On the Fixed-Length-Burst Levenshtein Ball with Unit Radius

Gennian Ge; Yuanxiao Xi; Yubo Sun

REVIEW 3 major objections 4 minor 18 references

On the Fixed-Length-Burst Levenshtein Ball with Unit Radius

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper establishes a closed-form formula for the exact size of any unit-radius fixed-length burst Levenshtein ball, then derives extremal, average, and concentration results.

desk verdict Strong exact-count contribution with a genuine but fixable arithmetic slip in the expectation theorem; worth refereeing once corrected. read the letter →

arxiv 2506.07817 v1 pith:S5HWTHEM submitted 2025-06-09 cs.IT math.COmath.IT

Yuanxiao Xi , Yubo Sun , Gennian Ge This is my paper

classification cs.ITmath.COmath.IT MSC 05A1594B60

keywords fixed-lengthLevenshteinballburstdeletioninsertionb-runs2b-periodicsubstringscardinalityconcentrationboundssynchronizationerrors

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

A fixed-length burst Levenshtein ball collects all words reachable from a given center word by deleting one consecutive block of $b$ symbols and then inserting a consecutive block of $b$ symbols. This paper claims that, for every center word and every alphabet size, the exact size of that ball is given by a closed formula built from simple pattern counts of the center: the number of $b$-periodic runs, the number of positions where the period-$b$ equality fails at two nearby lags, and the number of long $2b$-periodic stretches bounded by such failures. If the formula is correct, ball sizes become computable in linear time rather than by enumeration, and the same counting machinery yields the smallest and largest possible balls, the average ball size over all centers, and a high-probability concentration bound around the average. The result extends the known radius-one analysis of ordinary Levenshtein balls to the burst case, where the balls are exponentially larger in $b$ and the intersection structure is more complicated.

What carries the argument

The load-bearing machinery is a structural description of $X_i\cap X_j$, the intersection of the two b-burst-insertion balls obtained after deleting a b-block from the i-th and j-th b-runs of the center. For two insertion-ball centers that agree outside a block of length $d$, Lemma 2.5 says the intersection is either a family of $q^{b-d}$ words when $d\le b$, or is contained in two explicitly described words when $d\ge b+1$. The paper repackages this as a disjoint union $X_i\cap X_j = A_{i,j}\sqcup B_{i,j}$, where $A_{i,j}$ is a small set anchored at $x$ and $B_{i,j}$ is either empty or a $q^{b-d}$-sized set produced by a $2b$-periodic alignment; Claims 4.2 and 4.3 then prove that all triple-overlap corrections reduce to the $B_{i,k}$ pieces, so the final inclusion-exclusion collapses into the b-run and $2b$-periodic sums of Theorem 3.3.

What would settle it

A direct brute-force check for small parameters settles the formula: for instance, with $q=2$, $n=6$, $b=2$ (and $q=3$, $n=7$, $b=2$), enumerate every center $x\in\Sigma_q^n$, compute $|L^b_1(x)|$ by explicitly deleting each length-$b$ block and inserting each length-$b$ block, and compare with the right-hand side of Theorem 3.3; any mismatch refutes the formula and would pinpoint which intersection piece in Claim 4.1 fails.

Watch

Extended reading notes

Core claim

The central claim is Theorem 3.3: for $n\ge b+1$ and every center $x\in\Sigma_q^n$, the fixed-length b-burst Levenshtein ball of unit radius has size $$|L^b_1(x)| = $q^{{b-1}}$\bigl((n-b+1)(q-1)-1\bigr)r_b(x) + $2q^{{b-1}}$ - \sum_{j=1}^{b-1} $q^{{b-j-1}}$ f_{b,j}(x) - \sum_{i=1}^{n-2b} g_{b,i}(x),$$ where $r_b(x)$ counts the maximal $b$-periodic substrings (b-runs) of $x$, $f_{b,j}(x)$ counts positions $i$ at which both $x_i\ne x_{i+b}$ and $x_{i+j}\ne x_{i+j+b}$, and $g_{b,i}(x)$ counts positions $j$ for which $x_i\ne x_{i+b}$, $x_j\ne x_{j+b}$, and the block $x_{[i,j+b]}$ is $2b$-periodic. The argument writes the ball as the union of unit-radius b-burst-insertion balls over all distinct b-burst deletions of $x$, decomposes every pairwise intersection into two disjoint pieces, and shows that after inclusion-exclusion only the second-kind pieces of non-adjacent intersections survive. From this exact count the paper derives the maximum ball size and the centers that attain it for $q\ge 3$, the exact expected size over a uniformly random center, and a concentration bound of order $\sqrt{n^3\log_q n}$.

