{"id":"19881fe0-d3fb-4cd6-9b60-5749b253dfc5","arxiv_id":"2507.19376","paper_version":1,"verdict":"UNVERDICTED","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A position paper outlines the Archiverse framework for VR-based cultural heritage reconstruction and reports the Volterra theater use case.","lead":"This paper is a panel overview describing a five-step workflow for turning scanned ruins into immersive VR reconstructions, demonstrated on the Roman theater in Volterra. It argues that such transdisciplinary VR projects improve how researchers and the public understand lost architecture, but provides no user study or evaluation.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claim that the pipeline yields 'accurate' reconstruction is unsupported: the input BIM is admitted to be simplified and variably confident, and no step validates fidelity against physical evidence or user perception.","rationale":"I read the paper as a panel and position statement whose central claim is the efficacy of the five-step Archiverse pipeline for accurate immersive reconstruction. The most load-bearing condition for that claim is the fidelity of the final experience to the lost original. The text itself admits that the BIM model was intentionally simplified and carries varying per-element confidence, yet neither the pipeline nor the discussion addresses how those simplifications affect spatial accuracy, nor does it offer an external check against the physical structure or independent archaeological evidence. Without such a check, 'accurately represents' is an overstatement; with a validation step added, the claim would become testable. The reader's weakest-assumption identification, the validity of the BIM model and the absence of validation against physical structure or user perception, captures the same risk, and I agree with it. I do not see an internal inconsistency in the described workflow itself; the weakness is missing empirical support for the comparative and accuracy claims. I also note that the prose says 'five-step process' while enumerating four listed stages; this could be read as five if 'data acquisition' and 'modeling' are counted separately, so I do not rest the critique on that wording. Given the absence of a user study, accuracy metrics, or released artifacts, the appropriate verdict remains UNVERDICTED rather than ACCEPT or REJECT.","tokens_in":4891,"tokens_out":2460,"duration_ms":29916,"concrete_test":"Independently assess fidelity by measuring deviations between the VR environment's geometry and the source point cloud and anastylosis documentation for each architectural element, reporting per-element error and confidence; if deviations exceed the admitted simplification level or cannot be bounded, the 'accurate representation' claim fails. Separately, run a between-subjects study where participants judge distances, scale, and spatial relations from the VR environment versus annotated 2D drawings, using objective accuracy metrics rather than self-reported preference; only if VR users show measurably better spatial understanding is the 'more accurate understanding' claim supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central assertion in Section 4 is that the five-step Archiverse process produces 'an immersive virtual experience that accurately represents lost architectural heritage.' For that assertion to hold, the reconstructed geometry and narrative must be faithful to the lost original, and this fidelity must be checkable. The paper itself undercuts this condition: Section 1 states that Fuchs' model was 'not a digital replica' and that 'some architectural details were simplified,' and that the reconstruction has 'varying levels of confidence regarding the individual elements.' Section 4's pipeline contains no step for estimating, propagating, or communicating this uncertainty, and no step for validating the finished environment against the physical remains or against independent archaeological evidence. The 'more accurate understanding' claim relative to 2D depictions is likewise asserted rather than demonstrated: no user study measures spatial comprehension in VR versus 2D. Absent such validation, the strongest claim reduces to a workflow description plus a plausibility argument, which is consistent with the UNVERDICTED verdict.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper describes Archiverse, a framework for creating immersive virtual reality (VR) experiences of lost architectural heritage, with the Roman theater in Volterra as its use case. The authors argue that VR/XR can overcome the limitations of static two-dimensional depictions by providing interactive, three-dimensional access to reconstructed spaces. They outline a five-step pipeline from data acquisition and modeling through real-time conversion, immersive environment development, and narrative composition, and they report on a panel and exhibition associated with the project. The paper explicitly acknowledges that the underlying BIM model is not a digital replica, that some details were simplified, and that reconstruction confidence varies across elements, yet Section 4 nevertheless claims that the process yields an environment that 'accurately represents lost architectural heritage.'","tokens_in":5004,"tokens_out":4087,"duration_ms":50395,"significance":"If its central claims were empirically supported, the paper would be a useful practitioner-oriented contribution to virtual heritage, particularly for teams working with BIM-derived VR reconstructions. The manuscript has several strengths: it is grounded in a concrete archaeological use case, it draws on prior published work by