Previous Up No next

Diary, July 2026



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


Wednesday, July 1, 2026

More on convex shapes of hexagons

I continue looking into convex shapes of hexagons and looking into the conjecture that the sequence deviates at most one from the sequence A216522. With respect to this sequence and the sequence A135711 the following are true:

A216522(n) = ceiling(sqrt(12n + 9)) - 3
A135711(n) = 2.ceiling(sqrt(12n - 3))
A135711(n) = 2.A216522(n-1) + 6
A216522(n) = A135711(n+1)/2 - 3
(It is important to note that the sequence A135711 is one based where the sequence A216522 is zero based.) If we take the sequence A216522 as a reference then the perimeter should be defined as the distance between the hexagons on the outside where the distance between the center of two hexagon is defined as one. The table below gives for a given perimeter (for which k is positive) the number of hexagons is can enclose. So for k is one, seven hexagons can be enclosed with a perimeter of six.

perimeter:  sides:                         #triangles:     #hexagons:
6k          k,   k,   k,   k,   k,   k     6k²             3k² + 3k + 1
6k + 1      k+1, k-1, k+1, k,   k,   k     6k² +  2k - 1   3k² + 4k + 1
6k + 2      k+1, k,   k,   k+1, k,   k     6k² +  4k       3k² + 5k + 2
6k + 3      k+1, k  , k+1, k,   k+1, k     6k² +  6k + 1   3k² + 6k + 3
6k + 4      k+1, k+1, k,   k+1, k+1, k     6k² +  8k + 2   3k² + 7k + 4
6k + 5      k,   k+2, k,   k+1, k+1, k+1   6k² + 10k + 3   3k² + 8k + 5
If we use a zero based variant of sequence A216522, the formulea becomes: ceiling(sqrt(12n - 3)) - 3 for the number of perimeters. Lets denote this number with p, then the equations with this formulea are:
p + 3 - 1 < sqrt(12n - 3) ≤ p + 3
(p + 2)² < 12n - 3 ≤ (p + 3)²
p² + 4p + 4 < 12n - 3 ≤ p² + 6p + 9
p² + 4p + 7 < 12n ≤ p² + 6p + 12
Now if we fill in the values for the perimers and matching number of hexagons, we get:
6k:      36k² + 24k +  7 < 36k² + 36k + 12 ≤ 36k² + 36k + 12
6k + 1:  36k² + 36k + 12 < 36k² + 48k + 12 ≤ 36k² + 48k + 19
6k + 2:  36k² + 48k + 19 < 36k² + 60k + 24 ≤ 36k² + 60k + 28
6k + 3:  36k² + 60k + 28 < 36k² + 72k + 36 ≤ 36k² + 72k + 39
6k + 4:  36k² + 72k + 39 < 36k² + 84k + 48 ≤ 36k² + 84k + 52
6k + 5:  36k² + 84k + 52 < 36k² + 96k + 60 ≤ 36k² + 96k + 67
It is quite obvious that for all positive k the above equations are true. This means that for all those minimal perimeter for the given number of hexagons is equal to the given formulea. So, there is not a value, above which the perimeter might be always one larger than given by the formulea. This seems to support the conjecture that it is at most one higher than the value given by the formulea. It should be noted that the number of hexagons for which there is a simple solution for the minimal perimeter appear further and further apart leaving more and more room for complicated cases.

End of heat wave

Today, is the last day of a two week long regional heat wave. In the past two weeks the maximum temperature was at least 25° Celsius. There were eight days on which the temperature was 30° or higher of which one day higher than 35° with a new 37.9° record for the day. There was a total of four days with temperature records for the day of the year. There was also one night that the minimum temperature was 21.6°, which is also rather exceptional. The national heat wave was only eleven days, but nevertheless a record number of days. In the South of the country there was even a 'super heat wave' with four days with temperatures of 35° or higher.


Thursday, July 2, 2025

Graduation Show at Rietveld Academie

I traveled to Amsterdam to attend the Graduation Show 2026 at Gerrit Rietveld Academie. I found the works of the following students noteworthy, which is very subjective and often based on the first impression, in the order I encountered them:

Museums

Afterwards, I went to Stedelijk Museum Amsterdam. There I first saw the exhibtion Kho Liang Ie, which contained the following works from his collection: From some of the other exhibtions: I also saw the exhibition Danh Vo - πνεῦμα (Ἔλισσα).

At FOAM, I saw the exhibitions Foam Talent 2026 and Martin Parr - Very Modern and Rather Ugly.


