{"id":"f37f6d77-1732-4a66-bb1e-efac385cb025","arxiv_id":"2505.04934","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A systematic survey that organizes blockchain cross-chain interoperability techniques into a classification and compares them with industry practice, while identifying open challenges.","lead":"This paper is a survey of technologies that let different blockchains exchange assets and data. It classifies approaches into categories like atomic swaps, sidechains, and light clients, and compares academic research with industry platforms.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'most comprehensive to date' claim rests on a non-auditable corpus; without the full source list and selection criteria, the central contribution cannot be verified.","rationale":"The central claim of the paper is comprehensiveness: 'the most comprehensive overview to date' and 'the most extensive classification to date'. For that claim to hold, the surveyed corpus must be representative of the field. The paper's stated method (Section I.B) is a Google Scholar keyword search (2016-2024) with 'prioritizing highly cited works', followed by unspecified screening and quality assessment, yielding 'over 150 pertinent studies'. This procedure is not auditable: there is no exact source count, no list of included works, no citation threshold, and no inclusion/exclusion rules. The manuscript itself admits 'potential publication bias and the risk of omitting significant studies'. A survey whose differentiator is breadth cannot substantiate that breadth without a reproducible corpus. This is a load-bearing concern because if the corpus is biased or incomplete, the 'most comprehensive' claim fails and the classification may omit established approaches. A secondary but concrete correctness issue is the internal inconsistency in Definition 6 (Section III.D). The formal statement of Weak Atomicity maps Tx_CC.In to L_T and Tx_CC.Out to L_S, while the surrounding text and the informal description map In to the source chain and Out to the target chain. The prose description of the implication is also reversed relative to the formal statement. This is a genuine technical error in the paper's security definition, though it is orthogonal to the comprehensiveness claim. Similarly, the security metric formulas in Section IV (M = Min(M1,M2,M3), M = M1⊕M2, M = Max(M1,M2)) are asserted without justification and the operator '⊕' is undefined; these are unsupported approximations that should be moved to explicit assumptions or removed. Given the reader's CONDITIONAL verdict and the fact that these issues are correctable (release the corpus and selection criteria, fix Definition 6, qualify the security metrics), no change to the verdict is warranted. The stress-test confirms the reader's weakest assumption as the primary load-bearing concern.","tokens_in":1260,"tokens_out":1506,"duration_ms":63144,"concrete_test":"Request the exact list of the >150 analyzed sources and the full selection protocol (query string, date range, citation threshold, screening rules) from the authors. Then independently rerun the same Google Scholar query for 2016-2024, apply the stated 'highly cited' priority, and compute the overlap with the provided list. Specifically: (1) verify every provided source appears in the reconstructed query results; (2) verify that all highly-cited interoperability papers cited in the comparison surveys listed in Table II (e.g., Belchior et al. 2021, Wang et al. 2023, Ren et al. 2023) are present in the corpus; (3) count how many of the top-cited papers in the reconstruction are missing from the provided list. If the reconstruction cannot reproduce the corpus or if significant missing items are found, the 'most comprehensive' claim is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's headline contribution—'the most comprehensive overview to date' and 'the most extensive classification to date' (Section I.C)—depends entirely on the representativeness of its literature corpus. The method section (I.B) describes only a Google Scholar keyword search for 'Blockchain&(Interoperability|Cross-Chain)' over 2016-2024, 'prioritizing highly cited works', followed by 'initial screening and quality assessment' that selected 'over 150 pertinent studies'. No reproducible inclusion/exclusion criteria, citation threshold, screening protocol, or full list of selected sources is provided, and the number 'over 150' is not exact. The survey itself concedes 'potential publication bias and the risk of omitting significant studies' (Section I.B). Since the paper's differentiation from prior surveys (e.g., Belchior et al. 2021, Wang et al. 2023) is precisely its breadth of coverage, an unverifiable corpus