{"id":"e058451a-72b2-466d-9b2f-8cfe64591a63","arxiv_id":"1908.09549","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The Oprema was an 8,313-relay, plugboard-programmed floating-point computer built at Carl Zeiss Jena in about 14 months, and it was the first program-controlled computer in the GDR.","lead":"This paper reconstructs the history of the Oprema, a relay computer built at Carl Zeiss Jena in 1954 and 1955, from archival diaries, memos, and technical notes. It gives new details on the machine's arithmetic circuits and argues it was the first computer in the GDR and the seventh in Germany.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 14-month timeline claim is internally inconsistent with the paper's own admission of design work before April 1954.","rationale":"The reader's weakest assumption focused on the external reliability of Kortum's May 1958 denial and the Lenski diary attribution. My concern is closely related but more precise: even taking those sources at face value, the paper's own text shows significant Oprema design work occurred between Kämmerer's return on 21 November 1953 and the first LWD entry on 5 April 1954. The 14 1/3 month figure therefore measures only part of the design effort, making the abstract's 'about 14 months' claim internally inconsistent. This does not undermine the existence of the Oprema, its technical description, or its status as the first GDR computer, but it does affect the 'remarkably short time' narrative that is part of the central claim. Because the paper otherwise handles uncertainty well and the correction is straightforward, a conditional acceptance—requiring the timeline claim to be revised or explicitly qualified—is the appropriate verdict rather than rejection or an unchanged accept.","tokens_in":50225,"tokens_out":8185,"duration_ms":80064,"concrete_test":"Recalculate the design-and-construction interval using the first documented design activity after Kämmerer's return to Jena on 21 November 1953 as the start date, drawing on the sources cited for endnote 31 (UACZ/BACZ No. 19570, memo 29.3.1955, and Kämmerer's 'Mein Weg zum Computerpionier'). If the resulting interval exceeds 14 months by more than 10%, the abstract's 'about 14 months' should be revised to a qualified range or the start date explicitly redefined.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The abstract's central timeline claim—that basic experiments, design, and construction of machine-1 were done in 'the remarkably short time of about 14 months'—is not consistently supported by the body text. The 14 1/3 month figure is computed from the first Lenski work diary entry (5 April 1954) to mid-June 1955, but the Design and Construction section states that 'Kämmerer had already done some significant design work before April 1954, especially after his return to Jena on 21 November 1953' (endnote 31). If that design work is counted, the interval from 21 November 1953 to the first productive use on 1 August 1955 is about 20 months, making the 'about 14 months' claim overstated. The paper acknowledges the ambiguity ('there is no clear-cut method to determine such time periods') but does not reconcile this with the abstract's unqualified 'about 14 months'. This is load-bearing because the speed narrative is part of the headline claim (7th computer in Germany, built remarkably fast), and the inconsistency is internal rather than a mere difference from external consensus.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper is an archival history of the Oprema (Optikrechenmaschine), a relay computer developed at Carl Zeiss Jena (CZJ) in 1954–1955 under Wilhelm Kämmerer's technical leadership and Herbert Kortum's managerial sponsorship. The paper claims that the Oprema was the seventh universal computer in Germany and the first in the GDR; that it consisted of two identical machines, each containing 8,313 relays, roughly 45,000 selenium rectifiers and about 250 km of cable; and that machine-1 was designed and built in about 14 months, measured from the first entry of Gerhard Lenski's work diary (5 April 1954) to its ready-for-use date around mid-June 1955. It reconstructs the machine's plugboard-based program storage, its four-address 27-bit instruction format, its decimal floating-point arithmetic with special values for zero, infinity and indeterminate, and its arithmetic algorithms, including a distinctive rounding scheme. It also recounts the biographies of Kämmerer and Kortum, their work in the KoKor development lab, their seven years of detention in the Soviet Union, the approval of the project by minister Rau in May 1954, the machine's operational record from 1955 to 1963, and the historiographical debate over whether Kortum should count as a German computer pioneer.","tokens_in":50394,"tokens_out":13077,"duration_ms":109381,"significance":"If the reconstruction holds, the paper secures the Oprema's place in the chronology of German and GDR computing and provides the most detailed