Friday, July 3, 2025

Graduation Show at KABK

Today, I went to the Graduation Show 2026 at KABK. I found the works of the following students noteworthy, which is very subjective and often based on my first impression, in the order I encountered them:

Wildsam Field Guides: Los Angeles

In the train home, I finished reading the booklet Wildsam Field Guides: Los Angeles, which I started reading on March 22 after I bought it on February 18. Initial I read some pages during commercial breaks. In the past days, I finished reading it in the train. Although it has many pages with fun and mundane facts, it also has some serious content giving you a good impression of the city.


Tuesday, July 7, 2026

Flowers again

Just like last year and some years before, our magnolia has some flowers again. The three plants that I planted about a month ago are still doing well. They still rather small, but I understand that that is normal as they focus on growing roots first. I regularly give them water with the waste water from rinsing vegatables.


Saturday, July 11, 2026

Cyber Ægg

In the afternoon, I went to TkkrLab to help with the 'sweatshop' for assembling the Cyber Ægg badge for BornHack. I first unpacked some PCBs, next I helped with attaching the E-Ink display, because that was the bottleneck in the whole assembly line. I did the same in 2017 for the SHA2017 badge. I felt it was a little bit more difficutl this time. I did a bit over thirty before I left.

Exhibitions

I went into the city and at photo gallery Objektief, I saw the exhibition Meesterwerken 2026 with students finishing some the photography training. I saw the presentations by:

At Concordia, I saw the exhibition AKI Finals '26 with works by the following graduates:

I also saw the exhibition Expositie Boswinkel with amateur art from the neighborhood with that same name. I was rather surprised by the quality as the neighborhood is not known as one of the best neighborhoods in Enschede.


Sunday, July 12, 2026

Some interesting relationships

I found some interesting relationships with respect to convex shapes of hexagons. I realized that the each hexagon on a triangluar grid can be viewed as the intersection of two triangles. I also realized that there is a simple relationship with respect to the sizes of the hexagon. Let the sizes of the hexagon, be denoted by a, b, c, d, e, and f, then the following equations are true:
    a + b = d + e
    b + c = e + f
    c + d = f + a
Note that the third equation can be derived from the first two. If n is the sum of all sides, the following equation can be derived:
	d = n - a - 2b - 2c
	e = -n + 2a + 3b + 2c
	f = n - 2a - 2b - c
Now it is also possible to calculate the number of enclosed triangles from the values of a, b, c, and n with the formulea:
     n(4(a + c) + 6b) - n² - 4(a² + c²) - 8b² - 10b(a + c) - 6ac
The formulea shows that a and c can be interchanged, which is what one would expect. This formulea can also be used to find generic formuleas on k of the form 6k² + nk + l, where l is equal to the above formulea for enclosed triangles. I wonder if there is an efficient method to discover for giving values for n and l there is a solution for some values of a, b, and c or not. We could for example, turn the formulea in a quadratic equation of b, which results in (if I am not mistaken):
     -8b² + (6n - 10(a + c))b - n² + 4n(a + c) - 4(a² + c²) - 6ac - l = 0
The determinant of the quadratic equation on b is (if I am not mistaken):
    -28a² - 28c² + (128n - 120)a + (128n - 120)c + 8ac + 4n² - 32l
For there to be a integer solution for b, the determinant must be a square number.

Remark July 13: I was mistaken with respect to last formulea.


Monday, July 13, 2026

I was mistaken

The formulea that I presented yesterday for the determinant (related to convex shapes of hexagons) is incorrect. The correct formulea, written slightly different, is:
    -28(a² + c²) + 8ac + 8n(a + c) + 4n² - 32l
Taking out a factor of four, we get:
    -7(a² + c²) + 2ac + 2n(a + c) + n² - 8l
So, how do we find integer values for a and c for given n and l such that the above formulea returns a square? Even if l is a negative number, there are only a limited number of pairs to be evaluated, due to the negative factor with the quadratic terms. The values for n are always positive. It looks like all positive values are within a certain circle or elipse. The function is symmetric around the line a = b, The highest value probably lays on that line. If we replace a and b we get the formulea:
    -12x² + 4nx + n² - 8l
From the first derivative we know that the highest point is at x = n/6. This might be used to determine the center of the circle. I wrote some code to investigate values for which the formulea returns squares for various values of n and l and discovered that only for some multiples of 3 for n and multiples of -6 for l there are no squares. The others have many squares and it seems that the number increases with smaller values for l.