makes the 'most comprehensive' assertion an article of faith rather than a demonstrated result. Table II's comparison is also self-assessed: the 'Our Survey' row marks every criterion as fully covered (●), but no external audit or inter-rater protocol is described. A secondary but concrete correctness issue: Definition 6 states Weak Atomicity as 'Tx_CC.In∈L_T ⇒ Tx_CC.Out∈L_S', which maps the input transaction to the target chain and the output to the source chain, contradicting the preceding description where Tx_CC.In is on Chain #1 (S) and Tx_CC.Out on Chain #2 (T). The accompanying prose says the opposite ('Tx_CC.Out appears on L_T only if Tx_CC.In has been already written into L_S'), so the formal definition and its explanation are mutually inconsistent. Additionally, the security metric formulas in Section IV (M = Min/Max/⊕) are introduced without derivation or definition of '⊕'. These issues do not bear directly on comprehensiveness, but they undercut the survey's claimed systematic rigor.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This survey aims to systematize blockchain cross-chain interoperability (CCI). It proposes a three-way classification of verification modes (native, local, external), defines a formal interoperability paradigm and security properties, reviews ten or more technology categories (HTLC, adaptor signatures, notaries, light clients, sidechains, chain relays, rollups, burn-and-mint, hierarchical chains, sharding), analyzes representative industry platforms (Interledger, Cosmos, Polkadot, HyperService, LayerZero, RSK, and several permissioned frameworks), and discusses interdisciplinary avenues such as redactable blockchains, asynchronous consensus, and the metaverse. The paper claims to be 'the most comprehensive overview to date' and 'the most extensive classification to date' based on a Google Scholar search yielding 'over 150 pertinent studies' from 2016 to 2024.","tokens_in":50153,"tokens_out":4611,"duration_ms":43695,"significance":"If the descriptive synthesis is taken as the main contribution, the survey is genuinely useful: it assembles a broad set of mechanisms, compares them across trust models and verification modes, and connects academic protocols with industrial practice. The comparison tables (e.g., Tables V, VII, VIII, IX, XI), the historical timeline in Fig. 6, and the explicit treatment of open challenges are likely to help newcomers and interdisciplinary readers. However, the paper's analytic apparatus is not yet reliable enough to carry the load placed on it: the formal atomicity definition is internally inconsistent, the security-metric formulas in Section IV are asserted without derivation or citation, and the corpus underlying the 'most comprehensive' claim is not auditable. These issues affect the credibility of the paper's central claims and require substantive revision.","major_comments":[{"comment":"The weak-atomicity formula is internally inconsistent with the surrounding prose. The text defines Tx_CC.In as the input affecting the source ledger and Tx_CC.Out as the output affecting the target ledger, and the failure cases in Fig. 9 place Tx_CC.In on Chain #1 and Tx_CC.Out on Chain #2. Yet Definition 6 states 'Tx_CC.In∈L_T ⇒ Tx_CC.Out∈L_S', and the strong-atomicity formula uses the same reversed mapping. As written, the formulae assert that the input transaction appears on the target ledger and the output on the source ledger, contradicting the preceding description. Please correct the formal statements, and the accompanying explanation, so that Tx_CC.In is associated with L_S and Tx_CC.Out with L_T, or explicitly redefine the notation if the labels were intended to be different.","section":"Section III.D, Definition 6"},{"comment":"The security metric M = Min(M1, M2, M3) for external verification, M = M1⊕M2 for local verification, and M = Max(M1, M2) for native verification is asserted without derivation or citation, and the operator ⊕ is never defined. These formulas are used to position technologies in Fig. 11 and to support the claim that the external verifier is the weakest link; without a stated adversary model or composition semantics, the formulas are not supported by the survey's own security definitions (Defs. 4 and 6). Please either derive them from the cited formal framework, replace them with qualitative statements that are referenced, or cite the source that establishes them.","section":"Section IV, paragraph following Fig. 11"},{"comment":"The headline claim of being the most comprehensive overview and classification to date rests on a literature corpus that is not auditable. The method section reports only a Google Scholar keyword search ('Blockchain&(Interoperability|Cross-Chain)', 