Load-bearing premise

The cardinality formula rests on the cited characterization of the intersection of two unit-radius b-burst-insertion balls (Lemma 2.5); if that characterization misses any case or imposes any unstated condition, the disjoint-union claim $X_i\cap X_j=A_{i,j}\sqcup B_{i,j}$ and every later inclusion-exclusion count would have to be revised.

Editorial extensions

If this is right

For any fixed $b$, the formula evaluates $|L^b_1(x)|$ in $O(n)$ time from the indicators $x_i\ne x_{i+b}$, so candidate centers for burst-error-correcting codes can be screened exactly rather than by enumerating all reachable words.
For $q\ge 3$, the maximum-size balls are exactly those whose centers satisfy $x_i\ne x_{i+b}$ and $x_i\ne x_{i+2b}$, and the paper also records a binary maximum of $2^{b-1}n^2 - 2^{b/2}n^{3/2} + O(n)$.
The expected ball size over uniform centers is exactly computable and equals $q^{b-2}(q-1)^2 n^2 + O(n)$, giving a benchmark against which any proposed code can be compared.
With probability at least $1-3/n$, a uniformly random center has ball size within $C\sqrt{n^3\log_q n}$ of the average, so almost all centers are nearly average, which is the quantitative input needed for probabilistic code constructions.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Beyond the paper: the same $A_{i,j}\sqcup B_{i,j}$ intersection decomposition could likely be adapted to radius $t>1$ once a multi-burst analogue of Lemma 2.5 is available; the $2b$-periodic sums would then index longer periodic blocks rather than intervals of length $j+b$.
Beyond the paper: because the exact expectation formula is built from linear expectations of the $f$ and $g$ counts, the same machinery could compute the variance of ball sizes, which the paper does not do and which would sharpen the concentration constant $C$.
Beyond the paper: the formula isolates the $2b$-periodic correction as the only negative contribution beyond the leading b-run term, suggesting a testable design heuristic—insert long $2b$-periodic segments into a center to reduce the number of distinct words at burst distance one.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Worth your time: the paper gives the first exact count of unit-radius fixed-length b-burst Levenshtein balls, and that result is probably right. But the expectation theorem has a q^2 indexing error and the binary max is only announced. This is a conditional accept, not a reject.

The genuinely new piece is Theorem 3.3. Previous exact ball-size work stopped at b=1; here the authors express |L^b_1(x)| through the number of b-runs, a break-position term f_{b,j}, and a 2b-periodic count g_{b,i}. The inclusion-exclusion over insertion balls of the b-deletion products is carried out in enough detail for a skeptical reader to check. The q>=3 maximum, with the balanced-sequence extremal condition, also looks right, and the concentration result on a code of density 1 - 3/n is a real addition.

The soft spots are real. Theorem 3.5 subtracts sum_{i,j} k_{i,j}/q^{j+b-i-1}, but Lemma 6.3 establishes E[g_{b,i}(x)] = sum_j k_{i,j}/q^{j+b-i+1}. Same k, same range, denominator exponent differs by 2. For b=1, q=2, n=3 the printed formula gives 5; direct enumeration gives 5.75. That is a concrete error in a stated exact result, almost certainly a typo, but it has to be fixed. The abstract promises extremal bounds including maximum for all alphabets, but for the binary case the paper only states max = 2^{b-1}n^2 - 2^{b/2} n^{3/2} + O(n) and then says the details are omitted because the discussion is 'convoluted' and the improvement is 'small.' Calling that a comprehensive solution is a stretch. Dependence on Lemma 2.5 from the authors' own prior paper is fine--it is a cited, published characterization--but it does mean the core formula is only as solid as that intersection lemma. No evidence here of a missing case; it is a dependency, not a demonstrated flaw.

