Flag-transitive $4$-designs and $PSL(2,q)$ groups

Huili Dong

REVIEW 3 major objections 5 minor 15 references

Flag-transitive $4$-designs and $PSL(2,q)$ groups

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read For flag-transitive 4-designs with automorphism group $PSL(2,q)$, only eight parameter sets survive for $5\le\lambda\le10$ (two left undecided), and several $\lambda>10$ families are impossible.

desk verdict A plausible but conditional extension of the Dai–Li classification; sound counting, under-documented Magma exclusions, and typos around the computations. read the letter →

arxiv 1908.00760 v3 pith:BMY4EK3K submitted 2019-08-02 math.CO

Huili Dong This is my paper · ORCID

classification math.CO MSC 05B0505B2520B25

keywords 4-designflag-transitivePSL(2q)blockdesignprojectivespeciallineargroupsubgroupclassificationorbitlengthsdivisibilityconditions

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

The paper asks which flag-transitive 4-designs (designs in which the automorphism group acts transitively on incident point-block pairs) exist when the automorphism group is the simple group $PSL(2,q)$. It claims that in the block-index range $5\le\lambda\le10$ the answer is almost a closed list: up to isomorphism, only eight such designs exist, with two parameter sets left undecided, and that for $\lambda>10$ several large families of block stabilizers admit no designs at all. A sympathetic reader would care because it turns a potentially infinite family of designs into a finite, checkable list and leaves the remaining uncertainty at two explicitly named parameter sets.

What carries the argument

The central object is the permutation action of $G=PSL(2,q)$ on the $q+1$ points of the projective line, with $G_B$ the setwise stabilizer of a block $B$ and $G_{xB}=G_x\cap G_B$ the stabilizer of an incident point-block pair. The argument is carried by two divisibility identities obtained from standard design counting: $\lambda(q-2)=k(k-1)(k-2)(k-3)/(n|G_B|)$ and $q=(k-1)(k-2)(k-3)/(\lambda n|G_{xB}|)+2$, where $n=\gcd(2,q-1)$. Combined with the fact that $G_B$ must be one of twelve subgroup types of $PSL(2,q)$, these identities cut the infinite search space down to short finite lists of candidate parameter sets; the survivors are then separated by checking orbit-length partitions and, in a few cases, by computer enumeration.

What would settle it

For the dismissed parameter set $(v,k,\lambda)=(48,12,11)$ with $G=PSL(2,47)$, enumerate all orbits of a block stabilizer $A_4$ on the 48 points and check whether any union of orbits forms a 4-design; if one does, Theorem 1.1 fails. More generally, rerun the full computer-assisted elimination for every candidate the paper discards, and separately decide the two undecided cases by checking whether the specified groups admit a 4-design with the given block size.

Watch

Extended reading notes

Core claim

The paper establishes two theorems. Theorem 1.1 states that if $\mathcal D$ is a flag-transitive $4$–$(q+1,k,\lambda)$ design with $10\ge\lambda\ge5$, $q+1>k>4$, and $G=PSL(2,q)$ a simple automorphism group, then, up to isomorphism, the design is one of $4$–$(24,8,5)$, $4$–$(9,8,5)$, $4$–$(8,6,6)$, $4$–$(10,9,6)$, $4$–$(9,6,10)$, $4$–$(9,7,10)$, $4$–$(12,11,8)$, or $4$–$(14,13,10)$, with block stabilizers $D_8$, $E_8\rtimes C_7$, $D_6$, $E_9\rtimes C_4$, $PSL(2,2)$, $D_{14}$, $E_{11}\rtimes C_5$, or $E_{13}\rtimes C_6$, respectively, except for two undecided cases, $(PSL(2,761),E_{761}\rtimes C_{380},S_4,24,7)$ and $(PSL(2,512),E_{512}\rtimes C_{511},D_{18},18,8)$. Theorem 1.2 rules out all such designs for $\lambda>10$ when $G_B$ is one of $A_4$, $S_4$, $A_5$, $PGL(2,q_0)$ with $q_0^g=q$ and $g>1$ even, or $PSL(2,q_0)$.

Load-bearing premise

The classification depends on the correctness of the computer-assisted eliminations in Lemmas 3.1, 3.4, and 3.5, which are stated without code, logs, or certificates; if any one elimination is wrong, the list of eight designs is incomplete.