2016-2024) with 'over 150 pertinent studies' selected after 'initial screening and quality assessment', but it gives no inclusion/exclusion criteria, no citation threshold, no screening protocol, and no list of the selected sources. The paper itself concedes 'potential publication bias and the risk of omitting significant studies.' In addition, Table II rates the 'Our Survey' row as fully covered on every criterion using only the authors' self-assessment. Please make the corpus reproducible (for example, by adding an appendix with the full source list and selection criteria), and temper or precisely qualify the comprehensiveness claim if the corpus cannot be fully enumerated.","section":"Section I.B and Section I.C, Table II"}],"minor_comments":[{"comment":"The corpus is described as 'over 150 high-impact sources' in the abstract and 'over 150 pertinent studies' in Section I.B; an exact count, or at least a consistent figure, would improve precision.","section":"Abstract and Section I.B"},{"comment":"The y-axis label appears garbled as '/glyph1197umber' and should read 'Number'.","section":"Fig. 3"},{"comment":"In the bullet list comparing adaptor signatures with HTLC, 'Mulit-path' should be 'Multi-path'.","section":"Section IV.A.2"},{"comment":"The entries for 'SSPV' under 'Backward Compatibility' and 'No Trusted Setup' read 'Constant', which is not a meaningful rating for those criteria; please re-check the intended symbols or notations.","section":"Table VII"},{"comment":"The proposed asynchronous-consensus interoperability framework is presented only as a sketch without a security or feasibility analysis; either expand it with such a discussion or label it explicitly as a research agenda.","section":"Section VI.B, Fig. 26"}],"recommendation":"major_revision","confidential_remarks":"Several entries in the comparison tables are authored by the current authors (e.g., [135], [174], [178], [205], and [6] in the introductory material). I did not find evidence that the central taxonomy depends on these results, so I treat this as a transparency concern rather than a reason for rejection. A full list of the surveyed corpus in the revision would address both the self-citation concern and the reproducibility issue raised in Major Comment 3."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a genuinely useful survey of blockchain cross-chain interoperability. Its descriptive core is accurate and well-sourced, and the industry platform comparison (permissionless vs. permissioned) plus the sidechain/light-client/relay tables are a solid contribution. The native/local/external verification taxonomy is mostly a relabeling of existing trust-model categories, as the authors themselves concede, but it is clearly explained and does no harm. For a newcomer, this would be a good entry point.\n\nThe problems are real but fixable. The headline claim of being \"the most comprehensive overview to date\" is not auditable: the search procedure is described only loosely, no selection criteria or full source list is given, and the paper itself acknowledges potential publication bias. That makes the claim an article of faith. Table II is also self-assessed, with the \"Our Survey\" row marked as fully covering everything.\n\nThere is also a concrete formal error. Definition 6 states weak atomicity as Tx_CC.In∈L_T ⇒ Tx_CC.Out∈L_S, which swaps the source and target ledgers relative to the prose and the preceding example. The explanation after the definition says the opposite (Tx_CC.Out on L_T only if Tx_CC.In on L_S). This is an internal inconsistency that should be corrected before the paper is used as a reference. The security metric formulas in Section IV (M = Min/Max/⊕) are asserted without derivation and ⊕ is never defined; either they should be justified or dropped.\n\nThe asynchronous-consensus framework in Section VI.B is only a sketch, so it does not add much beyond pointing at an interesting direction. The self-citations used in comparison tables are not load-bearing, so I do not see a circularity problem.\n\nMy overall take: the survey is honest in its limitations, the descriptive content holds up, and the flaws are concentrated in a few places. Fix the definitional error, provide a reproducible corpus or temper the comprehensiveness claim, and either support or remove the security metrics. Then this becomes a dependable reference.