The 'Crossbos' (Dutch for 'cross forest') is a small park a few minutes walking
from our home. See this map for its location. On June 13, 1926, a concrete cycle track named
'The Station' was opened in the norther part of the park, but due to the
Great Depression
it did not last long. After it went broke in 1928, it was reopened on July 17,
1932, but this too did not last long and it was closed around 1935. It was
also used as an ice rink. (Dutch source.) Around 1960 there was not much left of the track, except
for some elevations in the shape of the track. The area got overgrown
by trees, but the centre area of the track remained open. Around 1990, the
norther part of the Helmerhoek quarter was build around the area, which was
left untouched. Some asphalt walking paths were constructed through it and
children started to bike around it, making a cross track over the elevated
parts, That is how it got its name. In the past years, the park has undergone
some improvements by the local government after some people from the
neighbourhood made some requests for this. For a long time, I have wanted to
make a map of the part. Today, the weather was grey and I took 556 pictures in
the park in the hope to see if I can use some Structure from Motion techniques to reconstruct a map.
In the past weeks, I made some good progress with respect to developing a C compiler as an replacement for the MES-compiler used in Live-bootstrap. The version of the Tiny C Compiler compiled with this compiler now runs on some simple
program, but does not generate any code, because it cannot find the file
crt1.o. The letters 'crt' stand for C Runtime. The file
tcc-0.9.26/pass1.kaem gives the commands that are used to build
the file using the GNU Mes
compiler and using the sources of the C standard library that are included
in it. I have to figure out how to convert this to using my C compiler.
This afternoon, I found the website Partijgedrag (Dutch for 'party behaviour') which is based on information
about how Dutch political parties voted on motions submitted by members of the
Dutch House of Representatives. It has a voting guide ('stemwijzer' in Dutch)
that gives you a rating which party matches best your vote on a number of
motions from the set of motions. It also gives a table with percentages of
how often parties voted the same on all the motions. It shows the parties in
alphabetical order. I wanted to know how the parties are aligned from right to
left. The table below is based on motions from the period July 2. 2024 to
today, which is during the Schoof cabinet. There were fourteen political parties at the start of
this period. I have tried to put the parties in a right to left order, which
is shown in the table. There are a bit difficult to place. The party
Forum voor
Democratie is a far-right political party but often with strange opinions,
sometimes rather fascistic, at other times based on various conspiracy
theories. The Partij voor de dieren is not the most left among the left wing parties,
but most in disagreement with right wing parties. The strongest agreement is
found among three left wing parties: Democraten 66, Volt, and GroenLinks–PvdA.
I also watched the videos in the exhibition Landscape, disrupted,
which I already briefly saw on Saturday, September
27. One of the videoas was not on display due to some talk being planned
with an artist. I left around the time that the talk started.
In the evening, I went to opening of the textile art expo FÆES-›!T by Overlap at Creative Broedplaats Enschede. There were also a meet-up around a table
of (young) people drawing. I understand that they meet every week at this spot.
The puzzle game Pinguins on Ice was laying around and when
some people were playing with, I joined in and suggested a systematic way to
solve the puzzle, but we failed to find a solution. Then when studying the
piece, we noticed that the piece could be morphed. Each piece consisted of two
parts that can slide along each other. Make use of this, we quickly found a
solution. At the exhibition, I saw the following works:
Slipjes, Zachtlandschap, Wild Yoni Flower,
Botanical Yoni, Maria, Vlinderbloem, Yoni
nestjes, and Yoni Flower by Willemien Nijman.
blablabla, Turn me on, Ode aan Paula Moderson-Beche,
and Frida Kahlo by Willemein Nijman.
Then, I went to B93 to see the exhibition
PHOTOAUTOMAT, a
collaboration between Patrick
Jonkman, who took pictures in Berlin, and Sabine Peters, who took nature
pictures in Italy. The made prints that combined fragments and silhouette from
the pictures that they took. I bought one print for
€ 5.00.
