A mathematical thought experiment
If mathematics
took up decompwlj
What might emerge if many mathematicians, sustained funding, and substantial computation converged on Rémi Eismann’s weight × level + jump construction?
Imagine the construction becoming a shared research object. Analytic number theorists investigate its counting functions; probabilists design comparison models; computational teams maintain reproducible censuses; students search for the first explanation of an unfamiliar column. The central opportunity would be to turn a visual and experimental language into a cumulative theory.
The exact starting point. For a strictly increasing positive integer sequence, put d(n) = a(n+1) − a(n). Set l(n) = a(n) − d(n) when this exceeds d(n), and zero otherwise. When l(n) > 0, the weight k(n) is its smallest divisor exceeding d(n), and L(n) = l(n)/k(n). Otherwise k = L = 0 and the term is unclassified. The displayed identity applies to decomposable terms.[1]
This is a canonical selection rule: the additive gap chooses a multiplicative threshold. For the consecutive primes 19 and 23, the gap is 4, the shifted term is 15, and the smallest divisor above 4 is 5. Thus 19 = 5 × 3 + 4. Across a sequence, that simple operation becomes a family of questions about divisors and spacing.
The future imagined here: a productive research field whose value comes from explanations, transferable methods, and newly answerable questions. The following pages describe possibilities, not forecasts or claims of discoveries already made.
Scope: the definitions and project references provide the foundation; the proposed research agenda is exploratory. This report does not independently certify the treatise’s analytic proofs.
A plausible research programme
From a common picture
to a theory of differences
The most fertile question may be: which features of the weight–level landscape are shared, and which remember the arithmetic of the original sequence?
1. A theory of the background
Suppose researchers compared sequences with similar density but different construction: primes, constrained integer sets, and explicitly defined random samples. They could vary one ingredient at a time—density, parity, residue restrictions, or correlations between gaps—and ask which changes alter the classification.
A successful theory might specify when different sequences have the same leading statistical behavior, and when they separate. Such a theorem would need explicit hypotheses: density alone does not determine an integer sequence’s divisor structure. The treatise’s random controls motivate this question; their numerical resemblance to primes does not establish a universal law.[2]
2. A theory of the arithmetic deviations
After identifying a suitable baseline, the residual patterns become research targets. Why is one weight column more populated than another? How do congruences favor or suppress a level? Can one derive the dependence between a shifted term’s divisors and the requirement that the next sequence element really is the next?
Here a large collaboration could produce a systematic catalogue: exact identities for individual columns, conditional counting laws, and error estimates that predict where those laws fail. The deepest payoff would be a reusable technique for counting divisor events subject to an adjacent-gap condition. That technique could matter beyond the original classification.
A precise bridge. For a decomposable term, k > L exactly when l has no divisor in (d, √l]. Indeed, kL = l makes k > L equivalent to k > √l, and k is the first divisor above d. The classification is therefore a divisor-window problem evaluated at a gap-dependent integer.
3. A ladder of genuinely new results
The programme could advance through sharper upper bounds, laws for particular level ranges, and eventually lower bounds that force infinitely many examples. These are different achievements: proving that a class becomes sparse does not show that it continues forever. A convincing lower-bound method would be especially consequential because it must preserve enough of the original arithmetic constraints.
Status anchor. The ninth-edition ledger marks rarefaction as internally audited but awaiting external refereeing, the level-one asymptotic as conditional, and infinitude of the level class as open. Its §5.5 also identifies a decoupling assumption in the asymptotic argument. These are starting points for scrutiny, not completed guarantees for the imagined programme.[2]
What people and compute could change
A shared observatory
for integer sequences
Imagine an atlas where every visible feature can be traced back to exact terms, an algorithm, a comparison model, and a mathematical explanation.
The existing project already supplies a common representation and an online atlas.[3] A research community could develop this into infrastructure for comparison. Each dataset would record its sequence definition, range, successor convention, and verification method. A visitor could select a thin band in a plot and retrieve the integers responsible for it.
| Shared capability | What it could make possible |
|---|---|
| Comparable censuses | Separate changes caused by range or sampling from persistent differences between sequences. |
| Controlled models | Test whether an apparent pattern needs arithmetic, gap correlations, or only the classifier itself. |
| Exact witnesses | Turn an anomaly into a reproducible counterexample, a sharper hypothesis, or a lemma. |
| Formal proof records | Connect a computational observation to a precise statement and, where feasible, a machine-checked proof. |
Discovery by comparison
Imagine unrelated sequences sharing the same missing region of the plane. Researchers could seek a common obstruction, test it on deliberately constructed sequences, and prove a criterion explaining the absence. The atlas would then organize mathematical mechanisms, not simply visual resemblance.
Discovery by disagreement
Imagine a prediction fitting an aggregate count while failing in a particular residue class. A larger census could locate the discrepancy; specialists could identify the dependency that averaging concealed. More precise experiments would guide the next theorem.
Make computation ask better questions
Substantial compute would support ensembles of random controls, independent implementations, targeted searches, and measurements over separate ranges. Students and automated conjecture systems could propose identities, then challenge them on separate data. Every proposed law would still need proof or a clearly stated empirical status.
The visual discipline matters too. Weight–level plots should use square canvases and equal axis scales. In ordinary coordinates, each point lies on kL = l; in logarithmic coordinates, its displacement from the diagonal records the logarithm of the ratio k/L. Linked views could add the jump without obscuring that geometry.
More speculative possibilities
What might eventually
travel beyond decompwlj?
A general theory of selection by a threshold
One ambitious direction would treat “choose the least divisor above a prescribed threshold” as an object in its own right. Which divisor configurations create abrupt changes when the threshold moves? The canonical WLJ rule would remain the distinguished case where the sequence supplies the threshold through its next gap. A broader theory could explain why whole families of plots develop similar boundaries.
Extensions with a clearly defined meaning
Another group might explore polynomial rings or other arithmetic settings. This requires more than replacing integers with new symbols: there must be a meaningful successor, a notion of size, and a rule for selecting a divisor. A successful analogue would resolve ordering and tie-breaking choices and recover useful theorems. A failed extension could also clarify what makes the integer construction distinctive.
A language that helps specialists meet
One specialist might describe a gap condition, another a missing divisor interval, and a third a geometric region. Decompwlj could supply the dictionary that makes them recognize the same question. Its largest impact might be methodological: a habit of studying adjacent additive structure and multiplicative structure together, carried into problems where the original notation eventually disappears.
| Possible outcome | What would substantiate it |
|---|---|
| A durable niche | Independent proofs, useful examples, maintained tools, and a stable research vocabulary. |
| A substantial field | Theorems spanning several sequence families, with predictive laws and explicit limits. |
| A wider influence | A method developed through WLJ solves a previously inaccessible problem elsewhere. |
Attention alone guarantees no breakthrough. Reformulating twin primes as an infinite weight column preserves the difficulty. Wider influence would require techniques that handle the coupled constraints more effectively.
In the most fruitful future, decompwlj helps mathematicians ask—and answer—new questions together.
Sources and provenance. Prepared 9 September 2026. Future scenarios are proposals, not established results.
[1] R. Eismann, founding paper, arXiv:0711.0865 — construction. [2] Pedagogical treatise, ninth edition, 16 August 2026 — current status, §§5, 7–8. [3] Project atlas. Also consulted: supplied eighth edition and PARI/GP kernel. Ninth edition governs status.