{"id":"cf17363f-3fb1-4036-8360-096b1eefe805","arxiv_id":"2607.24042","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A design space taxonomy for quantum circuit visualizations, built from 194 analyzed examples, organizes design choices into view, component, management, narrative, and interaction levels.","lead":"Researchers analyzed 182 static and 12 interactive pictures of quantum programs and organized them into a structured 'design space' describing how such diagrams are drawn, from simple qubit lines to on-chip heatmaps and animations. The goal is to give tool builders a shared vocabulary so future quantum-circuit visualization tools can work across IBM, Google, and other platforms.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The design space's completeness rests on a single-coder open-coding process and a self-defined saturation criterion; without an independent coding pass, the taxonomy may reflect one analyst's lens rather than the range of existing visualizations.","rationale":"The reader's weakest assumption is on target. I would sharpen it: corpus representativeness is a stated limitation (§7 describes the corpus as a snapshot), but the central claim's validity also depends on whether the coding process could have been performed reproducibly. The authors report open-coding and three review rounds, which is good-faith qualitative practice, but with a single coder there is no way to separate category structure inherent in the visualizations from categories imposed by the analyst's prior experience. The saturation rule is defined in terms of new tags produced by the same analyst, so it is not an external check. A concrete inter-rater/holdout test would settle whether the taxonomy is robust. Because the paper is otherwise transparent, provides a gallery, and appropriately frames the design space as a first step, the conditional accept remains appropriate; I would not reject or upgrade before the test.","tokens_in":19671,"tokens_out":5290,"duration_ms":50308,"concrete_test":"Conduct an independent coding exercise: hold out 30 cases (stratified by source: tutorial, paper, presentation; 24 static + 6 interactive) from the corpus before the final codebook is applied; have two coders with no prior exposure to the design space independently label these cases using the published codebook, and compute Cohen's kappa per design-space level and the proportion of codes that do not fit existing categories. If kappa falls below 0.6 for any level, or if the coders' new tags cannot be mapped to the published taxonomy, the design space's reliability and saturation claims fail. A pass would not prove comprehensiveness but would show the taxonomy is not just an artifact of a single coder's lens.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central artifact is the design space itself. For it to characterize the range of existing quantum circuit visualizations, the 194 cases must be coded consistently and the category set must be saturated. The corpus-sampling caveat is acknowledged in §3.1 ('we do not claim a comprehensive corpus') and §7 ('a snapshot'), but the coding reliability issue is more load-bearing: every case was coded by the authors, and the three-round consistency pass in §3.2 can only eliminate internal inconsistencies, not systematic blind spots. Saturation was declared when 'no new open-coding tags' appeared; because the same person who generated the tags also judged saturation, the stopping rule conflates coder convergence with theoretical saturation. This matters concretely because several categories are interpretive judgments about intent: 'Abstraction' (programming/presentation/illustration), 'Layout' (traditional/on-machine/node-edge), 'Dynamicity' of bit lines, and 'Progression' all require inferring the creator's purpose. A second coder applying the final codebook could draw these boundaries differently. Since the paper's contribution is the taxonomy itself, unstable coding undermines the central claim even if the corpus were perfect.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a design space for quantum circuit visualizations by curating and open-coding 182 static and 12 interactive cases drawn from vendor tutorials/documentation, conferences, research publications, and prior systems. The resulting design space is organized into view-, component-, management-, narrative-, and interaction-level choices, and the authors illustrate how the scheme captures representative examples, relate it to established visualization principles (overview+detail, view consistency, collaborative visualization), and propose future directions toward a platform-agnostic grammar. The paper explicitly frames the corpus as a curated snapshot and restricts scope to gate-based quantum computing.","tokens_in":20025,"tokens_out":3763,"duration_ms":40516,"significance":"If the taxonomy is stable, this would be one of the first systematic, cross-tool design spaces for quantum circuit visualization, with clear utility for future grammar-based systems and for communication between visualization and quantum-computing communities. The paper is transparent about its scope, provides a public design gallery, and the taxonomy is grounded in an external corpus rather than derived from a fitted model, so the circularity concern raised in the stress-test note does not materially apply. The primary risk is methodological: the design space is the central contribution, and its reliability rests on a single-coder open-coding procedure with a self-defined saturation criterion.","major_comments":[{"comment":"Coding reliability is load-bearing for the central claim, but the manuscript reports only the authors' three internal review rounds. No inter-rater reliability statistic, independent second coder, or audit of the final codebook is provided. Several categories are interpretive judgments—e.g., 'Abstraction' (programming/presentation/illustration) in §4.2, bit-line 'Dynamicity' in §4.3, and 'Progression' in §4.2—so a second coder applying the same codebook could plausibly draw different boundaries. Since the design space is the paper's contribution, the absence of any reliability evidence leaves the taxonomy potentially reflecting one analyst's lens. The paper should either report an independent coding pass (even on a subset) with agreement measures, or be revised to present the result as an author-curated