technical account of the machine available, while correcting at least one figure in the secondary literature (Cortada's 25,000-relay count for the two machines combined). The paper's main strength is its archival grounding: the central claims are anchored in specific primary sources, including Lenski's work diary, memos in the CZJ and Deutsches Museum archives, Kortum's May 1958 report, and Kämmerer's 1955–1960 technical publications. The author consistently flags attribution inferences, contradictions between sources, missing documents, and gaps in the technical record: the attribution of the anonymous work diary to Lenski is argued step by step; the tension between Helga Kämmerer's recollection of Oprema discussions on Gorodomlia and Kortum's 1958 denial is reported frankly; and the incompleteness of the anonymous arithmetic notes, and the contradiction in the rounding description in Kämmerer's 1960 book, are stated rather than concealed.","major_comments":[{"comment":"The abstract's headline timing claim is not fully supported by the paper's own body text. The abstract says that 'basic experiments, design and construction of machine-1 were all done, partly concurrently, in the remarkably short time of about 14 months,' without qualification. In the Design and Construction section, however, the author states that 'Kämmerer had already done some significant design work before April 1954, especially after his return to Jena on 21 November 1953' (supported by endnote 31), and the same section later concedes that 'there is no clear-cut method to determine such time periods.' The 14 1/3-month figure is computed from the first Lenski work-diary entry (5 April 1954) to mid-June 1955, and the text also records the alternative start date of 17 May 1954 (minister Rau's approval). The stress-test concern therefore lands in part: the abstract overstates what the body's own evidence supports, and the inconsistency is internal. The fix is local: the abstract should specify the interval being measured (e.g., 'from the start of the documented design-and-construction work in April 1954') and should acknowledge the earlier design work discussed in the body. The body's own discussion is already appropriately hedged and does not require substantive change.","section":"Abstract; §Design and Construction of the Oprema"}],"minor_comments":[{"comment":"The sentence 'During these weeks the machine was operated 13286 hours per week' contains a garbled figure; endnote 86 shows the intended comparison is between 136, 1360 and 1320 hours, so the main text should read approximately '132 hours per week' (or state the 1320-hour ten-week total) and the endnote should be reconciled with the corrected wording.","section":"§Operation and Use"},{"comment":"The isolated phrase 'on average almost 10 % unnoticed errors' appears as a dangling fragment immediately before Table 2; it should either be integrated into the preceding paragraph about Kämmerer's human-versus-machine experiment or deleted.","section":"§Technical Properties, Reliability subsection"},{"comment":"Several typos and typesetting artifacts should be cleaned up: 'a relay computer were electromechanical relays' should be 'where'; 'the Havard Computation Laboratory' should be 'Harvard'; 'German Metereological Service' should be 'Meteorological'; 'who's now head of his own company' should be 'whose'; 'Seminumerical Algoritms' (endnote 30) should be 'Algorithms'; and 'GoldsƟne' (endnote 28) should be 'Goldstine' to match the spelling used elsewhere.","section":"Throughout"}],"recommendation":"minor_revision","confidential_remarks":"This is a well-sourced archival paper that fits the scope of a history-of-computing journal and reads as a finished contribution. The two self-citations (endnotes 33 and 93) refer to the author's earlier shorter treatments and do not carry the central claims, so circularity is not a concern. The only substantive revision I request concerns the abstract's timing claim, which is a one-sentence fix; I would support publication after that change."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a careful history of the Oprema relay computer, and it deserves a serious referee. What is genuinely new: Winkler uses the Lenski work diary and anonymous technical notes to reconstruct the instruction format, division, square root, and a distinctive rounding rule, and he corrects Cortada's published relay count (25,000 per machine to 8,313). The archival work is real. He flags contradictions, missing documents, and uncertain dates rather than smoothing them over, and the arithmetic reconstruction is explicitly labeled as based on incomplete notes. The claim that Oprema was the seventh universal computer in Germany and the first in the GDR looks supported, with appropriate caveats about how hard it is to date these machines.