{"id":"2fdeea37-edba-44e4-b8d9-a009df26fd50","arxiv_id":"2509.01617","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The schema-centric bias in conceptual modelling is a historical artifact, not a necessity, and data pipelines like bCLEARer can restore the information base into conceptual modelling.","lead":"This paper argues that the field of conceptual modelling has unjustifiably separated the data (the information base) from the schema, and that modern tools make it possible to model both together. It offers a historical re-framing and an example pipeline (bCLEARer) to show the separation is a contingent choice, not a fundamental one.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'not fundamental' conclusion rests on bCLEARer as a working counterexample; the paper does not show that bCLEARer genuinely avoids schema/base separation, so the central claim is unverified.","rationale":"The reader's verdict CONDITIONAL is appropriately cautious. I agree that the causal story in Section 5.3 is asserted rather than demonstrated, but I do not think that is the most load-bearing point. The sentence 'we showed that the turn is not fundamental' (Section 10) is supported by two pillars: (i) the historical narrative that the schema turn arose from contingent technological constraints, and (ii) the existence of bCLEARer as a working inclusive schema-and-base pipeline. Of these, (ii) is the true refutation of fundamentality, because a single genuine counterexample would establish that the separation is not necessary, regardless of why the status quo arose. The historical narrative is interesting but would not by itself show non-fundamentality—a practice can be contingent in origin and still be the only viable one. Conversely, if bCLEARer is not a genuine counterexample, the paper's central claim loses its constructive support. The bCLEARer evidence is weak as presented: it is self-reported, high-level, and linked but not analyzed. The paper does not show that the schema and base are integrated under its own Section 3.2 definitions, nor that the claimed benefits are due to integration rather than to pipeline automation generally. This is a concrete, checkable issue. The open-source repositories make a decisive test feasible. If the test shows bCLEARer does keep schema and base as one integrated artifact, the CONDITIONAL verdict could move toward ACCEPT; if not, it would move toward REJECT or UNVERDICTED. For now, CONDITIONAL is right, so I recommend no change.","tokens_in":16472,"tokens_out":7124,"duration_ms":78064,"concrete_test":"Inspect the bCLEARer repositories (e.g., github.com/boro-alpha/bclearer_boson_1_1) and trace a single pipeline run. Determine whether at each stage the 'conceptual model' is stored as one integrated graph containing both schema-level and instance-level nodes with no separate schema-only artifact, or whether the pipeline creates a stand-alone schema (e.g., a relational catalog/ontology file) separate from the base data. If any stage separates schema from base, then bCLEARer does not demonstrate the inclusive style; if no stage separates them, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's conclusion that the schema turn is 'not fundamental' (Section 10) depends on showing that an inclusive schema-and-base approach is actually possible. The only evidence for this is bCLEARer (Section 8). Yet the description is self-reported and underspecified: Section 8.2 says the pipeline is 'automated' and links to repositories, but it never states whether the conceptual schema and information base form a single integrated artifact under the paper's own taxonomy (Section 3.2: separated/aggregated/integrated based on extensional separation of data objects). Section 8.3 attributes benefits (automation, empirical design) to the inclusive style, but these could arise from ordinary data-pipeline engineering regardless of whether schema and base are unified. If bCLEARer in fact constructs a schema, stores it separately from base instances, and then maps between them—as standard ETL does—it is not a counterexample to the schema turn. The historical causality (Section 5.3) is secondary: even if that story is correct, non-fundamentality requires a viable alternative; without a verified alternative, the central claim is unproven.