design vocabulary rather than a validated design space.","section":"§3.2"},{"comment":"The saturation criterion as stated is internally circular and therefore does not establish completeness. The authors stopped collecting cases 'as we reached a point where we were not adding any new open-coding tags,' but the same people who generated the tags also judged when no new tags appeared. Combined with the acknowledged convenience sampling (vendor tutorials, specific conferences, and Nature journals after January 2025), this stopping rule conflates coder convergence with theoretical saturation. The §7 limitation ('a snapshot based on the corpus curated by us') is welcome, but the abstract and RQ1/RQ2 frame the contribution as 'a design space' without that qualifier. A concrete remedy would be to validate saturation with a held-out set coded blind, or to explicitly downgrade the claim to 'a design vocabulary derived from a convenience sample' throughout the title/abstract.","section":"§3.1"}],"minor_comments":[{"comment":"The third bullet says the design space encompasses 'four aspects (view, information, narrative, and interaction),' but Figure 6 and Section 4 describe five levels: view, component, management, narrative, and interaction. Please align the wording.","section":"§1 contributions"},{"comment":"Typo: 'proivde' should be 'provide.'","section":"§7"},{"comment":"Typo: 'techinques' should be 'techniques.'","section":"§4.4"},{"comment":"The labels contain the literal text 'uni27E9' instead of rendered angle brackets (⟩). This is likely a rendering/encoding artifact and should be fixed in the camera-ready version.","section":"Figure 2"},{"comment":"The statement that the full codebook is in the Supplementary Material is helpful, but the main text would benefit from a short example of the code consolidation process (e.g., the 'Format' key split described) to make the methodology more self-contained.","section":"§3.2"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is substantively within TVCG's scope and the taxonomy appears useful. My concern is not with the authors' honesty—the limitations are acknowledged—but with the evidentiary weight placed on the coding methodology. The fix is feasible: add a reliability pass or reframe the central claim as a survey. I saw no problematic citation or novelty issue beyond the incidental inclusion of the first author's own Patoka system, which is appropriately labeled and is only one of twelve interactive cases."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a real contribution, not a reframing. Kim and Battle have done the first systematic design space for quantum circuit visualizations I know of, based on 182 static and 12 interactive cases, and organized it into five levels: view, component, management, narrative, and interaction. The taxonomy is descriptive, carefully linked to examples (with a gallery), and grounded in prior visualization design-space literature. The authors also correctly identify the constraints that make quantum circuits different from classical circuits and from generic spatiotemporal visualizations: qubits don't merge, connectivity is fixed, time competes with 2D chip layout for position channels, and logical-to-physical transpilation creates an extra mapping problem. That is genuinely useful scaffolding for anyone building or comparing QC visualization tools.\n\nWhat I liked: the separation of layout (traditional vs. on-machine vs. hybrid) from format (circuit vs. heatmap) is clean and does real work; the treatment of narrative elements (annotations, attachments, program info) goes beyond typical circuit-diagram surveys; and the limitation statement is honest — they call the corpus a snapshot, not a set of 'good' designs. The Patoka self-citation is appropriate; it is one of the 12 interactive cases and illustrates the animation format. No circularity concern there.\n\nThe soft spots are the ones the stress test flags. The taxonomy is built by the authors' own open-coding, and the saturation criterion ('no new open-coding tags') is applied by the same person who generated the tags. There is no inter-rater reliability pass, so the completeness of the category set is unverified. Several categories are interpretive (abstraction, dynamicity of bit lines, progression), and a second coder might draw those boundaries differently. That does not kill the paper — design-space surveys commonly proceed this way, and the authors are transparent about scope — but it means the right reading is 'a well-organized, empirically grounded starting point,' not 'a proven exhaustive enumeration.' The corpus is also a convenience sample (vendor tutorials, select venues, Nature after Jan 2025), which they acknowledge.\n\nI'd send this to peer review. It is already accepted at TVCG per the header, and that is the right venue. The paper deserves a serious referee: the contribution is new, the writing is clear, and the claims are mostly calibrated. My main referee request would be either an independent coding pass or a more explicit account of what would falsify the category set. I would cite this in the next year as the default reference for QC visualization design choices.","headline":"A genuinely new, well-organized design space for quantum circuit visualizations; the categories are credible and useful, though single-coder open coding leaves completeness unproven.","tokens_in":20375,"tokens_out":2767,"would_cite":true,"duration_ms":29233,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper establishes a design space for quantum circuit visualizations, organizing the diverse encoding choices found across 194 existing examples into five levels: view, component, management, narrative, and interaction.","keywords":["quantum circuit visualization","design space","quantum computing","visualization grammar","open coding","interaction design","narrative visualization","spatiotemporal visualization"],"falsifier":"Survey a fresh set of 50 quantum circuit visualizations from sources outside the original corpus (for example, recent research papers from different venues and documentation from other hardware providers) and open-code them using the paper's taxonomy; if new codes emerge that do not fit any of the five levels, the design space is incomplete. A single clear counterexample — a mainstream quantum circuit visualization whose essential encoding