\n\nI would send this to peer review, expecting major revision but not rejection. It deserves referee time because a well-corrected version would serve the community. For my own work, I would cite it for its industry/platform coverage, but not for the formal definitions.","headline":"Useful survey with an unverifiable comprehensiveness claim and a flawed formal definition; correctable but needs revision.","tokens_in":50761,"tokens_out":1625,"would_cite":true,"duration_ms":18942,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This survey claims that blockchain interoperability is best understood through a three-way split—native, local, and external verification—and that this split supports the most extensive classification of cross-chain technologies, industry…","keywords":["blockchain interoperability","cross-chain communication","atomic swaps","light clients","sidechains","notary schemes","verification modes","scalability trilemma"],"falsifier":"Reproduce the stated keyword search over the 2016–2024 window, apply the stated 'highly cited' filter, and audit whether every selected source maps into one of the survey's ten technology categories; if a substantial cluster (for example, oracle-based or intent-based cross-chain systems) falls outside the native/local/external verification trichotomy, the 'most comprehensive classification to date' claim is unsupported.","tokens_in":49606,"feed_emoji":"🔗","tokens_out":6956,"duration_ms":69012,"temperature":0.7,"pith_summary":"Blockchains store value and data in silos, and this survey tries to organize the many mechanisms that let different chains exchange assets and information. Its central claim is that every interoperability technology can be classified by where verification happens: natively on the chains themselves, locally among the transacting parties, or externally through intermediaries. Under that trichotomy it catalogs more than ten technology families—atomic swaps, notaries, light clients, sidechains, chain relays, rollups, burn-and-mint protocols, hierarchical designs, sharding, and others—and claims this is the most extensive classification to date. If true, researchers and industry builders get a shared vocabulary for comparing the security, trust, and efficiency trade-offs of cross-chain designs. The survey also formalizes weak and strong atomicity for cross-chain transactions and connects interoperability to scalability, arguing that interoperability is an essential prerequisite for service scalability.","feed_headline":"One taxonomy sorts all cross-chain technologies into three modes","feed_subtitle":"Native, local, and external verification organize atomic swaps, sidechains, light clients, and more into a single map.","key_machinery":"The load-bearing machinery is a verification trichotomy drawn as a triangle: native verification, where the source or target chain's own consensus checks the cross-chain state; external verification, where a third-party validator, notary, or committee vouches for it; and local verification, where the transacting parties check each other directly. The paper adds a generic interoperability paradigm—setup, commit on the source, verify, commit on the target or abort—and a security definition in terms of weak and strong atomicity. The taxonomy assigns every surveyed technology to a point or combination in this triangle, which is what carries the 'most comprehensive classification' claim.","core_discovery":"The paper's systematization, offered as a survey rather than a new protocol, is that blockchain interoperability is not an unstructured grab bag of bridges. It proposes a generic paradigm in which a cross-chain transaction commits on the source chain, is verified, then commits on the target chain or aborts, and it defines security through weak atomicity (the target-side effect appears only if the source-side effect is already stable) and strong atomicity (both sides appear together or not at all). Against this framework it lays a verification-based taxonomy: native verification (e.g., light clients and sharding), external verification (e.g., notaries and validator networks), and local verification (e.g., state-channel-style direct-party checks), with chain relays, rollups, and sidechains combining modes. It then claims that this taxonomy, applied to over 150 sources, yields the most comprehensive classification to date, including at least ten technology types, and that treating interoperability as computational offloading helps balance the blockchain trilemma.","pith_inferences":["Beyond the paper, if the trichotomy is adopted by the field, newer 'chain abstraction' and intent-based cross-chain systems would likely fall under external verification, which would expose their trust assumptions rather than presenting them as a brand-new category.","The paper's open challenge of cross-chain redaction suggests a natural next frontier: rewriting a transaction on one chain should cascade to dependent transactions on other chains while preserving atomicity, a problem the survey only sketches.","A testable extension would be to run the