the authors (including Fuchs's geometric analyses), it transparently discloses funding and institutional support, and it acknowledges—at least in Section 1—that the reconstruction has varying confidence levels and simplified details. However, the current version is essentially a workflow description supported by plausibility arguments rather than by validation: no user study, no accuracy metric, and no comparison against 2D representations is reported. The significance is therefore potential rather than demonstrated, and the paper would need substantial additional evidence or a reframing of its claims to meet the standard of a research contribution in human-computer interaction.","major_comments":[{"comment":"The central claim that the proposed process produces an environment that 'accurately represents lost architectural heritage' is not supported by the evidence in the paper. Section 1 explicitly states that Fuchs's model 'was not a digital replica,' that 'some architectural details were simplified,' and that 'the project has varying levels of confidence regarding the individual elements.' The pipeline described in Section 4 contains no step for estimating, propagating, or communicating this uncertainty, and no validation of the finished environment against the physical remains or against independent archaeological evidence is reported. As written, the accuracy claim is an assertion, not a demonstrated result; it should either be weakened to a claim about plausible reconstruction or the pipeline should be extended with an explicit validation/uncertainty-communication step.","section":"Section 4 (Discussion and Conclusions), first paragraph"},{"comment":"The paper asserts that static 2D depictions 'provide an incomplete and potentially misleading understanding of spatial relationships' and that interactive 3D simulations give a 'more accurate understanding.' The only cited support for this comparative claim is reference [11], which is an opinion/position chapter rather than empirical evidence, and the paper reports no user study, no spatial-comprehension measure, and no direct comparison between VR and 2D conditions. For a paper submitted in the field of human-computer interaction, this load-bearing comparative claim requires either a reported experiment or at least a careful citation of existing controlled studies on spatial understanding in VR versus 2D.","section":"Section 4, paragraph beginning 'Despite the above-mentioned technological advancements'"},{"comment":"The proposed 'five-step process' is described at a level of abstraction that prevents evaluation or replication. The text lists only four items ('data acquisition and modeling,' 'conversion for real-time environments,' 'development of immersive environment,' 'composition of narrative interactive experience') and gives no specifics on model-conversion criteria, real-time rendering constraints, interaction design, narrative structure, or handoff artifacts between interdisciplinary team members. As a result, the paper's claim that this process is what is 'required' for accurate representation cannot be assessed, and the framework is difficult to apply or test in its current form.","section":"Section 4, 'Developing an immersive virtual experience...'"}],"minor_comments":[{"comment":"The text calls the process 'five-step' but enumerates only four items; either add a fifth step or correct the number to four.","section":"Section 4, first paragraph of the process description"},{"comment":"'Roman theater in Voltera use-case' contains a typo: 'Voltera' should be 'Volterra.'","section":"Section 2, first paragraph"},{"comment":"Figures 3 (graybox prototype) and 4 (immersive visualization) are referenced but not described in the text; a sentence explaining how the point cloud, graybox, and final visualization relate would improve readability.","section":"Section 1 and Section 4, Figures 3 and 4"},{"comment":"The phrase 'highly detailed and objective data' overstates the capabilities of laser scanning and photogrammetry, which involve measurement error, occlusion, and processing decisions; 'highly detailed' would be more accurate than 'objective.'","section":"Section 4, paragraph on laser scanning"},{"comment":"The paper presents the panel and exhibition as part of the project but reports no outcomes of the panel discussion; if the panel is intended as a contribution, a summary of its conclusions or discussion points should be included.","section":"Section 2, 'Discussion Points'"}],"recommendation":"major_revision","confidential_remarks":"This manuscript reads more like a project report or panel position paper than a full research contribution. The central accuracy claim conflicts with the paper's own admission of model simplification and variable confidence, and the missing empirical validation is a substantive gap rather than a cosmetic one. However, the issue is addressable within the manuscript's scope: the authors could reframe the claim as a workflow description with explicitly stated limitations, add existing empirical evidence for VR versus 2D spatial understanding, and provide a more detailed specification of the pipeline. I would therefore treat this as a major-revision case rather than a reject, assuming the authors are willing to substantially revise their claims and evidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: This is a clearly written project report describing a five-step VR reconstruction pipeline applied to the Roman theater in Volterra. It is not a research contribution: the pipeline is standard digital heritage practice, and the Section 4 accuracy claim is asserted rather than demonstrated. The paper's most valuable asset is its honest acknowledgment that the underlying BIM model is simplified and has varying confidence per element, but that honesty undercuts the very claim of 'accurate representation.'