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper argues that the near-universal separation of a conceptual schema from its information base in computer-science conceptual modelling—what it calls the 'schema turn'—is historically contingent rather than fundamental. It develops a hylomorphic modularity framework with separated/aggregated/integrated styles, reads Codd/CODASYL, Mealy/Chen, and ANSI-SPARC/ISO as showing the gradual emergence of schema/base separation, attributes that separation to 1970s technology and greenfield projects, and uses the bCLEARer pipeline as an example of an inclusive schema-and-base alternative. It concludes that modern technology may make the schema turn a temporary evolutionary detour and that the design space for conceptual modelling is wider than currently assumed.","tokens_in":16761,"tokens_out":5754,"duration_ms":61166,"significance":"If the thesis were established, it would challenge a taken-for-granted assumption in database textbooks and open new research directions in pipeline-based conceptual modelling. The paper is commendably concrete in shipping open-source repository links and in grounding its historical narrative in primary documents. However, the central claim currently rests on an interpretive historical causal story and on a self-reported implementation with no independent evaluation; both need substantial strengthening before the non-fundamentality conclusion is warranted.","major_comments":[{"comment":"The conclusion that the schema turn is 'not fundamental' (Section 10) depends on bCLEARer as a working counterexample, but the paper never shows that bCLEARer instantiates the 'integrated' category of the authors' own taxonomy (Section 3.2, based on extensional separation of data objects). Section 8.2 describes an automated pipeline and links to repositories; it does not state whether the conceptual schema and information base form a single extensional artifact or are stored as separate datasets mapped by transformations. Section 8.3 attributes benefits (automation, empirical design) to the inclusive style, but those benefits can accrue to ordinary ETL pipelines regardless of schema/base unification. Without an artifact-level demonstration, the existence proof for an inclusive schema-and-base model is unverified.","section":"Section 8.2/8.3"},{"comment":"The causal claim that the schema turn was 'making a virtue of technical necessity' (Section 5.3) is asserted rather than demonstrated. No contemporaneous source is cited in which the originators justify the schema/base separation by scalability constraints or greenfield project conditions; the ANSI-SPARC/ISO text quoted in Section 5.4 gives 'practical arguments' for the distinction, which the paper interprets as technological, but the quote is generic. Since the 'not fundamental' thesis requires that the separation was caused by removable environmental constraints rather than by a conceptual type/instance necessity, the historical causality needs direct evidence or a counterfactual analysis.","section":"Section 5.3/5.4"},{"comment":"The taxonomy of 'separated/aggregated/integrated' is defined by a loose notion of extensional separation, with 'integrated' explained by analogy to Aristotelian mixis. There is no operational criterion for deciding when a system is 'truly mixed' as opposed to storing schema and data in separate but related structures. This makes the classification of both historical styles and bCLEARer unfalsifiable. A sharper, testable criterion is needed before the inclusive style can be presented as a distinct architectural possibility rather than a relabeled ETL design.","section":"Section 3.2"},{"comment":"The package-application-migration example shows that the physical base data persists through schema replacement (the germline analogy), but persistence of data does not imply that schema and base are conceptually unified in an inclusive model. The proposed two-stage mapping through a common conceptual model can be implemented with a separated schema and base; the paper does not identify a step that requires the inclusive style or show that it is empirically better. The example therefore supports at most the weak claim that base-first thinking is useful in migration, not the strong claim that a schema-and-base modularity style is viable.","section":"Section 7"},{"comment":"The summary states 'we then showed that the turn is not fundamental,' whereas the abstract and introduction frame the conclusion as a possibility ('it is possible that ... could be merely a temporary evolutionary detour'). The paper's evidence, even if accepted, establishes at most that an alternative is conceivable. The wording should be aligned with the evidential strength; as written the concluding claim overstates the support.","section":"Section 10 vs. Abstract/Section 1"}],"minor_comments":[{"comment":"Typo: 'molarity landscape' should be 'modularity landscape.'","section":"Section 10"},{"comment":"Duplicate phrase: 'persist persist.'","section":"Section 7.1"},{"comment":"The terms 'aggregated' and 'aggregate' are used inconsistently; please unify the terminology.","section":"Section 3.2"},{"comment":"The GitHub links are valuable, but the paper should state which exact files or modules demonstrate the purported integrated style; a pipeline architecture diagram would help the reader verify the claim.","section":"Section 8.2"},{"comment":"The historical argument would be strengthened by engaging independent scholarship on the history of databases; as it stands, many references are self-citations or the authors' own interpretive readings of primary sources.