REVIEW 2 major objections 3 minor 6 references
APD profiles and transfinite asymptotic dimension
T0 review · 2 major / 3 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read This paper proves that an $\infty$-pseudometric space has transfinite asymptotic dimension at most $\omega+n$ exactly when it admits a two-term integral APD profile $(n+1,f)$, and that $m+1$-term profiles give upper bounds $m\cdot\omega+n$.
desk verdict A useful dictionary between APD profiles and transfinite asymptotic dimension, with two small but real gaps that are easy to close. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central objects are the inclusive family $M(X,d)$ of finite sets of scales at which $X$ cannot be decomposed into subspaces of scale-$i$-dimension 0, and its ordinal rank $\operatorname{Ord} M$, which is exactly the transfinite asymptotic dimension. The paper's combinatorial bridge is Proposition 1.9: $\operatorname{Ord} M \le m\cdot\omega+n$ holds if and only if there is an $m$-strategy, a rule that after seeing disjoint finite scale sets $\sigma_0,\dots,\sigma_k$ fixes the size of $\sigma_{k+1}$, such that every complete play avoids $M$. An integral APD profile is a finite array of non-decreasing functions that supplies such a strategy: the function $\alpha_k$ determines the size of the next scale set from the maximum scale seen so far, and the decomposition from the profile witnesses that the play stays outside $M$.
What would settle it
Compute, for any proposed space, the values $f(k)=\operatorname{Ord} M^{\{k,\dots,k+n\}}+1$; if for some $k<\ell$ one finds $f(k)>f(\ell)$, the profile function is not non-decreasing and the converse of Theorem 3.2 fails for that space. More directly, exhibit an $\infty$-pseudometric space that has an integral APD profile $(n+1,f)$ but transfinite asymptotic dimension greater than $\omega+n$; such a space would refute the characterization, and the failure would show up in the monotonicity or heredity step of the proof.
Extended reading notes
Core claim
The paper's central claim is a characterization: an $\infty$-pseudometric space $X$ has an integral APD profile $(n+1,f)$ if and only if $\operatorname{trasdim} X \le \omega+n$. The forward direction decomposes any prescribed pair of scales $r_0\le r_1$ into one subspace with $n+1$ scale-$r_0$-dimension-0 pieces and one subspace with $f(r_0)$ scale-$r_1$-dimension-0 pieces, which forces the ordinal rank of the family of 'bad' finite scale sets to be below $\omega+n$. The converse builds the second profile function from that ordinal rank, setting $f(k)=\operatorname{Ord} M^{\{k,\dots,k+n\}}+1$. A longer integral APD profile $(n+1,\alpha_1,\dots,\alpha_m)$ yields the upper bound $\operatorname{trasdim} X \le m\cdot\omega+n$, by using the profile functions to build an $m$-strategy in the combinatorial game that characterizes $m\cdot\omega+n$.
Load-bearing premise
The load-bearing premise, never stated as a lemma, is that scale-$r$-dimension 0 is inherited at smaller scales: if a subspace has uniformly bounded scale-$r$ components and $j\le r$, then its scale-$j$ components are also uniformly bounded, and the constructions in Theorems 3.2 and 3.4 both rely on this monotonicity.
Editorial extensions
If this is right
- Two-term integral APD profiles are a complete certificate for transfinite asymptotic dimension at most $\omega+n$: every space at that level has one, and every space with one is at that level.
- Spaces with transfinite asymptotic dimension below $\omega+\omega$ are exactly those admitting an integral APD profile made of two functions, so the first infinite block of levels is profile-classifiable.
- Every space with an $(m+1)$-term integral APD profile has transfinite asymptotic dimension at most $m\cdot\omega+n$, giving a profile-based sufficient condition at every finite multiple of $\omega$.
- Asymptotic property D, defined as admitting some integral APD profile, forces transfinite asymptotic dimension below $\omega\cdot\omega$.
- A space with transfinite asymptotic dimension $\omega+1$, whose existence is cited from the literature, must admit a profile $(2,f)$ and cannot admit a profile $(1,g)$, so APD profiles separate $\omega+1$ from $\omega$.
