Previous Up No next

Diary, August 2019

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  31

Thursday, August 1, 2019

Received two books

I recieved the two books from Antiquariaat Kas Cornelis, for € 6.00 each:

Sunday, August 3, 2019


I went to Amsterdam, to see the graduates exhibition at gallery Ron Mandos, but they were closed, due to canal pride. I visited the following bookshops in Amsterdam and Utrecht:

Sunday, August 11, 2019

KABK wrapping paper

In the past week, I spend some time to analyze the wrapping paper of the Royal Academy of Art (Koninklijke Academie van Beeldende Kunsten in Dutch, KABK for short), which I took with me when I visited the graduation exhibition on Sunday, July 7, 2019. The wrapping paper is based on logo of the KABK, which consist of seven dots, four dots in a square, one dot in the middle of the top two dots, and two dots at the cross points of some lines between the other five dots. The seven dots can be connected with seven lines to create the shape resembling a crown. On the wrapping paper, there are 27 lines 18 columns of these logos, with seemingly randomly added lines between the dots. On the seventeenth row in the eleventh column, the logo resembling a crown can be found. I took a picture of the wrapping paper and analyzed this with a MySample script to extract the lines used in each of the logos. I next went on to analyze it with a program to see if there is any pattern in the lines added to the logos. I discovered that the first ten columns all logos have six lines added, and that in the remaining eight columns, all logos have seven lines added. There were no two logos with the same lines added. Futhermore, I could not find any patterns. It seems that the logos are purely selected and also placed at random. The design somehow reminds me of the PARR patterns, who were designed by Taconis Stolk, who at the moment is the head of the ArtScience department.

Tuesday, August 13, 2019

Third qualifying round (continued)

Gerard, the one who raised the question about the third qualifying round, figured out the mathematical formuleas behind the problem. (In the following, I will use the n over k notation for binomial coefficient n!/(k!(n-k)!).) First he derived the expressions for the numbers produced by the program I wrote:
11520 = (10 over 0)(10 over 2)2^8
53760 = (10 over 1)( 9 over 3)2^6
50400 = (10 over 3)( 8 over 4)2^4
10080 = (10 over 4)( 7 over 5)2^2
  210 = (10 over 5)( 6 over 6)2^0
Next he generalized this to the expression:
   (w over s)(w-s over s+w-t)2^(t-2s)
Where s is the number of games in which two 'League Path' teams play against each other, t is the number of 'League Path' teams, and w is the total number of games played, which equals half the sum of 'Champions Path' and 'League Path' teams.

Wednesday, August 14, 2019

KABK wrapping paper alternative

I wonder it would be possible to arrange the logos in the KABK wrapping paper such that logos that have similar lines added are not placed close together. I wrote a program, which basically exchanges two at random selected logos as long as it does not make things worse, and continues with this until there is nothing to improve anymore. It makes use of a badness measurement that defines when the line pattern of two logos are to similar when logos are a certain distance of each other. Also mirrored logos in close proximation are considered bad. The solution the program came up with can be found in this PDF. I want to investigate if there are alternative selections of logos that by themselves are less 'close'.

Thursday, August 15, 2019

Constant weight codes

While thinking about the KABK wrapping paper, I arrived at constant weight codes through error correction codes. The Hamming distances of two line patterns of a logo is equal to the places where there is or is no line. Thus moving one line, equals a Hamming distance of two. To avoid these the distance should be 4 and for logos with six and seven lines, we should look for the value A(11,4,5) and A(11,4,4) respectively. According to a table, these values are 66 and 35. There is a superscript s with the 66, which stands for: "shortened code (from code of length n+1 and weight w or w+1)." I downloaded the data for A(12,4,5) and wrote a small program to convert it into binary vectors, but I have no clue how to shorten the vectors into something useful. It seems I have to write a program myself to generate the vectors.

This months interesting links

Home | July 2019