For a coding theorist working on synchronization channels or burst-correcting codes, this is a useful and citable contribution once the expectation formula is corrected. A serious referee should spend time on it. My recommendation: send it to peer review with a request to fix the q^2 offset, reprove/verify the expectation, and either complete the binary maximum proof or demote it to a conjecture.

Referee Report

3 major / 4 minor

Summary. The paper studies the fixed-length b-burst Levenshtein ball of unit radius, i.e., the set of length-n sequences obtainable from a center by one burst of b deletions followed by one burst of b insertions. The main result is Theorem 3.3, an exact closed-form formula for the cardinality of this ball in terms of the number of b-runs of the center, certain alternating-position counts f_{b,j}, and counts g_{b,i} of 2b-periodic windows. Building on this formula, the paper derives an upper bound on the maximum ball size, claims an asymptotic value for the binary maximum, gives an exact expected-size formula (Theorem 3.5), and proves a concentration bound around the expectation (Theorem 3.6). The technical engine is a structural decomposition of intersections of burst-insertion balls, following the authors' earlier work.

Significance. If Theorem 3.3 is correct, it provides the first exact cardinality formula for a multi-deletion/multi-insertion Levenshtein ball, going substantially beyond the radius-one single-symbol case of Sala-Dolecek and Bar-Lev et al. The formula has a uniform structure with a number of summands independent of the center, which is a genuine improvement in tractability. The expected-size and concentration results also address natural questions for coding-theoretic applications. The paper is built on machine-checkable-style combinatorial counting and mostly gives complete proofs for the cardinality and upper-bound claims. However, the exact expectation formula in Theorem 3.5 is internally inconsistent with Lemma 6.3 by a q^2 factor, and the binary maximum asymptotic is asserted without proof; these issues currently prevent the paper from being fully reliable as a comprehensive treatment of extremal and average sizes.

major comments (3)

[Section 6, Theorem 3.5] The displayed formula in Theorem 3.5 subtracts sum_{i=1}^{n-2b} sum_{j=i+b}^{n-b} k_{i,j}/q^{j+b-i-1}, but the computation in Lemma 6.3 gives E[g_{b,i}(x)] = sum_{j=i+b}^{n-b} k_{i,j}/q^{j+b-i+1}. The denominator exponent in Theorem 3.5 is therefore off by 2, a factor of q^2. For a concrete check, take b=1, q=2, n=3: the printed formula yields E=5, whereas direct enumeration over the 8 binary centers gives E=5.75, and changing the exponent to j+b-i+1 gives 5.75. The proof of Theorem 3.5 as written does not follow from Lemmas 6.1-6.3, and the exact expected-size claim needs correction.
[Section 6, Lemma 6.3] The statement of Lemma 6.3 has an extra outer sum over i on the right-hand side: for a fixed i it reads E[g_{b,i}(x)] = sum_{i=1}^{n-2b} sum_{j=i+b}^{n-b} k_{i,j} q^{-(j+b-i+1)}. The proof of the lemma computes only the single sum over j, which is what the left-hand side requires. This typographical/notational error compounds the exponent mismatch in Theorem 3.5 and must be fixed before the expectation result can be evaluated.
[Section 3.C] The claimed binary-alphabet maximum, max_x |L^b_1(x)| = 2^{b-1} n^2 - 2^{b/2} n^{3/2} + O(n), is stated without proof, with the text saying only that a similar method works and that details are omitted. Since the abstract and the contribution list advertise extremal maximum bounds as one of the paper's deliverables, this unproved asymptotic is not sufficient. Either a proof should be supplied or the claim should be explicitly labeled as a conjecture or deferred result.

minor comments (4)

[Section 1] In the organization paragraph, 'retailed proofs' should be 'detailed proofs'.
[Section 3.B] In the definition of f_{b,j}(x), the range 'j ∈ [b−1]' should be written as 'j ∈ [1, b−1]' for clarity.
[Section 7, Lemma 7.3] The inequality g_{b,i}(x) ≤ 2 log_q n follows from the construction only up to an additive constant depending on b; the constant does not affect the claimed concentration bound, but the statement should use O(log n) rather than an exact bound.
[Section 3.B] The sentence 'In the rest of this paper... we assume n ≥ 2b+1' should be reconciled with Theorem 3.3, which is stated for n ≥ b+1; the assumption is fine, but the transition should be explicit that the formula remains valid with empty sums for smaller n.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the ball-size formula is derived by union/inclusion-exclusion, not assumed; self-cited intersection lemmas are external theorems about insertion balls.