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"This is a thought-provoking position paper, but the load-bearing bCLEARer evidence is self-reported and the historical causality is thin. I would suggest asking the authors for a detailed artifact analysis or an independent case-study evaluation before publication. The paper may also benefit from a historian of databases as a reviewer."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The central claim—that the schema turn, the near-universal separation of conceptual schema from information base, is a contingent product of 1970s constraints rather than a conceptual necessity—is worth taking seriously. The historical argument is largely sound and genuinely new. The modularity taxonomy (separated/aggregated/integrated) gives the discussion a workable vocabulary, and the package-migration example makes the alternative concrete. The paper reads well, stays in its declared scope, and is honest about being a position piece rather than a formal result.\n\nThe real soft spot is the bCLEARer evidence. The conclusion that the turn is not fundamental requires showing a working inclusive approach. The paper asserts bCLEARer implements one, but it never shows under its own definitions that bCLEARer's schema and base are not extensionally separated inside the pipeline. A data pipeline that builds a schema, keeps base instances in a separate store, and maps between them is standard ETL, not an integrated schema-and-base model. The open-source repositories are real evidence that something exists, but a reader cannot tell from this paper which of the three mixing categories applies. This is the stress-test concern, and it lands.\n\nThe historical causality in Section 5.3 is also asserted more than demonstrated. The technology-constraint story is plausible, but the paper does not connect the cited papers' actual choices to specific measurable limits of 1970s hardware or software. That is a secondary issue, not a fatal one.\n\nOn the circularity charge: I would not overweight it. The framework is built by the authors, but the historical analysis is independently checkable, and self-reported lessons from a long-running pipeline are acceptable for a position paper if labeled as such. The paper does label them as lessons learnt. Circularity becomes a real problem only because the existence proof is underdescribed.\n\nVerdict: conditional, but worthwhile. A serious referee should engage. The paper would benefit from (1) a precise statement of bCLEARer's internal modularity under the paper's own taxonomy, ideally with a small schema+base example from the repos; (2) more direct historical support for the technology-constraint causal claim; (3) softening 'not fundamental' to 'not shown to be fundamental' if the evidence stays as is. This is a paper for conceptual-modelling researchers, database textbook authors, and people designing modernisation tooling. I would cite it and bring it to a reading group. Send it to review.","headline":"The schema-turn diagnosis is a genuinely useful reframing, but the paper's central claim outruns its evidence because bCLEARer is never shown to avoid schema/base separation under the paper's own taxonomy.","tokens_in":17239,"tokens_out":2702,"would_cite":true,"duration_ms":32439,"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":"The paper argues that conceptual modelling's schema-only style is not fundamental but a response to 1970s limits, reversible by modern pipelines.","keywords":["conceptual modelling","schema turn","information base","conceptual schema","modularity architectural style","hylomorphism","pipeline-based conceptual modelling","schema-and-base model"],"falsifier":"Run the same package-application migration twice: once with the schema-first two-mapping architecture (source schema to common schema to target schema, then map the base across twice) and once with a single inclusive schema-and-base pipeline mapping the physical system as a whole. Track when conceptual mapping errors are first detected and the cost of fixing them. If the inclusive pipeline does not detect errors earlier or lower total cost, the paper's practical claim fails; if an integrated schema-and-base model is infeasible at realistic base volumes even with modern tooling, the claim that","tokens_in":16333,"feed_emoji":"🗃️","tokens_out":8098,"duration_ms":86941,"temperature":0.7,"pith_summary":"Conceptual modelling in computer science has crystallized around the schema: a conceptual schema completely separated from its information base, to the point where 'conceptual model' usually means just the schema. This paper claims that the schema turn is not fundamental. It reads the early history—Codd, CODASYL, Mealy, Chen, ANSI-SPARC/ISO—and