Every four years, the world comes together to watch one of the most anticipated sporting events in history: the FIFA World Cup.

Behind all the anticipation, venue planning, and media fanfare, there are many artists and researchers who devote themselves to designing a new FIFA World Cup ball to be rolled out for the public eye (pun intended).

This post presents an overview of the geometric ideas behind the design of the FIFA 2026 "Trionda" ball, using Maple to visualize and explore these concepts in depth. The ideas presented here were inspired by this Scientific American Article. For more information and facts about the 2026 Trionda ball, as well how the shape of the ball impacts play on the pitch, I suggest you check it out!

FIFA ball designs are often inspired by one of the 5 Platonic solids. A Platonic solid is a convex polyhedron with each face being the same regular polygon with the same number of faces meeting at each corner.

This year, the Trionda ball was constructed from the simplest of these shapes, the tetrahedron, consisting of 4 triangles, with 3 faces meeting at each corner. Of the five Platonic solids, this shape has the fewest faces, making it the least sphere-like. Turning such a simple polyhedron into a smooth ball is therefore a surprisingly challenging geometric problem.

So how can we turn our pointy tetrahedron into something that rolls? Rather than trying to transform the entire tetrahedron at once, we can start by redesigning a single triangular face. The goal is to create a curved triangle that will fit perfectly with three identical copies of itself while covering the surface of a sphere.

 

Hi Maple community and others,

I'm very proud to present my code.

Sequences are fun,
for those who know, about them

consider Fermat numbers, of the form,
F(n) = (2^(2^n)) + 1.
goes like

3, 5, 17, 257, 65537, 4294967297, 18446744073709551617, 
340282366920938463463374607431768211457, ...

in oeis.org database at
https://oeis.org/A000215 .

Similarly we can have base 3,

B(a) = (3^(3^a)) + 1.
goes like, this,
4,28,19684, ...
online, in database, with Universal Resource Location (URL)
https://oeis.org/A129290

There could also be base 4, that grows even faster
 

double_exponential_2_and_3_and_4.mw

That is all that I have, for now.

Thank you for this free forum.
regards,
Matt

Hi Maple community, and all,

Here is a little Maple worksheet, shoing an interesting property of prime numbers.

Numerical evidence supports Andreca's conjecture.

see    

_Andricas_conjecture.mw

good fun

see, also
Andrica's Conjecture -- from Wolfram MathWorld
Enjoy
regards,
Matt

PS online at https://MattAnderson.fun/

PPS Have a good day, everybody.

Main worksheet [not all is displayed below]:

Download RootAnalysis4.mw

Download RootAnalysis5.mw

A note on what I've been working on for the past while. Some of you may have seen the announcement on LinkedIn yesterday; this is for the home audience.

The question I've been chasing is the one that's underneath the Physics package, the dsolve / pdsolve formal methods and heuristics, the advanced Mathematical Functions and FunctionAdvisor, and most of what I've written for Maple over the years. How can mathematicians and physicists speed up significantly their work using Computer Algebra Systems (CAS) and at the same time trust the result a computer hands back? The new chapter is what happens when AI sits between the human and the CAS, and the answer to that, in my view, turns out to be a much harder problem than the AI hype suggests.

Why? Because AI is increasingly the driver of computational mathematics in research, engineering, and education. And the unsolved problem isn't whether AI can do mathematics. It can. The problem is that an incorrect AI result arrives with the same confidence as a correct one.

On 100 challenging problems of undergraduate mathematics we tested, six independent state-of-the-art AIs returned mathematically equivalent answers on only 21% of them, and even within a single AI, repeated runs disagreed with themselves on 3% to 57% of the problems (details). The gap this validation crosses, between probabilistic inference and certified mathematical computation, is epistemological, not technological. It won't close with more training data. It needs validation across multiple AIs and multiple CAS, with no single engine having the final word.