Yesterday evening, I played a bit with Pinguins on Ice puzzle game where you have to fit five pieaces of five
squares each into a five by five square. The pieces can morph in the sense that
the consist of two parts that can slide along each other. Each piece also has
a small pinguin on one of the five squares. In the booklet with the puzzle,
there are 60 puzzles, where you have find a solution where the pinquins are at
some locations. This got the idea that there are at least 60 solutions. Today,
I tried to figure out the number of solutions of the puzzle. I used the
program gen_dpfp with the seventeen
different configurations of the five pieces. I did some editiong to combine the
different configurations for each piece together and next used the program
dpfp2ec to transform the program into an
Exact Cover problem that I fed into my solver to
return only 44 solutions (not counting rotations). The solutions are shown
below:
Last Saturday, we got some curly kale in the harvest from Herenboeren Usseler Es. Today, we made some curly
kale hotchpotch with it using some potatoes we also got from the harvest. I
had some slices of gherkins and some small sausages with it. The sausages were
sold as 'Catalan' sausages, but although they had some unions, garlic and
peper as ingredients, they were not like the traditional Fuet.
Last Sunday, I came across the Flummoxagon puzzle game by Nervous System as it was mentioned and
discussed on Hacker
News. I wrote a comment stating that it would be nice if they would publish the
Exact Cover, such that I my solver to come up an
estimate of the number of solutions. I got an answer that they ran their
program for a couple of days and found tens of thousands of solutions. I
decided to write my own program, flummoxagon2ec.c, for generating an exact cover given an example
solution for the puzzle on the standard input. In the input the colours should
be specified with the letter 'g', 'b', 'o', and 'p', matching the hexagonal
shape of the puzzle, where there is one space between elements in the
horizontal line and where there is a shift of one character for consecutive
lines. For example:
o g g
o o p p
o p b
Because the output may contain duplicated vectors, one has to sort the output
and remove duplicates with the command 'sort -u'. I first let my
solver reduce the number of vectors by removing those that cannot occur in any
solution. Then I started the program yesterday. After it ran for about twenty
hours, it had only found three solutions and estimated that the number of
solutions was about 1049. I was a little disappointed about the low
number of solutions. The exact cover solver does not know about the nature of
the problem and thus it could either decided to try all pieces to cover a
certain location or try to place a piece at all possible places and
orientations. If the latter is decided, it is possible that it is placed in the
'middle' of the 'open space' such that there are no possible solutions. This
afternoon, I tried to make a modification such that it will only try to cover a
location with all possible pieces that are left. That did work a lot better,
finding about 20 solutions per hour. Still not the numbers that were found by
the program by the people of Nervous Systems. Maybe their program fills up the
locations in a fixed pattern, starting from one side to the other. Below the
first 45 solutions are shown:
With this run, the program estimates that there are 10148 solutions,
which is a lot more. It might be that the result is incorrect due to the
limitation that I introduced. It is also possible that the actual number is
somewhere between the two estimates. I have always had the idea that a generic
exact cover solver that always explores the position with the least number of
vectors first, would be the most efficient. But after today, I have to admit
that that is not the case. Fascinating.
Conny noticed a flower in our back garden. I am a
bit surprised by the number of flowers you can still see outside. I have no
idea what kind of plant it is, but the flower does have white and pink petals
and a yellow center. It could be a Dahlia 'Twyning's Smartie'.
At 14:55, we received a poster tube with the map
'Verdwenen Dorpen in Nederland' (Dutch for 'Disappeared Villages in the
Netherlands') as poster giving the locations of 563 villages that have
disappeared due to natural causes (primarily flooding), war, and/or economic
changes (such as disappearing peat production). The map can also be accessed
online in Dutch. Conny bought it as a gift
for my upcoming birthday from Cartomatie.
When I left home to go to the city, the sun was shining, but about half way,
when I biking on a priority road and seeing some dark clouds, I put on my
rain poncho while biking. Not a harmless act, but it went rather smooth. (I
believe that if you cannot biking on upright riding bike without hands, there
is either a problem with the bike that needs to be fixed or you are not good
enough at biking to bike safely.) At Fotogalerie Objektief, I saw the exhibitionStillness Between Moments with photographs and
mixed media by Nikolina Karoulla. The sun was shining again when I biked home after
having paid some more visits.