Editorial extensions

If this is right

For $5\le\lambda\le10$, the classification is complete up to two explicitly named undecided cases; no other parameter sets are possible.
The two undecided cases are $(PSL(2,761),E_{761}\rtimes C_{380},S_4,24,7)$ and $(PSL(2,512),E_{512}\rtimes C_{511},D_{18},18,8)$; settling them decides whether the list of eight designs is the whole truth.
For $\lambda>10$, designs with block stabilizer $A_4$, $S_4$, $A_5$, $PGL(2,q_0)$ with $q_0^g=q$ and $g>1$ even, or $PSL(2,q_0)$ do not exist.
The eight surviving designs are explicitly identified or constructed, including a $4$–$(24,8,5)$ design, $4$–$(9,8,5)$, $4$–$(8,6,6)$, $4$–$(10,9,6)$, $4$–$(9,6,10)$, $4$–$(9,7,10)$, $4$–$(12,11,8)$, and $4$–$(14,13,10)$.

Reading between the lines

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

The two undecided parameter sets, with roughly $9.2\times10^6$ and $7.5\times10^6$ potential blocks, are exactly the cases where block-count enumeration is infeasible; a non-enumerative arithmetic argument is what would settle them.
The same divisibility-and-orbit method could be applied to flag-transitive $4$-designs with other rank-one simple groups of Lie type, using their subgroup classifications in place of the twelve types for $PSL(2,q)$.
A practical consequence is that any future search for highly symmetric $4$-designs with $PSL(2,q)$ in this $\lambda$ range can restrict attention to the eight listed designs plus the two open cases, rather than rescanning all possible parameters.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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

Dear —,

Quick take: this paper is a real next step in the Dai–Li program, but the headline classification is conditional on Magma runs the manuscript does not document. I would not desk-reject it; I would also not let the theorem through without a computational appendix or referee-verified scripts.

What is new: [9,10] settled λ=3,4; this covers 5≤λ≤10 for flag-transitive 4-designs with PSL(2,q), listing eight isomorphism types and openly leaving two undecided cases. The counting identities in Lemma 2.1 are correctly derived, the subgroup/orbit data from [10,13] is legitimate external support, and the explicit constructions for the small designs are a concrete plus. For a continuation paper it is a useful step, though not a new method.

Where it is soft: the non-existence rulings for Table 1 rows 1–6 and 8–10 and for the eliminated sets in Lemmas 3.4–3.5 are asserted with 'Using Magma...' and no code, logs, or certificates. In a classification theorem those exclusions are load-bearing; each thrown-away row is a potential missing design. The two undecided cases are admitted, so the theorem is already conditional; the undocumented eliminations are the part that could silently change the list.

The typos are also concentrated exactly where the proof depends on computation. Lemma 3.6 gives Gx=E761⋊C380 for PSL(2,512), but 761 does not divide the group order; Lemma 3.3 says the blocks of size 6 produce a 4-(9,8,5); Lemma 3.4 has a 2-adic divisibility line whose exponent does not match the displayed equation. Individually minor, collectively corrosive.

One scope note: Theorem 1.2 is a conditional nonexistence statement for five stabilizer types with λ>10, not a full λ>10 classification; the cyclic/dihedral/elementary abelian cases are only analyzed for λ≤10. That is fine if read literally, but the abstract invites over-reading.

I agree with the reader's conditional verdict. No circular reasoning that I can see—parameters are checked against external lists and designs are built explicitly—but the evidence is not yet reproducible. The audience is the flag-transitive design classification community; for them this result matters. Recommendation: send to a serious referee with a request to rerun the Magma checks, and require the scripts. I would not cite it as settled until then.

Referee Report

3 major / 5 minor

Summary. The paper classifies flag-transitive 4-(q+1,k,λ) designs with automorphism group G=PSL(2,q), under the assumptions q+1>k>4 and λ≥5. Theorem 1.1 states that for 10≥λ≥5, up to isomorphism the design must be one of eight listed types, with two explicitly undecided exceptions; Theorem 1.2 rules out λ>10 for stabilizers of types A4, S4, A5, PGL(2,q0) and PSL(2,q0). The proofs use a counting identity (Lemma 2.1) and a case analysis over the twelve possible subgroup types of PSL(2,q), supported by Magma and GAP/DESIGN computations for orbit-length checks and design constructions.