ExaktAI aims to address that gap. It guides AI through mathematical computation, validates each step against Maple and Mathematica, and automatically generates and opens a corresponding CAS document where the validation can be audited and reproduced for every step, and where one can continue working on the problem. The goal: AI-mathematics that is validated, with the human in the loop.

ExaktAI is now well developed (TRL 6: System prototype demonstration in a simulated environment, on the ISED / Innovative Solutions Canada TRL scale). At the end an image. A Beta is scheduled for late summer / fall 2026; details at exaktai.ai.

In summary: ExaktAI is my present, and if you work on AI for mathematics and computer algebra, or the validation problem for AI, I'd love to hear your perspective.



Edgardo S. Cheb-Terrab
ExaktAI
Research Fellow Emeritus at Maplesoft.

Hi MaplePrimes, and all,

Here is a new, to me, set of numbers
defined by ,
the first three numbers are {1,2,3}
and then, 
the next number is the sum of the three 
previous numbers,
so,
{1,2,3,6,11,20, ... }
but can only calculate a finite number of numbers
the, so called, Tribonacci numbers
could start with {0,1,0}
see online
https://oeis.org/A001590

and
triple_recursive_sequence_simple_first.mw

triple_recursive_sequence_simple_first.pdf

regards,
Matt
 

Hi again Maple community, and others,

want to share
tiwn_and_cousin_prime_numbers.mw
tiwn_and_cousin_prime_numbers.pdf
(spelling error in file name)

~

just want to share,
some successful code

The lesser of the twin primes are listed
{3,5,11,17,29,41,59,71}
https://oeis.org/A001359
Prime numbers p such that p+2 is also a prime number

also, the lesser of the cousin primes are listed
{3,7,13,19,37,43,67,79,97}
https://oeis.org/A023200
prime numbers p such that p+4 is also a prime number

good fun

also, my webpage has more details
https://mattanderson.fun
okay

regards,
Matt

I am very pleased to announce the publication of my new book written together with Nic Fillion, "Perturbation Methods Using Backward Error", which uses Maple heavily throughout.

You can find it at

the SIAM bookstore

and I hope that you find it useful and interesting.

You can also find a paper in Maple Transactions, written with Michelle Hatzel, that explains how we generated the image that was chosen for the cover.  Exploring Cover Designs for an Upcoming Book  In the end, the SIAM design people chose a different one than we had thought, but they did pick one of the ones we generated!  

This was fun to do.

Hi mapleprimes, and all,

did a little exploration with the Matrix() and ifactor() Maple commands.

prime_factorization_of_one_digit_numbers.mw

have a look
no warnings, and no errors

Regards,
Matt

Hi Maple community, and all,

My intrest in prime numbers continues.

Made a quick example file.

3_tuple_admissible_example.mw

3_tuple_admissible_example.pdf

also, see my webpage for similar content
https://mattanderson.fun
and Norman, in Germany
prime k-tuplets & Primzahlen

Enjoy

Matt

I was wondering whether a water molecule exhibits the intermediate axis theorem (IAT) effect.

I was also curious if a tool like MapleSim could be used for simulations, as it is designed with technical applications in mind. (It would have been handy to enter parameters in Angstroms and atomic mass, but MapleSim only provides length in mm and mass in grams for small objects.)  

The atoms of the H2O molecule form an obtuse triangle. An obtuse triangle exhibits the effect when rotated about its intermediate axis (horizontal axis in the image below) provided it consists of equal point masses (more details here) which is not the case for oxygen and hydrogen. 

Building and running a model is quick since geometry and masses are available online. Two molecules are modeled in the attached file with parameters from different sources. The left model in the image below represents the atoms as three discrete point masses without individual rotational inertia. The system's total inertia is derived entirely from the spatial distribution of these masses. On the right, the entire molecule is modeled as a single lumped mass located at the center of mass (CoM), with the molecule’s full rotational inertia tensor applied at that point.