\n\nThe soft spot is the timeline. The abstract says basic experiments, design, and construction were done in \"about 14 months,\" but the body admits that Kämmerer had already done significant design work before April 1954, after his return to Jena on 21 November 1953. Depending on where you start, that is closer to 20 months. The paper acknowledges that \"there is no clear-cut method\" and even gives alternative figures (official 7.5 months, his own 14 1/3 estimate), but it does not reconcile the abstract's unqualified claim with the earlier design work. That is an internal inconsistency, not a fatal flaw, but it should be fixed by qualifying the abstract. The speed narrative is a real part of the paper's claim, so an editor should ask for a revision rather than let it stand.\n\nOther concerns are minor. The primary sources are not publicly accessible, so independent verification is inherently hard, but that is true of most archival history. The digressions — BBC broadcasts, ENIAC comparisons, the Kortum-pioneer discussion — are mostly on point and do not undercut the core. Self-citations are limited and not load-bearing.\n\nWho is this for? Historians of computing and anyone studying early relay machines or GDR technology. It is not a physics or CS result, but it is honest, well-sourced, and worth publishing after a modest revision. I would send it to peer review and ask the author to soften the timeline claim in the abstract and add one sentence acknowledging the earlier design work.","headline":"Solid archival reconstruction of the Oprema with a genuine correction and new technical detail; the abstract's 14-month timeline overstates what the body itself concedes.","tokens_in":50899,"tokens_out":1433,"would_cite":false,"duration_ms":17106,"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":"The Oprema relay computer was built in 14 months at Carl Zeiss Jena and became the first universal computer in the GDR.","keywords":["Oprema","relay computer","Carl Zeiss Jena","history of computing","GDR","plugboard program storage","floating-point arithmetic","Wilhelm Kämmerer"],"falsifier":"Finding a dated Oprema design document, such as a wiring diagram, an instruction-format sketch, or a relay count, from before 5 April 1954 in any surviving collection of the participants' papers would refute the claim that the design work began in Jena in 1954.","tokens_in":50018,"feed_emoji":"🖥️","tokens_out":9841,"duration_ms":91523,"temperature":0.7,"pith_summary":"Oprema was a relay computer for optical calculations, initiated by Herbert Kortum and designed and built under Wilhelm Kämmerer at Carl Zeiss Jena. The paper tries to establish that the two identical machines were designed and built in roughly 14 months, from early April 1954 to mid-1955, with formal completion reported on 30 December 1954, and that this made the Oprema the seventh universal computer in Germany and the first in the GDR. It reconstructs the machine from a surviving work diary and scattered technical notes: 8,313 relays per machine, plugboard program and data storage, a four-address 27-bit instruction format, and decimal floating-point arithmetic with special values for zero, infinity, and indeterminate results. If these claims are right, the GDR's computing history begins in 1954 with a relay computer built by an optics company under tight political deadlines, not in the later decade with transistor machines.","feed_headline":"Oprema: the 14-month relay computer that launched East German computing","feed_subtitle":"Reconstruction of the Jena machine sets the start of East German computing in 1954, not later.","key_machinery":"The load-bearing source is Gerhard Lenski's handwritten work diary, 58 pages covering 5 April to 24 December 1954, which fixes the project's start, its weekly decisions, and the final completion push; its attribution to Lenski rests on handwriting, his role as head of physical design, and the testimony of a later team member. The second load-bearing document is Kortum's May 1958 report, which explicitly denies that any concrete Oprema design work was done while he and Kämmerer were detained in the Soviet Union; without that denial the designed-and-built-in-1954 claim loses its force. On the technical side, the central mechanism is the 3-phase pulse system with cam-operated contactors, which switches relay contacts only when no current flows, and the control sequences ('Führungsketten') that propagate an activation wave through chains of bistable latching relays to sequence instruction execution.","core_discovery":"The central claim is that the Oprema was a genuine program-controlled relay computer, built as two identical machines, and that its design and construction were completed within roughly fourteen months. Each machine held 8,313 relays, about 45,000 selenium rectifiers, 250 km of wire, and an estimated 500,000 solder joints; programs of up to 300 instructions were wired onto plugboards, and data lived in plugboard constants, cyclic memories, and 32 relay registers. The instruction format was four-address and 27 bits long, and the number format was normalized decimal floating point