Significance. If the classification is correct, it is a natural continuation of the λ=3 and λ=4 classifications and provides a concrete finite list of parameter sets together with explicit constructions for most of them. The counting identities in Lemma 2.1 are correctly derived, and the paper honestly names two undecided cases rather than overclaiming completeness. Its principal scientific value, however, depends on the correctness and reproducibility of the computer eliminations, which are currently not documented.

major comments (3)

[§3, Lemma 3.1 and Table 1; also Lemmas 3.4–3.5] The non-existence conclusions for Table 1 rows 1–6 and 8–10, and for the eliminated cases in Lemmas 3.4 and 3.5, rely on undocumented Magma computations. Lemma 3.1 states 'Using Magma[14] we can rule out all cases except case 7', and Lemmas 3.4–3.5 invoke commands such as 'Subgroups(G:OrderEqual:=b)' without giving the actual input groups, commands, or output logs. Each eliminated candidate could in principle support a 4-design, and one wrong ruling would make the list in Theorem 1.1 incomplete. The manuscript should provide the full Magma scripts and output, or an independent certificate of non-existence for each eliminated case.
[§3, Lemma 3.4] The 2-adic argument in Lemma 3.4 is not correctly written. In the subcase (λ,n,|GB|,k)=(6,1,2c,2c), the displayed equation '3 2f−1 / c = 4c2−12c+11' should read 3(2^f−1)/c = 4c^2−12c+11, and the later identity should be 2^{f−1}−1 = (4m+1)(48m+13)(48m+11), without the extra '+1'. More importantly, the claimed contradiction '2^14 | ((4m+1)(48m+13)(48m+11)+1)' is only the statement that product+1 is a sufficiently large power of 2, which is automatic from 2^{f−1}=product+1 and is not contradictory by itself; an additional modular or growth argument is needed. In the subcase (6,1,c,c), the formula 2^f−1 = l(6l−1)(3l−1)+1 for c=6l+1 does not agree with the preceding equation. Since Lemma 3.4 feeds directly into Lemma 3.6 and Theorem 1.1, this is a load-bearing gap.
[§3, Lemma 3.3 and Lemma 3.6] The notation for the design D1 is inconsistent. Lemma 3.3 states that D is a 4-(9,6,10) design with G_B=PSL(2,2), denoted D1, but its proof says that the construction 'returns a 4-(9,8,5) design D1'. Lemma 3.6 then says that the 4-(9,6,10) design in Table 2 case 3 is isomorphic to D1. The statement and proof must be aligned so the reader can tell whether D1 is the 4-(9,6,10) design or the 4-(9,8,5) design.

minor comments (5)

[§3, Lemma 3.6, case 2] The point stabilizer for q=512 is printed as 'E761⋊C380'; from |G_x|=261632=512·511 it should be E512⋊C511, as stated in the abstract and Theorem 1.1.
[Table 2, case 6] The notation '1 2,82' is not explained; from the context it presumably means orbit lengths 1^2 and 8^2, but this should be made explicit.
[§3, Lemma 3.1, Table 1] The column labelled 'position' should be explained; as typeset it is not self-contained and appears to refer to PrimitiveGroup identifiers from GAP/Magma without giving the group library or command.
[§3, Lemma 3.6] The sentence 'For the remaining cases, the same as before, we can deal with them by the same method' does not specify the computations used for Table 2 rows 4–12; for reproducibility, the construction data and verification commands should be supplied.
[Abstract and throughout] The manuscript needs a careful proofread: the abstract says 'Depend on the fact' instead of 'Depending on the fact', and there are several missing parentheses in displayed formulas in Lemma 3.4.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the classification follows from standard design equations, an external subgroup census, and explicit computations; Magma rulings are computational checks, not fitted predictions.