IAT_H2O.msim

Both models are rotating about the assumed intermediate axis at 3 THz (3 trillion rotations per second), which is about two orders of magnitude less than near-infrared light (just beyond the visible spectrum). To generate the IAT effect an orthogonal tiny initial “kick” of 0.001 Hz was added.

Visualizing was not so easy since the models are in the picometer range. It was difficult to locate the molecules by zooming with the mouse wheel (the button “Fit scene” on the 3-D Playback window did not work). Running an animation to visualize the IAT effect was not possible because the display speed could “only” be slowed down by a factor of 2^31 (see A in the image below). However, the slider C to move back and forth in time worked. The time (B) was always rounded to zero due to the very short simulation time span of 1E-10 seconds (i.e. 10 picoseconds).

Export of an animation movie worked better. Over the displayed timespan of 10 ps the molecules flip 2 times back and forth. This corresponds to a flip frequency of 200 GHz.

Time is still not displayed in the movie but that is pretty much all what did not work. Overall, a good performance. Also the numerical results did not show signs of loss of fidelity despite the atomic scale  of the objects and the very short time span.

In reality, the rotation of a dipole results in the emission of electromagnetic radiation, which dissipates energy and decelerates the molecule. How can this damping effect be modeled in MapleSim? Someone has an idea?

For decades, Maple has been built around one of the world’s most powerful mathematics engines—helping students, educators, engineers, and researchers explore ideas, solve complex problems, and communicate mathematics clearly.

Maple 2026 builds on that foundation with major advances in the math engine, expanding the kinds of problems Maple can solve while improving reliability and performance.

At the same time, Maple 2026 introduces new AI-powered tools that help you work faster—finding commands, generating visualizations, explaining concepts, and helping you explore ideas. The key difference is that these tools sit on top of Maple’s math engine, so the results are grounded in real computation rather than guesswork.

If you’ve been following along with our recent Mathy teaser videos and sneak peek posts, you may already have seen hints of some of these features. Now I’m excited to finally share them in full.

One of the most exciting additions in Maple 2026 is the new AI Assistant.

AI tools are incredibly useful for exploring ideas, writing code, and learning new topics. But when the mathematics becomes more involved, relying on AI alone can be risky. The Maple AI Assistant brings those productivity benefits into Maple while keeping the mathematics grounded in Maple’s trusted computation engine.

You can ask the AI Assistant questions in natural language and have it help you:

• find Maple commands or formulas
• generate Maple code
• create visualizations
• explain mathematical concepts
• draft examples, worksheets, or reports

Because Maple performs the underlying computations where appropriate, the results are grounded in Maple’s powerful math engine. The AI Assistant becomes a productivity partner that helps you accomplish tasks in Maple faster and more easily, combining the flexibility of AI with mathematics you can trust.

Watch the AI Assistant in action.

Another feature I’m particularly excited about is Document Import.

Many of us have years of mathematical content stored in PDFs, lecture notes, journal articles, slides, or even handwritten pages. Traditionally these documents are static—you can read them, but you can’t interact with the mathematics inside them.

With Maple 2026, that changes.

Document Import allows Maple to convert many document formats—including PDFs, DOCX files, and presentations—into Maple worksheets where the mathematics becomes live and executable. 

The image below illustrates the transformation.

On the left (“Before”), scribbled handwritten notes from a Calculus III lecture were saved in a Word document. The notes include hand-drawn sketches, formulas, and written explanations.

After importing the document into Maple (“After”), the mathematical expressions were recognized and converted into live, editable Maple mathematics. The text was preserved, and the hand-drawn sketches were retained as images. The resulting worksheet supports evaluation, editing, and further computation.

Once imported, you can:

• evaluate expressions
• modify formulas
• extend derivations
• add visualizations
• explore variations of the mathematics

Instead of recreating examples from scratch, you can bring existing material directly into Maple and start exploring.

While the new AI features are exciting, the heart of Maple has always been its mathematics engine—and Maple 2026 delivers significant advances here.