same verification-based taxonomy on protocols published after this survey's 2024 cutoff and check whether any genuinely new family appears or whether the existing ten categories absorb them.","Because the survey's 'most comprehensive' claim rests on a corpus selected without a published inclusion list, a useful follow-up is to release the full annotated source list and criteria, turning the classification into an auditable dataset."],"forward_implications":["Researchers gain a standard way to place a cross-chain protocol: ask where verification happens, then read off the trust assumption and the dominant cost.","Security comparisons become formulaic: external verification is only as strong as its weakest link, while native verification inherits the stronger of the two chains.","The weak and strong atomicity definitions give protocol designers a precise target: a bridge is secure when it satisfies strong atomicity, and only weakly safe when target-side effects can lag behind source-side effects.","The survey's comparison of permissionless and permissioned industry platforms gives practitioners a menu for choosing between asset swap, asset migration, and data-transfer modes.","Interoperability is repositioned as a scalability tool: routing work across chains relaxes the blockchain trilemma instead of merely adding bridging features."],"supporting_citations":[{"why":"Supplies the cross-chain communication protocol model, the impossibility of TTP-free interoperability in asynchronous settings, and the atomicity framing the survey builds on.","marker":"[30]"},{"why":"Provides the earlier Public Connectors / Blockchain of Blockchains / Hybrid classification that the survey positions itself as extending.","marker":"[27]"},{"why":"Provides another baseline taxonomy (chain-based, bridge-based, dApp-based) and interoperability definitions used for comparison.","marker":"[28]"},{"why":"Contributes a performance-evaluation perspective and interoperability definitions that the survey integrates into its framework.","marker":"[29]"},{"why":"Supplies the security-and-privacy systematization and layered security analysis that inform the survey's security and privacy discussions.","marker":"[12]"},{"why":"Gives the Bitcoin Backbone model whose persistence and liveness properties ground the survey's security definitions.","marker":"[82]"},{"why":"Formalizes atomic cross-chain swaps, the basis for the atomic-swap technology section.","marker":"[59]"},{"why":"Provides the original Bitcoin design and the simplified payment verification concept that underlies light clients.","marker":"[1]"},{"why":"Supplies the Cosmos hub-and-zone architecture and IBC protocol used as a representative industry interoperability platform.","marker":"[23]"}],"fun_headline_variants":["Taxonomy maps all cross-chain tech into 3 verification modes","One framework sorts atomic swaps, sidechains, light clients","Most comprehensive cross-chain survey: 150 sources, one taxonomy","Bridge tech sorted: native, external, local verification"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The survey's claim to be the most comprehensive classification depends on the unstated assumption that a keyword search of academic and grey literature from 2016 to 2024, prioritizing highly cited works, produced a representative corpus of over 150 studies; the paper provides no reproducible inclusion criteria or full source list and concedes potential publication bias and the risk of omitting significant studies.","fun_headline_variants_meta":{"raw":{"variants":["Taxonomy maps all cross-chain tech into 3 verification modes","One framework sorts atomic swaps, sidechains, light clients","Most comprehensive cross-chain survey: 150 sources, one taxonomy","Bridge tech sorted: native, external, local verification"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001262,"raw_usage":{"total_tokens":5177,"prompt_tokens":963,"completion_tokens":4214,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":579,"completion_tokens_details":{"reasoning_tokens":4146}},"tokens_in":579,"tokens_out":4214,"duration_ms":28558,"temperature":1.0,"reasoning_tokens":4146,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T23:17:18.618880+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Reproduce the stated keyword search over the 2016–2024 window, apply the stated 'highly cited' filter, and audit whether every selected source maps into one of the survey's ten technology categories; if a substantial cluster (for example, oracle-based or intent-based cross-chain systems) falls outside the native/local/external verification trichotomy, the 'most comprehensive classification to date' claim is unsupported.","supporting_citations":[],"review_version":1}