\n\nWhat the paper does well: the Volterra case is concretely presented, with figures showing the point cloud, graybox, and immersive visualization. The emphasis on cross-disciplinary collaboration, and the need to convert BIM data for real-time engines, is sensible and will resonate with practitioners. The references to Fuchs' earlier geometric studies give the reconstruction a substantive archaeological basis.\n\nSoft spots: the load-bearing problem is the accuracy claim in Section 4. No user study compares VR against 2D spatial understanding, no fidelity metric connects the finished environment to the physical remains, and the pipeline contains no step for estimating or communicating uncertainty. The five-step description is a restatement of scan-model-convert-visualize-narrate. The citation pattern is mostly self-citations; that is not inherently a flaw when the prior work is the basis, but it supplies no independent grounding. If the paper were reframed as a project report, these would be minor caveats; as a research paper, they are central.\n\nWho this is for: practitioners and researchers in VR/cultural heritage who want a short description of one project's workflow, and a reminder that interdisciplinary collaboration is hard. It does not offer a testable method or a benchmark.\n\nRecommendation: I would not send this to a full peer-reviewed venue; it would be desk rejected for lack of validated contribution. It could be accepted in a workshop or short-paper track as a status report, but it doesn't need a serious referee. The authors should either add empirical validation or explicitly tone down the accuracy claim.","headline":"A clearly written project report whose central accuracy claim is unsupported by the paper's own admitted uncertainties.","tokens_in":5593,"tokens_out":4426,"would_cite":false,"duration_ms":52345,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Immersive VR can faithfully recreate lost heritage buildings only through a staged pipeline from field scanning to narrative experience, the paper argues.","keywords":["Virtual Reality","Cultural Heritage","Digital Reconstruction","Building Information Modeling","Immersive Visualization","Roman Theater","Volterra","Extended Reality"],"falsifier":"A controlled test in which participants judge distances, sightlines, and scale while navigating the VR reconstruction, compared with participants reading annotated 2D plans and photographs of the same theater, would settle the central claim: if the 2D group matches or beats the VR group, the asserted advantage of immersion is not supported.","tokens_in":4699,"feed_emoji":"🏛️","tokens_out":7816,"duration_ms":78158,"temperature":0.7,"pith_summary":"This paper argues that recreating vanished or ruined architecture as an immersive virtual experience is a staged engineering and research process, not a single modeling task. Using the Roman theater in Volterra as the test case, the authors lay out the Archiverse framework: acquire and model source data, convert it for real-time engines, develop the immersive environment, and compose a narrative interactive experience. The framing claim is that static two-dimensional depictions such as plans and photographs are incomplete and potentially misleading, while VR provides a more accurate understanding of spatial relationships, scale, and construction. The authors also stress that the pipeline depends on transdisciplinary collaboration among archaeologists, architects, modelers, developers, and humanities researchers. A reader should care because the paper offers a concrete, reusable recipe for turning ruins into explorable digital heritage.","feed_headline":"Five-step pipeline recreates a lost Roman theater in VR","feed_subtitle":"A panel paper argues 2D drawings mislead; BIM-based VR walkthroughs give a truer sense of scale and space.","key_machinery":"The load-bearing mechanism is the production pipeline itself, anchored on Building Information Modeling (BIM) — digital architectural models used as the geometric source for the virtual environment. The pipeline's stages decide what is measured, what is inferred, how the model is simplified for real-time rendering, and what narrative layer makes the result meaningful to a user. A supporting geometric insight in the Volterra case is the theater's regular composition based on a 21-sided polygon derived from a heptagon, which gives the reconstruction its architectural coherence. The same pipeline is what lets the project claim a shift from static images to interactive spatial understanding.","core_discovery":"The paper's central claim is that an accurate immersive reconstruction of lost architecture emerges only from a deliberate pipeline that starts with heterogeneous source data and ends with a narratively framed interactive environment. Using the Roman theater in Volterra as the use case, the authors describe how laser scans, photogrammetry, and drone imagery are combined with geometric-architectural analysis into a digital building model (BIM), which is then converted for a real-time graphics engine and developed into a walkable VR scene. They contend that 2D depictions such as plans, sections, and photographs, despite their apparent objectivity, are incomplete and potentially misleading for understanding spatial relationships, because they omit the dynamic, embodied experience of scale, lighting, and construction. The reconstruction's fidelity is presented as sufficient for