One particularly notable improvement is Maple’s expanded ability to solve linear recurrence equations. Through improvements to the rsolve command and major extensions to the LREtools package, Maple can now solve dramatically more recurrence relations than before, including many third- and fourth-order cases that were previously beyond reach.

In fact, Maple can now fully solve over 94% of the 55,979 entries in the Online Encyclopedia of Integer Sequences (OEIS) that that can be shown to satisfy a linear recurrence relation. These advances reflect ongoing research into linear difference equations and their algorithmic implementation in Maple, continuing Maple’s long tradition of advancing the state of computer algebra.

Beyond recurrence solving, Maple 2026 includes many improvements across its core symbolic and numeric algorithms. Maple’s assumption system has been strengthened to improve reasoning under mathematical assumptions, and enhancements to the simplify, combine, and evalc commands allow Maple to produce more compact and mathematically natural forms for a wider range of expressions.

There are also improvements to Maple’s differential equation solvers, polynomial system solving, and numerical solving routines such as fsolve, along with updates to other foundational parts of the math library used throughout the system.

Taken together, these improvements expand the range of problems Maple can solve and improve the robustness, correctness, and efficiency of the results.

Maple has always offered extensive control over plotting options, but achieving consistent visual styling across multiple plots could require specifying many settings each time.

Maple 2026 introduces Plotting Themes, which allow you to define a plotting style once and apply it across many plots with a single option.

Themes make it easy to maintain consistent visual styles in worksheets, teaching materials, reports, and publications, while still allowing individual plots to override specific options when needed.

The image below shows an example of creating and applying a custom plotting theme. 

Maple continues to be widely used in classrooms around the world, and Maple 2026 includes several improvements designed to support teaching and learning.

The Check My Work system has been enhanced so Maple can recognize a wider variety of valid student solution steps and provide more accurate feedback.

Maple 2026 also improves the generation of similar practice problems, making it easier to create variations of a problem while preserving its mathematical structure.

In addition, Maple’s step-by-step solutions have been expanded to support more types of expressions, helping students better understand the reasoning behind the mathematics they’re learning.

Maple 2026 also introduces improvements for developers building advanced applications, along with performance enhancements across the system.

One particularly interesting addition is the new VectorSearch package, which implements a vector database directly inside Maple.

If you’re not familiar with vector databases, one way to think about them is through recommendation systems like Netflix or Spotify. Each movie or song can be represented by a vector containing thousands of numbers describing its characteristics—things like genre, pacing, or mood. When you watch something, the system finds other items whose vectors are closest to it, which is how recommendations are generated.

With the new VectorSearch package, Maple can store thousands (or more) of vectors and efficiently find the ones most similar to a given vector. This makes it easier to build applications involving machine learning, data analysis, and modern AI workflows directly in Maple.

Maple 2026 also delivers significant performance improvements. For example, operations involving quantities with units have been greatly optimized—some computations now run over 90 times faster, making Maple even more efficient for engineering and scientific workflows.

Maple 2026 also expands the benefits available through the Maplesoft Elite Maintenance Program (EMP). The new benefits include access to additional Maplesoft products and services:

• Maple Learn, the online environment for teaching and learning mathematics
• Maple Calculator Premium, bringing the power of Maple to your phone with full access to features like Solution Steps and Check My Work
• Maple MCP, which allows you to connect Maple’s math engine to external AI tools so they can produce mathematical results you can trust

These additions extend Maple beyond the desktop, giving users powerful tools for learning, teaching, and exploring mathematics across web and mobile platforms, as well as through integrations with external AI tools.

This post only scratches the surface of what’s new in Maple 2026. There are many more improvements across the math library, programming tools, and performance.

To learn more about all the new features and enhancements in Maple 2026, visit the What’s New in Maple page on our website.