Reading between the lines
- The paper proves only one direction of the $m$-term generalization; a natural test is whether every space with $\operatorname{trasdim} X \le m\cdot\omega+n$ admits an $(m+1)$-term profile, which would make profiles a complete invariant below $\omega\cdot\omega$.
- The unstated monotonicity of scale-$r$-dimension 0 deserves a standalone proof; if it failed, the converse direction of Theorem 3.2 and the strategy construction of Theorem 3.4 would need repair, so a counterexample there would directly limit the scope of the theorems.
- The game-theoretic formulation suggests a definition of transfinite asymptotic dimension in terms of the size game, which could be used to search for a space with asymptotic property C but not asymptotic property D, the open question the paper poses.
- Because profile length $m+1$ gives only an upper bound, the sharpness question for each $m,n$ is open; testing spaces whose transfinite dimension is exactly $m\cdot\omega+n$ would show whether the bound is ever strict.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper develops APD profiles for ∞-pseudometric spaces and connects them with transfinite asymptotic dimension. The main results are Theorem 3.2, an equivalence between possessing an integral APD profile of the form (n+1,f) and having transfinite asymptotic dimension at most ω+n, and Theorem 3.4, which gives the upper bound trasdim X ≤ m·ω+n from an integral APD profile of length m+1. The proofs use Borst's ordinal rank Ord on families of finite subsets of N and a game-theoretic strategy characterization from Proposition 1.9. The paper also records corollaries relating asymptotic property C and asymptotic property D and discusses the Satkiewicz omega conjecture in light of the recent counterexample by Wu and Zhu.
Significance. If the results are correct, Theorem 3.2 is a clean and exact bridge between Dydak's APD-profile language and Radul's transfinite asymptotic dimension at the level ω+n, and Theorem 3.4 provides a natural sufficient condition at multiples of ω. The arguments are compact and proceed by direct ordinal manipulation, with no fitted parameters and no circularity between the two notions being compared. The contribution is moderate but useful for the ongoing study of transfinite asymptotic dimension and asymptotic property D. The paper does not provide machine-checked proofs, but the mathematical structure is transparent enough that the main equivalence is verifiable by a reader.
major comments (2)
- [Definition 3.1; used in Theorems 3.2 and 3.4] Definition 3.1 as printed says that X0 has scale-r0-dimension 0, which makes the constant α0 irrelevant. However, both Theorem 3.2 (where Y0 is decomposed into n+1 pieces) and Theorem 3.4 (where X0 is decomposed into |σ0| pieces) require X0 to have scale-r0-dimension at most α0−1. The definition should read 'at most α0−1 for i=0' (or an equivalent formulation), with α0=n+1 in the two theorems.
- [Theorems 3.2 and 3.4] The proofs rely on an unstated hereditary property of scale-r-dimension 0: if j≤r and a subspace has scale-r-dimension 0, then it has scale-j-dimension 0. This property is used when X_{r0},...,X_{r0+n} are declared to have scale-r0-dimension 0 from scale-i-dimension 0 for i≥r0 in Theorem 3.2, when f(k)=Ord M^{{k,...,k+n}}+1 is asserted to be non-decreasing in Theorem 3.2, and when X_{k,j} is declared to have scale-j-dimension 0 in Theorem 3.4. The fact is true and elementary, since every scale-j chain is a scale-r chain for j≤r, but it is nowhere stated or proved. Please add it as an explicit lemma before Theorem 3.2 and invoke it at the three places above.
minor comments (3)
- [Definition 1.8(3)] In condition 3 of the strategy definition, the case k=0 is written as '(σ0,...,σ_{k−1},τ)', which is undefined because σ_{−1} does not exist; the base case k=0 should be stated separately.
- [Theorem 3.2, first part] The sentence 'Using Lemma 1.2 and Corollary 1.7, we finish one part of the proof' should be expanded: the argument shows Ord M^σ < ω for every σ of cardinality n+1, and Corollary 1.7 is then applied with α=ω and p=n+1 to conclude Ord M ≤ ω+n.
- [Throughout] There are a few typographical errors: 'aymptotic' in Question 3.6, 'dime nsions' in the abstract, and the arXiv number in reference [3] appears to contain an extra digit (1612.067771v4 rather than 1612.06777v4).