argues the separation was a sensible response to two niche pressures: greenfield projects whose base data was not machine-readable, and 1970s technology that could not scale to instance-level data. Because modern technology removes both pressures, the authors argue, an inclusive schema-and-base conceptual model becomes practical, enabling more automated and empirically motivated practices. The paper develops a modularity framework and uses bCLEARer as a working example to show that this wider design space is real.","feed_headline":"Schema-first modelling is a historical detour, not a law","feed_subtitle":"Modern data pipelines let modellers keep schema and instance data together, enabling automated, evidence-driven design.","key_machinery":"Key machinery is the hylomorphic modularity framework: an adaptation of Aristotle's form/matter distinction, renamed formal/material, for describing the modular structure of computer applications. It separates the form of a representation from the content distinctions that motivate modularity, and classifies mixing into separated, aggregated, and integrated styles. The schema/base distinction is one material distinction—classes versus instances—that can be formally realized in different ways; the schema turn is the separated realization. The framework lets the paper plot historical and possible architectures on one landscape and identify the inclusive schema-and-base style as an available, u","core_discovery":"On the paper's own terms, the discovery is that the schema turn—the complete separation of the information base from the conceptual schema—is an evolutionary adaptation to a vanished niche, not a conceptual necessity. The evidence is historical: the earliest conceptual-modelling threads (the relational model, CODASYL, entity-relationship work, ANSI-SPARC) show conceptual models that included schema and data together, or describe a separation that was explicitly practical. The paper identifies the pressures that produced the separated style: greenfield development made the base unavailable in machine-readable form, and under-powered technology could not scale to large instance-level data. It","pith_inferences":["An implication the paper leaves implicit: if the schema/base split is a niche adaptation rather than an ontological necessity, then longstanding debates about whether instance data belongs in conceptual models are really disputes about tooling and project context, and should be settled by cost and error data.","Extending the germline analogy, in any long-lived information system the information base is the persistent lineage while schemas are replaceable vehicles; this would carry the paper's argument into data migration, data warehousing, and digital preservation, where the base is the invariant.","The paper's 'two worlds' rereading—conceptual and logical models as evolving prototypes that both show and tell—points beyond its explicit claims: the platform-independent/platform-specific split in model-driven engineering may itself be a residue of schema-centricity.","A testable extension the authors do not run: a side-by-side comparison of schema-first and schema-and-base pipeline modelling on the same reengineering task, measuring time-to-error-detection and cost of fixes."],"forward_implications":["If the schema turn is an optional style, then database textbooks and schema-centric CASE, UML, and MDE tools are teaching one historical practice, not the only possible form of conceptual modelling.","In package application migration, a single schema-and-base module mapped as a whole avoids the 'shift right' cost of finding conceptual mapping errors only during late volume testing.","Inclusive schema-and-base modelling moves the core of modelling from manual human editing into an automated data pipeline, making observability and inspectability necessary engineering practices.","With schema and base together from the start, design decisions can be tested empirically against the information base, using what the paper calls epistemic actions, instead of relying on purely rational analysis.","The same modularity landscape suggests that more novel conceptual-modelling architectures are likely discoverable beyond the inclusive schema-and-base style."],"supporting_citations":[{"why":"ISO/TR 9007 is the standard source of the conceptual schema / information base distinction the paper disentangles.","marker":"[1]"},{"why":"The database textbook that crystallizes the schema-centric practice the paper argues against.","marker":"[3]"},{"why":"The authors' earlier work on data-pipeline strategies and the bCLEARer implementation, cited as evidence that the inclusive approach is automated.","marker":"[4]"},{"why":"Shaw and Garlan's architectural styles supply the base notion of modularity style used throughout.","marker":"[18]"},{"why":"Codd's relational model anchors the historical thread