full rationale

The central result, Theorem 3.3, is not assumed anywhere in the derivation. The proof decomposes L^b_1(x) as the union of burst-insertion balls centered at the distinct burst-deletion results (Lemma 4.1), applies inclusion-exclusion (Claim 4.3), and translates the correction terms into the f and g counts by direct combinatorial counting. The formula is the output, not the input. The cited Lemma 2.5 characterizes intersections of burst-insertion balls, not Levenshtein balls, and is an external theorem from [16]; using it as a tool is ordinary mathematical practice rather than circularity. The same holds for Lemmas 2.3 and 2.4, which are independent statements about deletion/insertion ball sizes. The expectation section is a direct application of linearity of expectation using the already-derived explicit formula; no parameter is fitted and no quantity is renamed as a prediction. The only notable defect found is internal and arithmetic, not circular: Theorem 3.5 prints the g-sum with denominator exponent j+b-i-1, while substituting Lemma 6.3's E[g_{b,i}(x)] = sum k_{i,j}/q^{j+b-i+1} into the surrounding linearity expression yields exponent j+b-i+1, a q^2 discrepancy. This is a correctness/indexing inconsistency in the stated expectation, not a case of the conclusion being presupposed. The paper is not fully self-contained because it imports several lemmas from the authors' earlier work, but those cited statements are independent theorems about different objects, so self-citation does not constitute circularity here.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The paper contributes derivations on top of standard combinatorial background and four prior results about burst deletions and insertions, three of which come from the authors' own earlier papers ([10] and [16]). No free parameters are fitted to data, and no new entities are introduced.

assumptions (4)

domain assumption Size of the b-burst-deletion ball is |D^b_1(x)| = r_b(x), together with the b-run characterization in Lemma 2.2.
Taken from prior work [16, Claim 3.1] and [10, Lemma IV.4]; used in Lemma 4.1 and Theorem 3.1.
domain assumption Size of the b-burst-insertion ball is q^{t(b-1)} sum_{i=0}^t binom(m+t,i)(q-1)^i (Lemma 2.4).
Taken from [10, Theorem III.2]; used in Theorem 3.1 and Lemma 4.1.
domain assumption Intersection structure of two b-burst-insertion balls (Lemma 2.5).
Taken from [16, Claims 3.2 and 3.3]; the main proof of Claim 4.1 reduces X_i cap X_j to A_{i,j} and B_{i,j} using this lemma.
standard math Hoeffding's inequality and the union bound in Lemma 7.2.
Used to bound the probability that a random string falls outside the code C in Construction 7.1.

how reviews work

0 comments

Cite this review

Pith. "Pith review of On the Fixed-Length-Burst Levenshtein Ball with Unit Radius." pith.science (2026). https://pith.science/paper/S5HWTHEM

@misc{pith2026250607817,
  author       = {Pith},
  title        = {Pith review of: On the Fixed-Length-Burst Levenshtein Ball with Unit Radius},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S5HWTHEM}},
  note         = {Machine review of arXiv:2506.07817}
}

abstract

Consider a length-$n$ sequence $\bm{x}$ over a $q$-ary alphabet. The \emph{fixed-length Levenshtein ball} $\mathcal{L}_t(\bm{x})$ of radius $t$ encompasses all length-$n$ $q$-ary sequences that can be derived from $\bm{x}$ by performing $t$ deletions followed by $t$ insertions. Analyzing the size and structure of these balls presents significant challenges in combinatorial coding theory. Recent studies have successfully characterized fixed-length Levenshtein balls in the context of a single deletion and a single insertion. These works have derived explicit formulas for various key metrics, including the exact size of the balls, extremal bounds (minimum and maximum sizes), as well as expected sizes and their concentration properties. However, the general case involving an arbitrary number of $t$ deletions and $t$ insertions $(t>1)$ remains largely uninvestigated. This work systematically examines fixed-length Levenshtein balls with multiple deletions and insertions, focusing specifically on \emph{fixed-length burst Levenshtein balls}, where deletions occur consecutively, as do insertions. We provide comprehensive solutions for explicit cardinality formulas, extremal bounds (minimum and maximum sizes), expected size, and concentration properties surrounding the expected value.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