Fixing bug for the C compiler I am developing
as a replacement for the MES-compiler used in Live-bootstrap is taking more and more time as the bugs become harder to
find. All the low fruit is gone. Yesterday, I found a bug in how the switch
statement is translated. The implementation did not take into account the
possibility that the expression in the switch could had a side effect and
simply executed it multiple times. I changed the implementation, such that it
places the value on the stack. I also implemented some unit tests to check the
new implementation.
This year Advent of Code will only have 12
days and not 25 like before. I understand the reason for this. Another change
is that there will be no global leader board anymore. In the last years this
was filled more and more with people using AI tools to quickly find a solution
sometimes even using automated script to retrieve the puzzle text, generate
some code, run it and find a solution. I have created a private leader board for people who want to solve the puzzles in pure C
only using code they developed themselves. I have no way to check this. To
join this private leader board, the following code can be used:
1563228-d419ba6d.
I finished reading the book DRIVE: Train je
stoïcijnse mindset (Dutch for DRIVE: Train your stoic mindset)
by Mark Tuitert,
which I started reading around September 19. I guess that the contents of the
book is similar to the book The Stoic Mindset that he advertises on his
page The Stoic Mindset. I bought the book on Friday,
July 9, 2021 for €20.00. I am not sure if it was worth that price as I
did not get much out of the book. I did not take time to do the exercises. I
did like that the chapters did follow his personal life and that it is a kind
of autobiography.
I visited the Rijksmuseum Twenthe to vote in the
voting station that was in the goblin hall. After having voted, you were
allowed to visit the museum for free. At the front desk, they just let everyone
not even paying notice. I walked around the museum and found the following
works noteworthy all of which I have already seen before, some of them only
casually:
witches' monuments 2004-2013, herman de vries, 2013. Five frames
with the following poisonous plants that were believed to allow witches to
fly:
Ergot,
valerian,
henbare,
belladona,
and monkshood.
belladonna (digitized 16mm film), herman de vries, 1983.
Yesterday evening until half past one this night, I again helped counting
ballots for the 2025 Dutch general election. According to the exit polls, the social
liberal and progressive party Democrats 66 (D66) got 27 seats from the 150 in the Dutch House of Representatives and thus becoming the largest party, while the
previous largest party, Party for Freedom (PVV), only got 25 seats. However, when the numbers came
in, the difference between D66 and the PVV became smaller and at some points the
PVV was even ahead of D66. Now at 99.7% of the votes being counted, D66 is
about 15,155 votes ahead of the PVV. In the municipality of Venray the verification of the counting has been
delayed due to a fire. And there are also still mail votes that can be
included. The official result will be announced on November 7 by the
Electoral
Council. It is expected that D66 will remain the largest party. It is even
possible that they will get an extra seat due to the algorithm to distribute
the residual seats. Now the residual seat goes to the Socialist Party (SP) but
if the D66 gets more votes relative to the SP that seat will go to D66. If this
happens, Geert
Wilders, the leader of the PVV, will not be very happy, because he stated
today that there was no party with more seats than his party. Some people have
said that the loss of the PVV was a blow to the far right parties in the
Netherlands, but that is actually not true, because the two other far right
parties, the Forum for Democracy (FvD) and JA21 both got more votes and their
total number of seats increased with one. After this election, the parties need
to form a coalition to become the government. Any coalition needs to include at
least four parties, but some parties have already expressed that they exclude
some other parties or are not in favour to work together. Parties also fear
that if they form a coalition with some other parties that their voters will
turn to another party in an upcoming election.
Some table with numbers of seats. The first four rows are about forcasts. The
fifth column is a combined forcast based on the first three forcasts. The sixth
column is the exit poll. The seventh column is number of seats based on the
count of today where 99.7% of the ballots have been counted. It shows that the
forcasts are not very reliable even if they specify a 95% range of lower and
upper number of seats.