studying how the building was constructed and used, even though individual model elements carry varying levels of confidence.","pith_inferences":["The accuracy claim is testable empirically: comparing spatial-judgment performance in the VR reconstruction against 2D plans of the same theater would tell whether immersion actually delivers the asserted advantage; the paper reports no such test.","Because the BIM model itself carries varying levels of confidence, a natural extension the authors do not develop is to visualize or annotate that uncertainty inside the VR experience.","The same pipeline could generalize to urban-scale heritage sites, not just single buildings, if the data-acquisition and real-time conversion stages scale to larger footprints.","The staged framing implies a need for standard quality metrics at each step, which the paper leaves implicit."],"forward_implications":["A cultural heritage team can follow the same staged recipe to turn survey data of any ruined site into an interactive VR experience, rather than treating each reconstruction as a bespoke effort.","If the claim holds, VR-based study of ruins should supplement or replace 2D drawings and photographs wherever the research question concerns spatial layout, scale, or embodied use.","Because the pipeline ends in a narrative layer, the framework implies that heritage VR is a storytelling medium, not just a geometric model.","The paper's emphasis on real-time graphics constraints implies that hardware and engine limitations, not just archaeological uncertainty, will shape what can be faithfully shown."],"supporting_citations":[{"why":"Documents the original 1950–1953 excavations that identified the theater and its remains.","marker":"[1]"},{"why":"Supplies the geometric-language method for interpreting Roman theater design used to read the Volterra evidence.","marker":"[2]"},{"why":"Provides the geometric and architectural composition study of the Volterra theater on which the BIM model is based.","marker":"[4]"},{"why":"Describes the theater's architectural features and the anastylosis work, grounding the reconstruction's historical claims.","marker":"[5]"},{"why":"Makes the case for co-designing immersive XR systems with interdisciplinary teams, supporting the framework's collaboration stage.","marker":"[7]"},{"why":"Argues virtual heritage must go beyond static images to interactive experience, backing the paper's rejection of 2D depiction.","marker":"[11]"}],"fun_headline_variants":["VR pipeline rebuilds lost Roman theater from scans","Why 2D plans fail: VR walkthroughs revive Volterra's theater","From laser scans to VR: reconstructing ancient sites immersively","BIM to VR: a five-step route to lost architecture","Volterra's theater returns in VR, proving 2D isn't enough"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes that the digital building model of the theater — a model that intentionally simplifies some details and marks other elements as less certain — is still faithful enough to anchor a virtual experience described as accurately representing the lost original.","fun_headline_variants_meta":{"raw":{"variants":["VR pipeline rebuilds lost Roman theater from scans","Why 2D plans fail: VR walkthroughs revive Volterra's theater","From laser scans to VR: reconstructing ancient sites immersively","BIM to VR: a five-step route to lost architecture","Volterra's theater returns in VR, proving 2D isn't enough"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000185,"raw_usage":{"total_tokens":1280,"prompt_tokens":860,"completion_tokens":420,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":476,"completion_tokens_details":{"reasoning_tokens":330}},"tokens_in":476,"tokens_out":420,"duration_ms":5137,"temperature":1.0,"reasoning_tokens":330,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T14:21:15.952999+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A controlled test in which participants judge distances, sightlines, and scale while navigating the VR reconstruction, compared with participants reading annotated 2D plans and photographs of the same theater, would settle the central claim: if the 2D group matches or beats the VR group, the asserted advantage of immersion is not supported.","supporting_citations":[{"cited_title":"Notizie degli Scavi di Antichit` a (s VIII), 114–150 (1955)","cited_arxiv_id":null,"evidence_quote":"Documents the original 1950–1953 excavations that identified the theater and its remains."},{"cited_title":"Nexus Network Journal 21(3), 547–569 (2019)","cited_arxiv_id":null,"evidence_quote":"Supplies the geometric-language method for interpreting Roman theater design used to read the Volterra evidence."},{"cited_title":"Giorgio Bretschneider Rome, Italy (2021)","cited_arxiv_id":null,"evidence_quote":"Provides the geometric and architectural composition study of the Volterra theater on which the BIM model is based."},{"cited_title":"rilevazioni, caratteristiche architet- toniche e problemi di sistemazione","cited_arxiv_id":null,"evidence_quote":"Describes the theater's architectural features and the anastylosis work, grounding the reconstruction's historical claims."},{"cited_title":"In: IFIP Conference on Human-Computer In- teraction","cited_arxiv_id":null,"evidence_quote":"Makes the case for co-designing immersive XR systems with interdisciplinary teams, supporting the framework's collaboration stage."},{"cited_title":"In: Paradata and transparency in virtual heritage, pp","cited_arxiv_id":null,"evidence_quote":"Argues virtual heritage must go beyond static images to interactive experience, backing the paper's rejection of 2D depiction."}],"review_version":1}