Previous Up Next
Dutch / Nederlands

Diary, June 2009

Sun Mon Tue Wed Thu Fri Sat
      1   2   3   4   5   6
  7   8   9  10  11  12  13
 14  15  16  17  18  19  20
 21  22  23  24  25  26  27
 28  29  30

Wednesday, June 3, 2009

Playing Go

I went to the university to play Go. When I arrived, Taco was already there. He had not brought a set him. We decided to play with the Chinese set I got on Thursday, July 27, 2006 from Mrs. Guo, a friend of my mother-in-law. We decided on a seven stone handicap for Taco. I resigned the game when I lost one of my groups. Taco showed me that I still had enough possibilities to save the group. We replayed the game. At home I tried to recollect the game record. Although I did reconstruct some game record that has all the stones in the right place at the end of the game, I am not sure about the order of the moves.

Thursday, June 4, 2009


This morning I went voting for the European Parliament and I had to vote with a red pencil for the first time in 28 years and for the second time in my life. The first time was on May 26, 1981, when I voted for the first time in my life for the House of Representatives of the Netherlands. All the other times, I voted with a voting machines. But some court order last year forbade the use voting machines, because that there is a small chance that they are not reliable or that could be tapped.

Sunday, June 7, 2009

Go tournament in Apeldoorn

Today, I went to the Go tournament in Apeldoorn. It is already for the fourth time that go to this tournament. I arrived as one of the first around half past ten. Before the tournament, I already watched a game between a twelve year old boy and his father and quickly noticed that the boy played quite strong. He only started playing about one and half year ago, but he already reached 3 kyu. In the first round, I played against Marten (11 kyu) and lost with 35 points. In the second round, I played against Jan-Willem (9 kyu) and lost with about six points, which is maybe not too bad. Then in the third round I played against Martijn with a nine stone handicap because he is only 23 kyu. Playing with a big handicap requires a completely different playing style then what I am used to (where my opponent gets a handicap). I did not manage to beat him. I started to doubt that I had been too optimistic to declare myself a 10 kyu player at the start of the tournament. Last summer, it look like my 10 kyu rating was quite solid. It is also a fact that if I have to play faster (half an hour thinking time per person), I play worse. But in the fourth round, I had to play against Huub. For him it was his first tournament and he had enlisted himself as 13 kyu. The game was quite interesting, until he made a mistake and I could kill one of his groups. He made some desperate attempts to trick me into a mistake, but I failed to do so. In the fifth and last round, I played against a lady called Pita. It was an exciting game, where we both used almost all of our time. In the end I won with 77 against 69½ points. I had to leave immediately, because it was already a quarter to six and I had to catch Andy.

Wednesday, June 10, 2009

Playing Go

This evening, when I arrived at the university to play Go, I found Taco playing against Vincent and Araldo playing against Rudi. I asked Rudi if he had won last Sunday's tournament, and he had, as usual. He also scored five out of five. I first watched the game between Araldo and Rudi, and several times almost felt asleep. When that game was over, I wanted to leave, but there was a very interesting fight going on between Taco and Vincent, so I watched their game come to an end as well.

Thursday, June 11, 2009


I had formatted my ScanDisk Cruzer® Mirco 16GB as a TrueCrypt device and even made some short cuts on the desktop for (dis)mounting it. Everything worked fine, until last Tuesday evening, I got some message about a corrupted header and was adviced to restore the header from a backup version stored elsewhere on the disk. After I had done this, the drive seemed to work fine again. But yesterday morning, after I had mounted it (without problems) the mounted drive showed up as an unformated drive. I tried to (dis)mount it several times, and even restored the header again, but the mounted drive remained unformated. I have no idea what is the cause of this problems, but I decided to format the USB drive as a normal drive again. Luckily, I did not lose any important data because of this failure.

Arabic geometrical patterns

This evening at 20:35:15, I bought two books from De Slegte about Arabic geometrical patterns. These are: Yesterday, I had already seen these books in the shop in the section of recently acquired books. This evening, I did not see them, and I was afraid that they would already have been sold. I made some inquiries. At first they had not idea, but then one of the personel remembered something and found the books on some shelve.

Friday, June 12, 2009

Saint Peter's Square

This evening, I went to game club Fanaat who are celebrating their Sixth lustrum. There I met some people I knew from the past, including Martin Medema. I first watched him and five others play a game of Zigeuner (gypsy) Risk, a variant of Fanaat Risk. Both of these variants of Risk are played without dices and designed by Martin. After this, I joined in a game of Sint-Pietersplein (Saint Peter's Square) of which a picture is shown on the right. The game is played with a maximum of six players, who start with ten pieces at their side of the board and the object of the game is to move those ten pieces to the opposite side of the board. Players move by turn. You role two dices where the outcome decides with how many pieces you may move that many places, where it is up to you decide which dice is used for what. If, for example, you throw four and five, you might either move five pieces at most four places or four pieces at most five places. Whenever a piece is moved, it moves towards the opposite, except for at most one move in another direction. The board we played on, was made by Christian Freeling. It is really a fun game, because in the middle of the square you get a great congestion, like tourists crossing a square and all meeting in the middle. There where many more people playing all kinds of board plains. The twins, Arjan and Erik, where playing the new Battlestar Galatica game.