Tuesday, July 14, 2026

Age verification

I understand, based on the page The EU approach to age verification, that the EU is thinking about the introduction of mandatory age verification through a privacy-preserving application (based on zero-knowledge proofs) for accessing websites that have an age restricted content. Paul Walsh gives an interesting reply in a tweet. I agree with him that it is the primary responsibility of parents to determine what their children may watch and that there are enough tools in place to implement those restrictions. I think that what type of content you find appropriate for your child at a personal convictions. I personally have more problems with violence than with sexual explicit content, while I get the idea that in some cultures and/or people with a religious background it is often the reverse. I wonder if the implementation will also work for linux. I am also thinking what will be consequences for private website, like this one. Will an age indication be required for all website? If not, how will it be determined if a website is required to have it?


Thursday, July 16, 2026

Asking Google AI

I decide to ask Google AI about the expression below, which I had found with respect to convex shapes of hexagons. I asked Google AI why for some n and l the values of the expression does not contain perfect squares for any value of a and c.
    -7(a² + c²) + 2ac + 2n(a + c) + n² - 8l
It did not give the answer to the question, which I had not expected, but it did return the boundary for a and c in which all positive values can be found:
    (n - √(7n² - 42l))/6 < a,c < (n + √(7n² - 42l))/6
It found this, by first rewriting the quadratic equation of c, like:
    -7c^2 + 2(a + n)c - 7a^2 + 2na + n^2 - 8l
and then derive the determinant for it having real solutions, resulting in:
    -192a^2 + 64na + 32n^2 - 224l
and then again derive the determinat of the quadratic equation of a, like:
    28672n^2 - 1702032l
The square root of this determinant can be simplified to:
    64√(7n² - 42l)
It also stated that 6l < n², which in our case is easily satisfied knowing that l is at most 4. I also developed the program convexhexagon3.cpp, which can find a solution for any number of hexagons, not having to 'construct' solutions for all previous numbers. It finds the solutions for the first million values. For 40550 of those values, the convex perimeter is one higher than the minimal value of there is no convex requirement. In for non of the values it was more than one higher. The fact that it is only a little over 4% of all values is one higher and skimming the output gives me the impression that my hypothesis that it is at most one higher is true, although this does not count as a mathematical proof of it. I think I will leave at this for the moment.


Friday, July 17, 2026

Zwolle

I went to Zwolle because I had an appointment at Kamer van Koophandel, which is the Chamber of Commerce in the Netherlands to finish the registration of my sole proprietorship named FFAA5E. I decided to early and visit some places first. I went to the Museum de Fundatie. I first watched the exhibition Ntu with works by Buhlebezwe Siwani. I saw the following works:

Next I saw the exhibition Umbild with installations by Jules van Hulst & Wieger Steenhuis. I saw the following works, all created in 2025-2026:

On the second floor, I saw the exhibition Cremer in context: The early years about the artist and author Jan Cremer. I found the following works noteworthy:

Photographs:

From the collection of the museum, I found the following works noteworthy:

I walked through the city and visited the following bookshops: Boekhandel Bloks (only secondhand books), A Lot of Books (new and secondhand), Boekhandel Westerhof (new) and Van der Velde Boeken (new and some sale in former church). I also visited the art shop and gallery Blauwdruck. At Academiehuis Grote Kerk I saw the exhibition PROTO MASTERS Art & Design. I found the following works noteworthy:


Monday, July 20, 2026

Interesting answers

Yesterday, I posted the question Hints for proving properties about an expression on the askmath subreddit about the expression I found with respect to convex shapes of hexagons and I got some interesting answers. From one comment it can be shown that the expression
     -7(a² + c²) + 2ac + 2n(a + c) + n² - 8l
can be rewritten as the following (equivalent) expressions:
     -3(a + c)² - 4(a - c)² + 2n(a + c) + n² - 8l
     -3(a + c - n/3)² - 4(a - c)² + 4n²/3 - 8l
From another comment it can be shown that the expression is equivalent with:
     (a + c + n)² - 8(l + a² + c²)
It also explains that if and only if -l does not contain any prime factor ≡ 3 (mod 4) with odd multiplicity, then there are integers a and c that make the second summand zero and thus the whole thing a perfect square. This definitely looks interesting, although I have not worked out all the details. The sequence of number that are the sum of two squares is A001481. This definitely interesting, although I still do not understand how it works out in detail.


Tuesday, July 21, 2026

Cold night

Last night, The temperature at Twente Airport dropped to 5.0° Celsius at 1.5 meter height, which does not beat the record of 4.2° on this day in 2012. At ground level, at a height of 10 cm, the temperature dropped just below zero to -0.1° which really rare for this time and the lowest ever since measurements started in 1971. The previous lowest temperature was 1.9° (also) in 2012.


This months interesting links


Home | June 2026