full rationale

The derivation chain is Lemma 2.1 (divisibility and parameter equations) followed by the subgroup classification from [10,13], then case-by-case elimination and construction. Lemma 2.1 is obtained from the standard design identity and the 2-transitivity of flag-transitive automorphism groups cited from [11,12]; it is not defined in terms of the target classification. The subgroup list of twelve kinds is attributed to the external census of Huber and to Dai--Li, not to the present paper's own results. The Magma computations that rule out candidate parameters are explicit computational checks on parameter sets already derived from the necessary conditions; they do not feed a fitted quantity back into the conclusion. The undecided cases are explicitly stated, and the two exceptional parameter sets are not hidden as predictions. Several apparent typos exist (e.g., the '4-(9,8,5)' label inside Lemma 3.3 and the misprinted point stabilizer for q=512), and the absence of Magma code or logs is a reproducibility concern, but these are correctness/verifiability issues, not circularity. I could not exhibit any equation or construction in the paper that is equivalent to its own input by definition, so the circularity score is 0.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The central claim rests on external subgroup classification [10,13], on the transitivity consequence of flag-transitivity [11], and on unshown Magma computations. No free parameters or invented entities are introduced.

assumptions (3)

domain assumption A 2-transitive, point-primitive, block-transitive action lemma for flag-transitive 4-designs (from [11]).
Used in Lemma 2.1 to apply standard orbit-counting formulas; the paper does not reprove Huber's result.
domain assumption G_B must be one of the twelve subgroup types of PSL(2,q) with the stated orbit lengths (from [10,13]).
The entire case split in Section 3 relies on this classification; the orbit-length data are quoted, not reproduced.
ad hoc to paper Magma and GAP/DESIGN computations are correct.
The eliminations in Lemma 3.1 and Lemmas 3.4-3.5 are reported without code or logs, so correctness is assumed.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Flag-transitive $4$-designs and $PSL(2,q)$ groups." pith.science (2026). https://pith.science/paper/BMY4EK3K

@misc{pith2026190800760,
  author       = {Pith},
  title        = {Pith review of: Flag-transitive $4$-designs and $PSL(2,q)$ groups},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BMY4EK3K}},
  note         = {Machine review of arXiv:1908.00760}
}

abstract

This paper considers flag-transitive $4$-$(q+1,k,\lambda)$ designs with $\lambda\geq5$ and $q+1>k>4$. Let the automorphism group of a design $\cal D$ be a simple group $G=PSL(2,q)$. Depend on the fact that the setwise stabilizer $G_B$ must be one of twelve kinds of subgroups, up to isomorphism we get the following two results. (i) If $10\geq \lambda \geq 5$, then except $(G,G_x,G_B,k,\lambda)=(PSL(2,761),{E_{761}}\rtimes {C_{380}},S_4,24,7)$ or $(PSL(2,512),{E_{512}}\rtimes {C_{511}},{D_{18}},18,8)$ undecided, $\cal D$ is a $4$-$(24,8,5)$, $4$-$(9,8,5)$, $4$-$(8,6,6)$, $4$-$(10,9,6)$, $4$-$(9,6,10)$, $4$-$(9,7,10)$, $4$-$(12,11,8)$ or $4$-$(14,13,10)$ design with $G_B=D_8$, ${E_8}\rtimes {C_7}$, $D_6$, ${E_9}\rtimes {C_4}$, $PSL(2,2)$, $D_{14}$, ${E_{11}}\rtimes {C_{5}}$ or ${E_{13}}\rtimes {C_6}$ respectively. (ii) If $\lambda>10$, ${G_B}=A_4$, $S_4$, $A_5$, $PGL(2,q_0)$($g>1$ even) or $PSL(2,q_0)$, where ${q_0}^g=q$, then there is no such design.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

15 extracted references · 15 canonical work pages

[14]

Bosma, J

W. Bosma, J. Cannon, C. Playoust, The MAGMA algebra system I : The user language. J. Symbolic Comput. 24:235-265(1997)

work page 1997
[1]

Delandtsheer, Flag-transitive ﬁnite simple groups

A. Delandtsheer, Flag-transitive ﬁnite simple groups. Arch. Mat h. 47:395-400(1986)

work page 1986
[2]

Alavi, M

S.H. Alavi, M. Bayat, A. Daneshkhah, Symmetric designs admitting ﬂag-transitive and point-primitive automorphism groups associate d to two dimensional projective special groups. Des. Codes Crypto gr. 79:337-351(2016)

work page 2016
[3]

Zhan, S.L

X.Q. Zhan, S.L. Zhou, Non-symmetric 2-designs admitting a two- dimensional projective linear group. Des. Codes Cryptogr. 86:276 5- 2773(2018)

work page 2018
[4]

