🔢 mathematics

Arithmetic properties of DSOME function

This paper derives a closed-form generating function for the $DSOME(n)$ function, enabling the discovery of new internal congruences modulo 4 and 8 that extend the recent work of Andrews and Ghosh Dastidar.

Original authors: Nayandeep Deka Baruah, Pankaj Gogoi

Published 2026-07-03

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Nayandeep Deka Baruah, Pankaj Gogoi

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you have a giant bag of numbered blocks. Your job is to build towers using these blocks, but there's a rule: the blocks in each tower must add up to a specific number, say 5. You can build a tower with a single block of 5, or five blocks of 1, or a 3 and a 2, and so on. In math, these towers are called partitions.

For a long time, mathematicians have been fascinated by counting how many different towers you can build for any number. But recently, two researchers (Andrews and Ghosh Dastidar) decided to look at these towers in a new way. Instead of just counting them, they started weighing them.

The "Weighing" Game: SOME and DSOME

They created two special "scales" to weigh these towers:

SOME(n): Imagine every odd-numbered block (1, 3, 5...) is made of heavy gold, and every even-numbered block (2, 4, 6...) is made of light plastic. This function adds up the weight of all the gold blocks and subtracts the weight of all the plastic blocks for every possible tower of size n.
DSOME(n): This is the same game, but with a twist. You can only build towers where no two blocks are the same size. You can't have two 2s, or two 3s. It's like building a tower where every step must be a unique size. Then, you do the same gold-minus-plastic weighing.

The paper focuses on this second, stricter version: DSOME(n).

The Big Discovery: A Secret Recipe

The authors of this paper, Nayandeep Deka Baruah and Pankaj Gogoi, wanted to understand the hidden patterns in these weights. They knew a complicated, messy recipe (a formula) existed to calculate DSOME(n), but it was hard to use to find patterns.

Their first major achievement was finding a "closed form." Think of this like taking a complex, 10-step cooking recipe with obscure ingredients and simplifying it into a single, elegant equation. They found a neat, compact formula that generates all the DSOME numbers at once.

Finding the Hidden Rhythms (Congruences)

Once they had this neat formula, they started looking for "rhythms" or repeating patterns in the numbers. In math, finding a rhythm often means discovering that certain numbers always result in a remainder of zero when divided by a specific number (like 4 or 8).

Here is what they found, explained simply:

The "Every 4th" Rule: They proved that if you look at DSOME numbers at certain intervals (like every 4th number in a specific sequence), the result is always perfectly divisible by 4. It's like a drumbeat that always lands on the downbeat.
The "Every 25th" Rule: They found that if you look at numbers that are 1 more than a multiple of 25 (like 26, 51, 76...), the result is always divisible by 4.
The "Deep" Rule (Modulo 8): They went even deeper. They discovered that for very specific, large numbers (like those ending in 26 when divided by 125), the result is divisible by 8.
The "Family Tree" Connection: Perhaps the most interesting finding is a relationship between different generations of these numbers. They showed that the value of a huge number (like 15,625) is mathematically linked to the values of much smaller numbers (like 25 or 625) in a specific equation. It's like saying the weight of a giant oak tree is exactly determined by the weights of its acorns and saplings in a precise formula.

The Crystal Ball (Conjectures)

At the end of the paper, the authors look at their data and make an educated guess (a conjecture). They suspect there are even more hidden rhythms they haven't proven yet. They guess that if you look at numbers ending in 21 (when divided by 50), the result is divisible by 8, and if you look at numbers ending in 71 (when divided by 100), the result is divisible by 16.

Summary

In short, this paper takes a complicated mathematical puzzle about weighing unique number towers, finds a simpler way to calculate the weights, and uses that simplicity to reveal hidden, repeating patterns in the numbers. They didn't just find one pattern; they found a whole family of rules that dictate how these numbers behave, and they made a guess about even more rules waiting to be discovered.