Circularity Check
No circularity: Theorem 3.2 is a direct equivalence between independently defined invariants; the flagged issues are an unstated monotonicity lemma and a typo in Definition 3.1, neither of which makes the conclusion an input to the proof.
full rationale
The paper does not contain a circular derivation. Transfinite asymptotic dimension is defined through Ord M (Definitions 2.3-2.6, following Radul), while APD profiles are defined separately through scale-r-dimension 0 decompositions (Definition 3.1, following Dydak). In Theorem 3.2, the forward direction assumes an integral APD profile (n+1,f) and bounds the cardinality of members of M^sigma, then applies Lemma 1.2 and Corollary 1.7 to conclude Ord M <= omega+n. The converse defines f(k) = Ord M^{{k,...,k+n}} + 1 and shows that the profile condition follows from tau notin M, i.e. from the existence of a cover by scale-i-dimension 0 subspaces. This is a normal equivalence proof, not an instance of one notion being defined in terms of the other. No parameter is fitted to data, no benchmark is author-defined, and no load-bearing self-citation occurs: references [3], [4], and [6] are external works. The proof does contain two genuine gaps that are correctness issues rather than circularity. First, the monotonicity of f in Theorem 3.2 ('It is not hard to check') and the conversions in Theorem 3.4 from scale-r_k-dimension 0 to scale-j-dimension 0 require an unstated hereditary property: if j <= r and a subspace has scale-r-dimension 0, then it has scale-j-dimension 0. Second, Definition 3.1 as printed says '0 for i = 0', which contradicts the proof's use of alpha_0 = n+1 in Theorem 3.2. Neither gap makes the target theorem equivalent to its own assumptions; both are repairable by adding a lemma and correcting the typo. The central claim is therefore self-contained against the independently defined notions, and the appropriate circularity score is 0.
Assumptions & free parameters
assumptions (4)
- standard math Ordinal arithmetic and transfinite induction are valid for Ord M.
- domain assumption The families A(X,d) and M(X,d) are inclusive.
- domain assumption Scale-r-dimension 0 is hereditary downward: if j is at most r and a set has scale-r-dimension 0, then it has scale-j-dimension 0.
- domain assumption The family M is monotone under increasing indices: if sigma is in M and tau is obtained by increasing indices, then tau is in M.
Cite this review
Pith. "Pith review of APD profiles and transfinite asymptotic dimension." pith.science (2026). https://pith.science/paper/3LOUQDCZ
@misc{pith2026190811620,
author = {Pith},
title = {Pith review of: APD profiles and transfinite asymptotic dimension},
year = {2026},
howpublished = {\url{https://pith.science/paper/3LOUQDCZ}},
note = {Machine review of arXiv:1908.11620}
}
abstract
We develop the theory of APD profiles introduced by J. Dydak for $\infty$-pseudometric spaces. We connect them with transfinite asymptotic dimension defined by T. Radul. We give a characterization of spaces with transfinite asymptotic dimension at most $\omega+n$ for $n\in\omega$ and a sufficient condition for a space to have transfinite asymptotic dimension at most $m\cdot \omega+n$ for $m,n\in\omega$, using the language of APD profiles.
Reference graph
Works this paper leans on
-
[1]
Borst, Classification of weakly infinite-dimensional spaces , Fund
P. Borst, Classification of weakly infinite-dimensional spaces , Fund. Math. 130 (1988), 1–306
work page 1988
-
[2]
Dranishnikov, Asymptotic topology, Russ
A. Dranishnikov, Asymptotic topology, Russ. Math. Surv. 55 (2000), 1085–1129
work page 2000
-
[3]
Time evolution of coupled spin systems in a generalized Wigner representation
J. Dydak, Matrix algebra of sets and variants of decomposition comple xity, arXiv:1612.067771v4
-
[4]
T. Radul, On transfinite extension of asymptotic dimension , Topology and its Applications 157 (2010), 2292–2296
work page 2010
-
[5]
Transfinite Asymptotic Dimension
M. Satkiewicz, Transfinite Asymptotic Dimension , arXiv:1310.1258v1
-
[6]
Y. Wu, J. Zhu, A metric space with its transfinite asymptotic dimension ω + 1, arXiv:1908.00434v1. 6
work page Pith review arXiv 1908
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.