choices cannot be expressed within the design space — would also refute the claim of a comprehensive characterization.","tokens_in":19606,"feed_emoji":"⚛️","tokens_out":5210,"duration_ms":44176,"temperature":0.7,"pith_summary":"Quantum circuits — the programs run on gate-based quantum computers — are drawn in many different ways, from traditional circuit diagrams to on-chip heatmaps to animated layouts, and each quantum computing platform tends to produce its own idiosyncratic rendering. This paper attempts to bring order to that variety: by analyzing 182 static and 12 interactive circuit visualizations, the authors derive a design space that describes every observed visualization as a combination of choices at five levels (view, component, management, narrative, interaction). The payoff of a correct design space is a shared vocabulary and a foundation for a platform-agnostic visualization grammar, so that techniques developed for one toolkit can be reused across all of them. A sympathetic reader would see this as a necessary first step toward treating quantum circuit visualization as a first-class, cross-platform concern rather than an afterthought of hardware-specific tools.","feed_headline":"194 quantum circuit visualizations yield one design space","feed_subtitle":"Five choice levels — view, component, management, narrative, interaction — point the way to platform-agnostic tools.","key_machinery":"The central object is the design space itself: a five-level taxonomy — view, component, management, narrative, interaction — of the choices a creator faces when making a quantum circuit visualization. It is built from open-coding of 194 examples and includes cross-cutting dimensions such as layout (traditional vs. on-machine) and format (circuit vs. heatmap). The design space does not prescribe good designs; it organizes observed ones, with the aim of giving practitioners a vocabulary and a target for future grammar-based tools.","core_discovery":"The central claim is that the range of existing quantum circuit visualizations can be characterized by a structured design space, derived from open-coding a corpus of 182 static and 12 interactive examples collected from tutorials, documentation, research publications, and prior systems. The design space organizes design choices into five groups: view-level decisions (abstraction, composition, layout, format), component-level representations (qubits, gates, measurement, connectivity), management techniques (selection, encoding, data transformation, space management), narrative elements (titles, annotations, emphasis, attachments), and interactions (exploration and composition). The authors a","pith_inferences":["This design space could be operationalized as a declarative specification language: a user would describe a quantum circuit visualization as a combination of the five choice levels, and a renderer would generate the view independently of the underlying quantum platform.","The five-level structure may generalize to other program-visualization domains with strict spatial and temporal constraints, such as classical circuit layout or dataflow debugging.","A concrete test of comprehensiveness would be to have independent coders apply the taxonomy to a fresh set of circuit visualizations and measure inter-rater reliability; the current analysis did not include such a check.","Because the corpus is static-heavy and interaction cases were drawn from a small set of tools, the interaction level is likely the least mature part of the design space and may need expansion as more interactive systems appear."],"forward_implications":["If the design space is accurate, quantum computing practitioners gain a shared vocabulary for describing and comparing circuit visualizations across tools and papers.","It provides a concrete foundation for building a platform-agnostic visualization grammar, so that encodings and interactions developed in one system can be specified and reused in another.","The taxonomy highlights underexplored areas — especially interactive and narrative features — that most current toolkits omit, signalling where future systems could add value.","The analysis connects quantum circuit visualization to established visualization principles (overview+detail, multi-view consistency, collaborative visualization), making research on other spatiotemporal domains transferable."],"fun_headline_variants":["Map of quantum circuit visualization design choices","194 visualizations reveal quantum circuit design space","Quantum circuit viz design space in five choice groups","How quantum circuit visualizations are designed: a map","Structured design space for quantum circuit visualizations"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the 194 visualizations the authors collected are representative of the full range of quantum circuit visualizations — and that the two authors' open-coding, without inter-rater reliability checks, correctly captured the choices in them.","fun_headline_variants_meta":{"raw":{"variants":["Map of quantum circuit visualization design choices","194 visualizations reveal quantum circuit design space","Quantum circuit viz design space in five choice groups","How quantum circuit visualizations are designed: a map","Structured design space for quantum circuit visualizations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000395,"raw_usage":{"total_tokens":1861,"prompt_tokens":648,"completion_tokens":1213,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":392,"completion_tokens_details":{"reasoning_tokens":1144}},"tokens_in":392,"tokens_out":1213,"duration_ms":9641,"temperature":1.0,"reasoning_tokens":1144,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T23:11:09.610695+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Survey a fresh set of 50 quantum circuit visualizations from sources outside the original corpus (for example, recent research papers from different venues and documentation from other hardware providers) and open-code them using the paper's taxonomy; if new codes emerge that do not fit any of the five levels, the design space is incomplete. A single clear counterexample — a mainstream quantum circuit visualization whose essential encoding choices cannot be expressed within the design space — would also refute the claim of a comprehensive characterization.","supporting_citations":[],"review_version":1}