Technical Summary: Arithmetic Properties of the DSOME Function

Problem Statement
This paper investigates the arithmetic properties of the function $DSOME(n)$, recently introduced by Andrews and Ghosh Dastidar. The function is defined as the sum of all odd parts in the partitions of a positive integer $n$ into distinct parts, minus the sum of all even parts in those same partitions. While Andrews and Ghosh Dastidar established generating functions for $DSOME(n)$ and proved several congruences modulo 4 and 5, the authors of this paper aim to derive a closed-form expression for the generating function of $DSOME(n)$. This closed form serves as the foundation for discovering new congruences and internal congruence relations modulo 4 and 8.

Methodology
The authors employ the theory of $q$ -series and partition identities. The core methodology involves:

Derivation of a Closed Form: The authors transform the generating function of $DSOME(n)$, originally expressed as an infinite series involving $q$ -products, into a closed form involving the standard $q$ -products $f_n = (q^n; q^n)_\infty$ .
Dissection Techniques: To extract coefficients corresponding to specific arithmetic progressions (e.g., $5n+r$ ), the authors utilize 5-dissections and 2-dissections of the functions $f_1$ , $1/f_1$ , and the Rogers-Ramanujan functions $G(q)$ and $H(q)$ .
Algebraic Manipulation: The proofs rely heavily on identities involving the Rogers-Ramanujan continued fraction $R(q)$ and its related function $T(q) = G(q)/H(q)$ . Specifically, the authors use a set of algebraic relations (Lemma 2.1) connecting $T(q)$ and $T(q^2)$ to simplify complex expressions.
Modular Arithmetic: The authors apply congruences modulo 2, 4, and 8 to the resulting series. They frequently utilize the property that $f_k^{2\ell} \equiv f_{2k}^{2\ell-1} \pmod{2^\ell}$ to reduce terms and isolate coefficients.

Key Contributions and Results

Closed-Form Generating Function (Theorem 1.1):
The primary theoretical contribution is the expression of the generating function for $DSOME(n)$ in a closed form:
$\sum_{n=0}^{\infty} DSOME(n)q^n = \frac{1}{8} \left( \frac{f_2}{f_1} - \frac{f_1^7}{f_2^3} \right)$
This expression allows for more direct manipulation than the original series representation.
New Congruences Modulo 4 (Theorem 1.2):
Using the closed form and 5-dissection techniques, the authors prove that for all $n \ge 0$ and $1 \le r \le 4$ :
$DSOME(25n + 5r + 1) \equiv 0 \pmod 4$
This extends previous results by identifying a broader family of vanishing coefficients.
New Congruences Modulo 8 (Theorem 1.3):
The authors establish a stronger congruence modulo 8. For all $n \ge 0$ and $1 \le r \le 4$ :
$DSOME(625n + 125r + 26) \equiv 0 \pmod 8$
This result demonstrates a deeper level of arithmetic structure within the function.
Internal Congruences (Theorem 1.4):
The paper derives recursive relations (internal congruences) linking values of $DSOME(n)$ at different scales. Specifically, it proves relations connecting $DSOME(15625n + 651)$ to linear combinations of $DSOME(25n+1)$, $DSOME(625n+26)$, and $DSOME(n)$ modulo 8.

Significance and Claims
The authors state that the closed-form expression of the generating function is the key enabler for these new discoveries. By expressing the function in terms of $f_1$ and $f_2$ , they were able to apply standard dissection methods to uncover congruences that were not apparent from the original series definition.

The paper modestly concludes that while these new congruences modulo 4 and 8 are significant, "far more congruences are yet to be discovered." Based on numerical calculations, the authors propose two conjectures for future investigation:

$DSOME(50n + 21) \equiv 0 \pmod 8$
$DSOME(100n + 71) \equiv 0 \pmod{16}$

The work positions itself as a continuation of the study initiated by Andrews and Ghosh Dastidar, providing a more robust algebraic framework for analyzing the arithmetic properties of partition-related functions involving distinct parts.