Zhan, S.Y

X.Q. Zhan, S.Y. Ding, S.Y. Bai, Flag-transitive 2 designs from P SL(2, q ) with block size 4. Des. Codes Cryptogr. (2019). https://doi.org/10.1007/s10623-019-00648-3

work page doi:10.1007/s10623-019-00648-3 2019
[5]

Cusack, S.W

C.A. Cusack, S.W. Graham, D.L. Kreher, Large sets of 3-designs from P SL(2, q ), with block sizes 4 and 5. J. Comb. Des. 3:147-160(1995)

work page 1995
[6]

Keranen, D.L

M.S. Keranen, D.L. Kreher, Shiue P.J.-S., Quadruple systems of th e projective special linear group P SL(2, q ), q ≡ 1( mod 4). J. Comb. Des. 11:339-351(2003)

work page 2003
[7]

Keranen, D.L

M.S. Keranen, D.L. Kreher, 3-designs of P SL(2, 2n) with block sizes 4 and 5. J. Comb. Des. 12:103-111(2004)

work page 2004

Show all 15 references

[8]

Liu, J.X

W.J. Liu, J.X. Tang, Y.X. Wu, Some new 3-designs from P SL(2, q ) with q ≡ 1( mod 4). Sci. China Math. 55:1901-1911(2012)

2012
[9]

S. J. Dai, S. Z. Li, Flag-transitive 4-( v, k, 4) designs and P SL(2, q ) groups. Utilitas Math. 105:3-11(2017)

2017
[10]

S. J. Dai, S. Z. Li, Flag-transitive 4-( v, k, 3) designs and P SL(2, q ) groups. Appl. Math. Comput. 332:167-171(2018)

2018
[11]

Huber, Flag-Transitive Steiner Designs, in: Birkh¨ auser Basel, Berlin, Boston, 2009

M. Huber, Flag-Transitive Steiner Designs, in: Birkh¨ auser Basel, Berlin, Boston, 2009. 8

2009
[12]

C. J. Colbourn, J. H. Dinitz, Handbook of Combinatorial Designs , CRC press, Boca Raton, FL, 2007

2007
[13]

Huber, A census of highly symmetric combinatorial designs

M. Huber, A census of highly symmetric combinatorial designs. J . Algebra Comb. 26:453-476(2007)

2007
[15]

L. H. Soicher, The DESIGN Package for GAP, Version 1.6. (2011 ). http://designtheory.org/software/gap-design/ 9

2011

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [14]

Bosma, J

W. Bosma, J. Cannon, C. Playoust, The MAGMA algebra system I : The user language. J. Symbolic Comput. 24:235-265(1997)

work page 1997

[2] [1]

Delandtsheer, Flag-transitive ﬁnite simple groups

A. Delandtsheer, Flag-transitive ﬁnite simple groups. Arch. Mat h. 47:395-400(1986)

work page 1986

[3] [2]

Alavi, M

S.H. Alavi, M. Bayat, A. Daneshkhah, Symmetric designs admitting ﬂag-transitive and point-primitive automorphism groups associate d to two dimensional projective special groups. Des. Codes Crypto gr. 79:337-351(2016)

work page 2016

[4] [3]

Zhan, S.L

X.Q. Zhan, S.L. Zhou, Non-symmetric 2-designs admitting a two- dimensional projective linear group. Des. Codes Cryptogr. 86:276 5- 2773(2018)

work page 2018

[5] [4]

Zhan, S.Y

X.Q. Zhan, S.Y. Ding, S.Y. Bai, Flag-transitive 2 designs from P SL(2, q ) with block size 4. Des. Codes Cryptogr. (2019). https://doi.org/10.1007/s10623-019-00648-3

work page doi:10.1007/s10623-019-00648-3 2019

[6] [5]

Cusack, S.W

C.A. Cusack, S.W. Graham, D.L. Kreher, Large sets of 3-designs from P SL(2, q ), with block sizes 4 and 5. J. Comb. Des. 3:147-160(1995)

work page 1995

[7] [6]

Keranen, D.L

M.S. Keranen, D.L. Kreher, Shiue P.J.-S., Quadruple systems of th e projective special linear group P SL(2, q ), q ≡ 1( mod 4). J. Comb. Des. 11:339-351(2003)

work page 2003

[8] [7]

Keranen, D.L

M.S. Keranen, D.L. Kreher, 3-designs of P SL(2, 2n) with block sizes 4 and 5. J. Comb. Des. 12:103-111(2004)

work page 2004