18 extracted references · 18 canonical work pages

[1]

On the Size of Balls and Anticodes of Small Diameter Under the Fixed-Length Levenshtein Metric,

D. Bar-Lev, T. Etzion, and E. Yaakobi, “On the Size of Balls and Anticodes of Small Diameter Under the Fixed-Length Levenshtein Metric,” IEEE Transactions on Information Theory , vol. 69, no. 4, pp. 2324-2340, 2023

work page 2023
[2]

Some general results of coding theory with applications to the study of codes for the correction of synchronization errors,

L. Calabi and W. E. Hartnett, “Some general results of coding theory with applications to the study of codes for the correction of synchronization errors,” Information and Control , vol. 15, no. 3, pp. 235-249, 1969

work page 1969
[3]

Coding for racetrack memories,

Y . M. Chee, H. M. Kiah, A. Vardy, V . K. Vu, and E. Yaakobi, “Coding for racetrack memories,” IEEE Transactions on Information Theory , vol. 64, no. 11, pp. 7094-7112, 2018

work page 2018
[4]

Block edit errors with transpositions: deterministic document exchange protocols and almost optimal binary codes,

K. Cheng, Z. Jin, X. Li, and K. Wu, “Block edit errors with transpositions: deterministic document exchange protocols and almost optimal binary codes,” in 46th International Colloquium on Automata, Languages, and Programming (ICALP) , vol. 132, 2019, pp. 37:1-37:15

work page 2019
[5]

Tight bounds on the number of string subsequences,

D. S. Hirschberg and M. Regnier, “Tight bounds on the number of string subsequences,” Journal of Discrete Algorithms, vol. 1, no. 1, pp. 123-132, 2000

work page 2000
[6]

The size of Levenshtein ball with radius 2: Expectation and concentration bound,

L. He and M. Ye, “The size of Levenshtein ball with radius 2: Expectation and concentration bound,” in Proceeding of the International Symposium on Information Theory (ISIT) , Taipei, Taiwan, 2023, pp. 850-855

work page 2023
[7]

Binary codes capable of correcting deletions, insertions, and reversals,

V . I. Levenshtein, “Binary codes capable of correcting deletions, insertions, and reversals,” Soviet Physics Doklady , vol. 10, no. 8, pp. 707-710, 1966

work page 1966
[8]

Asymptotically optimum binary code with correction for losses of one or two adjacent bits,

V . Levenshtein, “Asymptotically optimum binary code with correction for losses of one or two adjacent bits,” Systems Theory Research , vol. 19, no. 4, pp. 298-304, 1970

work page 1970

Show all 18 references

[9]

A characterization of the number of subsequences obtained via the deletion channel,

Y . Liron and M. Langberg, “A characterization of the number of subsequences obtained via the deletion channel,” IEEE Transactions on Information Theory, vol. 61, no. 5, pp. 2300-2312, 2015

2015
[10]

Sequence reconstruction under channels with multiple bursts of insertions or deletions,

Z. Lan, Y . Sun, W. Yu, and G. Ge, “Sequence reconstruction under channels with multiple bursts of insertions or deletions,” arXiv:2504.20460, 2025

2025 arXiv
[11]

On the number of subsequences when deleting symbols from a string,

H. Mercier, M. Khabbazian, and V . K. Bhargava, “On the number of subsequences when deleting symbols from a string,” IEEE Transactions on Information Theory, vol. 54, no. 7, pp. 3279-3285, 2008

2008
[12]

Random access in large-scale DNA data storage,

L. Organick et al., “Random access in large-scale DNA data storage,” Nature Biotechnology, vol. 36, no. 3, pp. 242-248, 2018

2018
[13]

Counting sequences obtained from the synchronization channel,

F. Sala and L. Dolecek, “Counting sequences obtained from the synchronization channel,” in Proceeding of the International Symposium on Information Theory (ISIT), Istanbul, Turkey, 2013, pp. 2925-2929