showing early data modelling without an explicit schema/base split.","marker":"[35]"},{"why":"Mealy's three-world model is an early conceptual-modelling thread that does not separate schema and information base.","marker":"[42]"},{"why":"Chen's entity-relationship model introduces the level divisions and two-stage integration that the paper analyzes as a concrete modularity style.","marker":"[43]"},{"why":"The ANSI-X3-SPARC report provides the three-level architecture and the conceptual schema / conceptual data pair that the turn builds on.","marker":"[44]"},{"why":"Partridge's Business Objects documents the early bCLEARer re-engineering process, the long-running implementation of the inclusive approach.","marker":"[55]"},{"why":"Kirsh and Maglio's epistemic/pragmatic action distinction justifies testing modelling decisions against the information base.","marker":"[58]"}],"fun_headline_variants":["Schema-first modeling is a historical accident, not a law","The schema turn: a temporary detour, not a necessity","Bring data back into conceptual modeling now","Why schema-only modeling is just a phase","Modern data pipelines end the schema-data split"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The argument depends on the causal claim that the schema turn emerged from 1970s technical limits and greenfield data scarcity rather than from something required by representation itself; if the type/instance split is a necessary feature of any model, removing those constraints would not make the turn non-fundamental.","fun_headline_variants_meta":{"raw":{"variants":["Schema-first modeling is a historical accident, not a law","The schema turn: a temporary detour, not a necessity","Bring data back into conceptual modeling now","Why schema-only modeling is just a phase","Modern data pipelines end the schema-data split"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000591,"raw_usage":{"total_tokens":2586,"prompt_tokens":703,"completion_tokens":1883,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":447,"completion_tokens_details":{"reasoning_tokens":1811}},"tokens_in":447,"tokens_out":1883,"duration_ms":17188,"temperature":1.0,"reasoning_tokens":1811,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T12:20:32.082807+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same package-application migration twice: once with the schema-first two-mapping architecture (source schema to common schema to target schema, then map the base across twice) and once with a single inclusive schema-and-base pipeline mapping the physical system as a whole. Track when conceptual mapping errors are first detected and the cost of fixing them. If the inclusive pipeline does not detect errors earlier or lower total cost, the paper's practical claim fails; if an integrated schema-and-base model is infeasible at realistic base volumes even with modern tooling, the claim that","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"ISO/TR 9007 is the standard source of the conceptual schema / information base distinction the paper disentangles."},{"cited_title":"McGraw-Hill, New York, NY (2020)","cited_arxiv_id":null,"evidence_quote":"The database textbook that crystallizes the schema-centric practice the paper argues against."},{"cited_title":"In: STIDS 2024, Twelfth International Conference on Semantic Technology for Intelligence, Defense, and Security","cited_arxiv_id":null,"evidence_quote":"The authors' earlier work on data-pipeline strategies and the bCLEARer implementation, cited as evidence that the inclusive approach is automated."},{"cited_title":"Prentice Hall Englewood Cliffs (1996)","cited_arxiv_id":null,"evidence_quote":"Shaw and Garlan's architectural styles supply the base notion of modularity style used throughout."},{"cited_title":"In: Proceeding of AFIPS 1967 Fall Joint Computer Conference","cited_arxiv_id":null,"evidence_quote":"Mealy's three-world model is an early conceptual-modelling thread that does not separate schema and information base."},{"cited_title":"ACM Transac- tions on Database Systems (TODS)","cited_arxiv_id":null,"evidence_quote":"Chen's entity-relationship model introduces the level divisions and two-stage integration that the paper analyzes as a concrete modularity style."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The ANSI-X3-SPARC report provides the three-level architecture and the conceptual schema / conceptual data pair that the turn builds on."},{"cited_title":"Butterworth-Heinemann, Oxford (1996)","cited_arxiv_id":null,"evidence_quote":"Partridge's Business Objects documents the early bCLEARer re-engineering process, the long-running implementation of the inclusive approach."},{"cited_title":"Cognitive Sci- ence","cited_arxiv_id":null,"evidence_quote":"Kirsh and Maglio's epistemic/pragmatic action distinction justifies testing modelling decisions against the information base."}],"review_version":1}