Saturday, June 13, 2009

Exact Covers of nonograms

In the past weeks I worked on improving my Exact Cover program. I started working on Exact Cover generated from nonograms. I started with a small 20 by 20 nonogram and did not encounter any problems. Next I tried the 80 by 95 nonogram from Kerrin Mansfield. But that surely proved too big because the program for transforming the nonogram in an Exact Cover did not terminate and produced huge files. I adapted the nonogram to Exact Cover program such that it take in account a partial solution. But with the partial solution I found on Tuesday, February 28, 2006 it still ran too long. Next I decided to start working on a 30 by 30 nonogram taken from the left top corner of this nonogram. (I wrote small program for generating the nonogram description.) When this was transformed to an Exact Cover it resulted in 116.743 vectors with a length of 960 (= 30*30 + 2*30). The total file size (including "names" for the generated vectors) was 117.696.606 bytes. I spend most of my time in developing an algorithm for reducing the number of vectors and positions. The last version of the program took several hours to reduce the Exact Cover to 4056 vectors with a length of 379. After this it took about eight minutes to find the 42 solutions, which are displayed here on the right. The image is produced with a program, which processes the output from the Exact Cover program. This makes use of include file, which I wrote today for creating 24-bits BMP files. In the image I did mark the solution matching the original image with a red line. It is interesting to know that the last version of my nonogram solver finds the 42 solutions within one minute. It seems that mapping Nonograms to Exact Covers is not a good idea if you want to find a solution efficiently.

Monday, June 15, 2009

Arabic geometrical patterns

This evening, I made an attempt to reproduce the first plate of the Arabic geometrial patterns by J. Bourgoin using PostScript. The result can be found in this PostScript file, from which an image is displayed on the right. It took me some time to learn some basic PostScript again using the freely available manuals. I was quite pleased with the reproduction as it looked on the screen. But when I turned the page and looked at the next page, I realized that I should have approached it in a different manner. I made use of a 21 by 18 points grid and made a rectangle pattern of 42 by 36 points that I rotated and copied to fill the page. The triangles on the page are not real equilateral triangles, but an approximation. Maybe I should reproduce the page starting from a triangular and/or hexagonal pattern, instead of a rectangle pattern.

Tuesday, June 16, 2009

Arabic geometrical patterns

This evening, I rewrote all the code for reproducing the first plate of the Arabic geometrial patterns by J. Bourgoin. After I had done this, the second page was not much trouble. I made some additional improvements to make the code more compact. The four lines that define the four patterns in the resulting PostScript file are the lines:
3 2 { 6 60 { 1 0 1.5 -1 utline 1 0 1.5 1 utline } nrotate } utalternate
3 2 { 6 60 { 0.5 -1 0 60 utarc 0.5 1 300 360 utarc } nrotate } utalternate
4.5 3 { 6 60 { 2 4 4 0 utline } nrotate } utalternate
3 6 { 6 60 { 0.5 -3 2.5 1 utline 2.5 1 3.5 1 utline } nrotate } utalternate

Sunday, June 21, 2009


This afternoon, while driving home from the Losserhof where we went to catch Andy, we saw several rainbows, including one bright, complete double rainbow when we turned from motorway N35.

Bourgoin, plate 13

This weekend, I worked on plate 13 from the Eléments de l'art arabe by Jules Bourgoin as reproduced in Arabic Geometrical Pattern & Design. This posed a little more trouble than the ones I had made so far, because it involved some constructed points. That means that I had to write some routines calculating the cutting point of two lines and for lines defined by a point and an angle. Friday night, I decided to write a limited PostScript interpreter for the subset of command that I used, purely for debugging purposes. This resulted in a C program based on IParse. The code quality of this program is not very high. Today, I finished reproducing plate 13. See this PostScript file for the result.

Saturday, June 27, 2009

Nonogram to Exact Cover

Yesterday and today, I thought about some other ways of mapping a Nonogram to an Exact Cover. The idea is to use two positions (columns) for each cell. For each possible row "1,0" is used for a filled cell and "0,1" for an empty cell. For the columns the exact opposite encoding is used. I was hoping that this might lead to an easier to solve Exact Cover. (I have to remark, that this encoding in some rare cases could lead to false solutions, because it does not include positions for enforcing that exactly one vector is selected for each row and column.) I extended my Nonogram to Exact Cover progam with an option "-01", to generate the alternative mapping. I tried this again on the 30 by 30 nonogram I used before. This resulted in 116.743 vectors with a length of 1800 (= 2*30*30). The total file size (including "names" for the generated vectors) was 215.760.726 bytes. The last version of my Exact Cover program took some hours to reduce the Exact Cover to an Exact Cover with 105 vectors with a length of 122 positions. After this it took five and half minute to find 42 solutions. I also have been thinking about another mapping, which is even longer, but likely far less vectors. The idea is to generate vectors for each possible placement of a block of filled cells in each row and column and for the white cells in between.

Sunday, June 28, 2009

Nonogram to Exact Cover

Today, I worked on the mapping that I mentioned yesterday. I extended my Nonogram to Exact Cover progam with an option "-logic", to generate this alternative mapping. This mapping generates a vector for each possible placing of a 'block' of filled cells in each row and column and the white cells in between. For the 30 by 30 nonogram this resulted in 3856 vectors with a length of 3422 positions. The last version of my Exact Cover program (with a small adaption that the define NR_POSITIONS has been changed into 4000) needed only half a minute to find the 42 solutions. The reduced Exact Cover contains 1517 vectors with a length of 1072 positions. Not a very dramatic reduction. But still it shows that the manner in which you map a certain problem to an Exact Cover can have a dramatic effect on the speed with which the Exact Cover can be solved. With an adapted version of the program to process the solution (which joins the partial solutions for each row and column), I generated the images shown on the right, in which I marked the original Nonogram, which happens to be the last solution found, with a red border.

This months interesting links


KMZ file of movements.

Home | May 2009 | July 2009