2013
[14]

Asymptotically Optimal Codes for (t, s)-Burst Error,

Y . Sun, Z. Lu, Y . Zhang, and G. Ge, “Asymptotically Optimal Codes for (t, s)-Burst Error,” IEEE Transactions on Information Theory , vol. 71, no. 3, pp. 1570-1584, 2025

2025
[15]

Codes correcting a burst of deletions or insertions,

C. Schoeny, A. Wachter-Zeh, R. Gabrys, and E. Yaakobi, “Codes correcting a burst of deletions or insertions,” IEEE Transactions on Information Theory, vol. 63, no. 4, pp. 1971-1985, 2017

1971
[16]

Sequence reconstruction under single-burst-insertion/deletion/edit channel,

Y . Sun, Y . Xi, and G. Ge, “Sequence reconstruction under single-burst-insertion/deletion/edit channel,” IEEE Transactions on Information Theory , vol. 69, no. 7, pp. 4466-4483, 2023

2023
[17]

Non-binary codes for correcting a burst of at most t deletions,

S. Wang, Y . Tang, J. Sima, R. Gabrys, and F. Farnoud, “Non-binary codes for correcting a burst of at most t deletions,” IEEE Transactions on Information Theory, vol. 70, no. 2, pp. 964-979, 2024

2024
[18]

On the size distribution of Levenshtein balls with radius one,

G. Wang and Q. Wang, “On the size distribution of Levenshtein balls with radius one,” Designs, Codes and Cryptography , vol. 92, pp. 2253-2265, 2024

2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

On the Size of Balls and Anticodes of Small Diameter Under the Fixed-Length Levenshtein Metric,

D. Bar-Lev, T. Etzion, and E. Yaakobi, “On the Size of Balls and Anticodes of Small Diameter Under the Fixed-Length Levenshtein Metric,” IEEE Transactions on Information Theory , vol. 69, no. 4, pp. 2324-2340, 2023

work page 2023

[2] [2]

Some general results of coding theory with applications to the study of codes for the correction of synchronization errors,

L. Calabi and W. E. Hartnett, “Some general results of coding theory with applications to the study of codes for the correction of synchronization errors,” Information and Control , vol. 15, no. 3, pp. 235-249, 1969

work page 1969

[3] [3]

Coding for racetrack memories,

Y . M. Chee, H. M. Kiah, A. Vardy, V . K. Vu, and E. Yaakobi, “Coding for racetrack memories,” IEEE Transactions on Information Theory , vol. 64, no. 11, pp. 7094-7112, 2018

work page 2018

[4] [4]

Block edit errors with transpositions: deterministic document exchange protocols and almost optimal binary codes,

K. Cheng, Z. Jin, X. Li, and K. Wu, “Block edit errors with transpositions: deterministic document exchange protocols and almost optimal binary codes,” in 46th International Colloquium on Automata, Languages, and Programming (ICALP) , vol. 132, 2019, pp. 37:1-37:15

work page 2019

[5] [5]

Tight bounds on the number of string subsequences,

D. S. Hirschberg and M. Regnier, “Tight bounds on the number of string subsequences,” Journal of Discrete Algorithms, vol. 1, no. 1, pp. 123-132, 2000

work page 2000

[6] [6]

The size of Levenshtein ball with radius 2: Expectation and concentration bound,

L. He and M. Ye, “The size of Levenshtein ball with radius 2: Expectation and concentration bound,” in Proceeding of the International Symposium on Information Theory (ISIT) , Taipei, Taiwan, 2023, pp. 850-855

work page 2023

[7] [7]

Binary codes capable of correcting deletions, insertions, and reversals,

V . I. Levenshtein, “Binary codes capable of correcting deletions, insertions, and reversals,” Soviet Physics Doklady , vol. 10, no. 8, pp. 707-710, 1966

work page 1966

[8] [8]

Asymptotically optimum binary code with correction for losses of one or two adjacent bits,

V . Levenshtein, “Asymptotically optimum binary code with correction for losses of one or two adjacent bits,” Systems Theory Research , vol. 19, no. 4, pp. 298-304, 1970

work page 1970