I am very pleased to announce that Volume 6 Number 1 of Maple Transactions has been published.  This is a Special Issue on Matrices and Polynomials in Computer Algebra, and the Guest Editors (our first ever!) were Marc Moreno Maza and Tomas Recio.  There are still some papers that are expected to be added to the issue when they come in, but at this moment there are 8 papers there for you to read (and a description of the issue in the Front Matter section, by the Guest Editors).

A link to this Special Issue

This post was inspired by the following discussion thread  https://mapleprimes.com/questions/242266-Count-The-Number-Of-Paths , which considered the problem of finding all Hamiltonian paths on an integer lattice in R^2 that connect two distinct vertices. The  AllPaths  procedure solves a more general problem: it finds all self-disjoint paths connecting two distinct vertices not only in the plane  R^2  but also in space  R^3 . Of course, it also finds all Hamiltonian paths or allows one to determine their absence. The procedure does not use commands from GraphTheory package (only direct manipulation of sets and lists).
Required parameters of the procedure: S  is a set or list of lattice vertices specified by their coordinates, Start and Finish are the initial and final vertices. Optional parameter R (defaults it's NULL  if all paths are searched) and any symbol (if only Hamiltonian paths are searched). S can be either a rectangular integer lattice or a union of several such lattices.

Code of the procedure:

restart;
AllPaths:=proc(S::{set(list),listlist},Start::list,Finish::list,R::symbol:=NULL)
local N:=nops(S), S1:=convert(S, set), L, n, m, k, i, j, s, p, q, P, a, b, c;

L:={[Start]};
for n from 2 to N do

if R=NULL then

P:='P';
m:=nops(L);
for k from 1 to m do
if nops(Start)=2 then
i,j:=L[k][-1][];
s:={[i-1,j],[i,j+1],[i+1,j],[i,j-1]} else
a,b,c:=L[k][-1][];
s:={[a-1,b,c],[a,b+1,c],[a+1,b,c],[a,b-1,c],[a,b,c-1],[a,b,c+1]} fi; 
s:=`intersect`(s,S1) minus convert(L[k],set);
if s={} and L[k][-1]=Finish then P[k]:=L[k] else
if s={} and L[k][-1]<>Finish then P[k]:=NULL else
P[k]:=`if`(L[k][-1]=Finish,L[k],seq([L[k][],s[i]],i=1..nops(s)))  fi; fi;
od;
L:=convert(P,set) else

P:='P';
m:=nops(L);
for k from 1 to m do
if nops(Start)=2 then
i,j:=L[k][-1][];
s:={[i-1,j],[i,j+1],[i+1,j],[i,j-1]} else
a,b,c:=L[k][-1][];
s:={[a-1,b,c],[a,b+1,c],[a+1,b,c],[a,b-1,c],[a,b,c-1],[a,b,c+1]} fi;
s:=`intersect`(s,S1) minus convert(L[k],set);
if n<N then 
if s={} or L[k][-1]=Finish then P[k]:=NULL else
P[k]:=seq([L[k][],s[i]],i=1..nops(s)); fi else 
if L[k][-1]=Finish and s<>{} then P[k]:=NULL else P[k]:=seq([L[k][],s[i]],i=1..nops(s));
fi; fi;  
od;
L:=convert(P,set)

fi; od;

L;
end proc:

Examples of use.

In the first example from the post above, we find the number of Hamiltonian paths in 

L:=CodeTools:-Usage(AllPaths({seq(seq([i,j], i=1..11), j=1..3)}, [2,2], [10,2], 'H')):
nops(L);

In this same example, we find the number of all paths from A to B and the possible lengths of these paths.

L:=CodeTools:-Usage(AllPaths({seq(seq([i,j], i=1..11), j=1..3)}, [2,2], [10,2])):
nops(L);
map(t->nops(t), L);
L1:=select(t->nops(t)=%[-1], L):
nops(L1);

In the following example, we find the number of all paths, as well as the number of Hamiltonian paths, and animate these paths (total 24 one's).

