default color pairs are useless because they're way too light and close to each other. I prefer stronger contrast. Like you, I also have one pair with different colors (used mostly for found placements and eliminations). Here's the setup I've been using since the beginning:

Hodoku colors: Show

colors.png (11.92 KiB) Viewed 545 times

I'd actually like to have a few more pairs, but those five are just enough for basic GEM (not my wished extensions, though). I guess they're about enough for humanly applicable multi-coloring too, but at least in my few recent experiments with that I've soon run out of colors. I don't expect multi-coloring to become a bread-and-butter solving technique for me anyway, but I like the concept, and I'm thinking it might have some use for initial reconnaissance.

Btw, too bad Hodoku doesn't support multiple concurrent colorings (except by starting multiple instances of the program), or even saving and restoring a coloring (as far as I know). That'd be useful in more complicated puzzles.

[SteveG48]

Btw, too bad Hodoku doesn't support multiple concurrent colorings (except by starting multiple instances of the program), or even saving and restoring a coloring (as far as I know). That'd be useful in more complicated puzzles.

I'd like to be able to erase candidate coloring while maintaining cell coloring, and vice-versa. <shrug> Can't have everything I guess; Hodoku is fantastic as it is.

[SpAce]

I'd like to be able to erase candidate coloring while maintaining cell coloring, and vice-versa.

I've never really missed that because I don't much use cell coloring anyway (mostly just for temporary highlighting of DP or fish cells etc). Now that you mentioned it, that behavior is weird anyhow. Both panels have separate R-buttons, so intuitively they should also work separately. Should be a very easy fix if someone bothered to find it in the code... and since I said that, I just tried to find it. Not exactly easy, but I think I found the ballpark.

Java stuff: Show

In the SudokuPanel.java file there's this method:

Code: Select all

    public void clearColoring() {
        coloringMap.clear();
        coloringCandidateMap.clear();
        setActiveColor(-1);
        updateCellZoomPanel();
        mainFrame.check();
    }


The first two lines apparently reset both cell and candidate colorings whenever this method is called, which is why they both happen together. It's called from the keyboard event handler for the 'R'-key (same file):

Code: Select all

            case KeyEvent.VK_R:
                clearColoring();
                break;


That's triggered whenever the 'R'-key is pressed. Obviously the same happens when the mouse is clicked on the visible Rs, but I haven't found where that event is handled (I guess it might be turned into a key event and handled here as well, but that would be kind of weird).

What this means is that if we wanted to have separate actions for the cell and candidate coloring resets, we'd just need to split the single clearColoring method into two (clearCellColoring and clearCandidateColoring, each with one or the other of the first two lines of the original) and have them called from the appropriate mouse click handlers for the two different Rs. That much is easy. Problem is, most of the UI code is generated and a mess. It seems very hard to find anything, like where the damn 'R's and their mouse events are defined.

Added. I think I found it. The relevant mouse click handler is in the MainFrame.java:

Code: Select all

   public void coloringPanelClicked(int colorNumber) {
        if (colorNumber == -1 || colorNumber == -2) {
            statusPanelColorResult.setBackground(Options.getInstance().getDefaultCellColor());
            sudokuPanel.setActiveColor(-1);
            if (colorNumber == -2) {
                sudokuPanel.clearColoring();
                repaint();
            }
        } else {
            statusPanelColorResult.setBackground(Options.getInstance().getColoringColors()[colorNumber]);
            sudokuPanel.setActiveColor(colorNumber);
        }
    }

colorNumber -2 apparently maps to the 'R' button, and -1 maps to the 'white' button. So, this method would need to be changed to call one or the other clearColoring method depending on which 'R' was pressed (that would need to be passed as a boolean parameter, which the caller already has). That's it, I think. Very easy and minor changes, but hard to locate because of the messy and undocumented UI architecture. Then again, can't really complain because:

<shrug> Can't have everything I guess; Hodoku is fantastic as it is.

Yes, it is. And that's good, because it's clearly not built or documented for anyone else to upgrade and to maintain But, based on this little exercise I don't think it would be impossible.