S:={seq(seq([i,j],i=1..5),j=1..3)} union {seq(seq([i,j],i=4..7),j=3..5)}: A:=[1,1]: B:=[7,5]:
P:=plots:-display(plots:-pointplot(S, symbol=solidcircle, color=blue, symbolsize=15, view=[0..7.5,0..6.5], size=[600,500], scaling=constrained), plots:-textplot([[A[],"A"],[B[],"B"]], font=[times,bold,22], align=[left,above])):
L:=AllPaths(S,A,B):
nops(L);
map(t->nops(t), L);
L1:=select(t->nops(t)=%[-1], L):
nops(L1);
plots:-animate((plots:-display)@(plottools:-curve),[L1[round(a)], color=red, thickness=4], a=1..%, frames=180, background=P, size=[700,500], paraminfo=false);

In the final example, we search for Hamiltonian paths in a lattice defined on the surface of a cube. Imagine a cube made of wires, and an ant must crawl along these wires from point  A(0,0,0)  to point B(2,2,2) , visiting all nodes of this lattice. Is this possible? We see that it is not. The length of the maximum path is 25, and this lattice has 26 vertices. An animation of one of the maximum paths is provided.

S:={seq(seq(seq([i,j,k],i=0..2),j=0..2),k=0..2)} minus {[1,1,1]}: A:=[0,0,0]: B:=[2,2,2]:
P:=plots:-display(plots:-pointplot3d(S, symbol=solidcircle, color=blue, symbolsize=20, scaling=constrained), plots:-textplot3d([[A[],"A"],[B[],"B"]], font=[times,bold,22], align=[left,above]), plottools:-curve([[2,0,1],[2,2,1],[0,2,1],[0,0,1],[2,0,1]], color=black,thickness=0),plottools:-curve([[1,0,2],[1,0,0],[1,2,0],[1,2,2],[1,0,2]], color=black, thickness=0),plottools:-curve([[2,1,0],[0,1,0],[0,1,2],[2,1,2],[2,1,0]], color=black,thickness=0), tickmarks=[3,3,3]):
L:=AllPaths(S,A,B):
nops(L);
map(t->nops(t), L);
L1:=select(t->nops(t)=%[-1], L):
nops(L1);
plots:-animate((plots:-display)@(plottools:-curve),[L1[-1][1..round(a)], color=red, thickness=4], a=1..25, frames=240, background=P, paraminfo=false, axes=box, labels=[x,y,z]);

We can see from this animation that the path does not pass through the vertex (0, 2, 2) .

Edit. A code error that could cause incorrect operation when using the  R  option has been fixed. Everything now works correctly. If there is no Hamiltonian path passing through all vertices, the procedure returns the empty set { } .

Paths1.mw

This system of nonlinear equations was proposed by an artificial intelligence named Alice.

f1 := x4^3+x5^2*x6-sin(x3)+exp(x1*x2)-5; 
f2 := ln(x4+x5)-x6^2+x3^3-x1^2*x2-2; 
f3 := cos(x4*x5)+x6*sin(x3)-x1*x2^2+x4^2*x5^3;

At the same time, AI accompanied it with comments. Here are some of them:
 

### System Characteristics
 **Underdetermination**: 6 unknowns with 3 equations 

### Solution Complexity

* Lack of a general analytical solution
* Need to use numerical methods
* Possibility of multiple solutions
* Complexity of visualization
* Sensitivity to initial approximations

### Solution Methods

* **Iterative methods**:
* Newton's method
* Simple iteration method
* Gradient methods
* **Numerical methods**:
* Finite difference method
* Monte Carlo method
* Genetic algorithms
* **Optimization approaches**:
* Lagrange multiplier method
* Constrained optimization methods

### Practical Application

Similar systems are found in:
* Quantum mechanics
* Field theory
* Economic modeling
* Bioinformatics
* Machine learning
* Engineering calculations of complex systems

Analyzing such systems often requires specialized software and powerful computing resources.