in excess-3 BCD, with a 39th bit encoding zero, infinity, and indeterminate values, features that put the Oprema in the same family as Zuse's special-value floating point and, later, IEEE 754. The reconstruction is anchored by Lenski's work diary, which shows design work beginning on 5 April 1954, by the 17 May 1954 ministry meeting at which Kortum promised completion by the end of 1954, and by Kortum's May 1958 report denying that concrete plans were smuggled out of the Soviet Union. The paper concludes that the Oprema was the seventh universal computer in Germany, after the Z3, Z4, G1, Z5, ALWAC, and G2, and the first in the GDR.","pith_inferences":["A simulation of the Oprema's instruction set and cyclic-memory mechanism would let a modern reader measure how much of the 1955–1957 productivity jump came from the machine itself rather than from the team's growing experience; this test is not in the paper.","The paper's rounding analysis suggests a hardware trick, roundNorm, which avoids the ripple-overflow adder needed by roundTiesToEven and could be worth revisiting in low-power designs where a full adder is expensive; this is an engineering consequence the paper does not draw.","The Oprema's cyclic memories, which feed successive values through registers with jump sockets, can be read as an index-register mechanism shaped by ray-tracing workloads; a comparative study of early index mechanisms across relay and tube machines would place it in a broader lineage."],"forward_implications":["Programs of up to 300 instructions were wired into plugboards, yet the machine could still execute conditional and unconditional jumps and loops; Oprema was program-controlled without using an electronic stored program.","The planned twin-machine checking mode, two machines running the same program and comparing results after each instruction, was never wired up because both machines worked reliably after debugging; the Oprema therefore always ran as two separate computers.","Productive use rose from 64% of possible hours in 1956 to 88% in the first half of 1960, which supports the paper's picture of a relay machine that, once debugged, was dependable enough for production computing.","At Carl Zeiss Jena, the share of lens-design samples that reached production rose from 7% in 1954 to 83% in 1957, a change the paper ties to the Oprema's ability to trace more rays and handle aspheric surfaces."],"supporting_citations":[{"why":"The Lenski work diary that fixes the 5 April 1954 start and daily progress through December.","marker":"29"},{"why":"Kortum's May 1958 report denying concrete design work in the Soviet Union, the key support for the 1954-only timeline.","marker":"24"},{"why":"Kämmerer and Kortum's 1955 published description of the Oprema as a program-controlled twin-relay machine.","marker":"30"},{"why":"Minutes of the 17 August 1946 meeting recording Kortum's early plan for a computer for optical calculations.","marker":"17"},{"why":"Surviving technical notes from which the paper reconstructs multiplication, division, square root, and rounding.","marker":"56"},{"why":"Patent for the 3-phase pulse system with current-free contact switching, the central technical mechanism.","marker":"60"},{"why":"Jung's '5 Jahre Oprema' report supplying productive-hours and reliability statistics for 1956–1960.","marker":"83"}],"fun_headline_variants":["Oprema: 14 months to build Germany's seventh computer","Oprema: 8,300 relays, 14 months, first GDR computer","Oprema: built in 14 months, became GDR's first computer","Oprema: 14 months, 8,313 relays, and a computing first","Oprema: the relay titan that made East Germany compute"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The 14-month timeline rests on Kortum's May 1958 statement that he and Kämmerer did no concrete Oprema design work in the Soviet Union, and on the attribution of the anonymous work diary to Gerhard Lenski as a complete daily record.","fun_headline_variants_meta":{"raw":{"variants":["Oprema: 14 months to build Germany's seventh computer","Oprema: 8,300 relays, 14 months, first GDR computer","Oprema: built in 14 months, became GDR's first computer","Oprema: 14 months, 8,313 relays, and a computing first","Oprema: the relay titan that made East Germany compute"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000712,"raw_usage":{"total_tokens":3294,"prompt_tokens":1126,"completion_tokens":2168,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":742,"completion_tokens_details":{"reasoning_tokens":2064}},"tokens_in":742,"tokens_out":2168,"duration_ms":15341,"temperature":1.0,"reasoning_tokens":2064,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:07:36.021590+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Finding a dated Oprema design document, such as a wiring diagram, an instruction-format sketch, or a relay count, from before 5 April 1954 in any surviving collection of the participants' papers would refute the claim that the design work began in Jena in 1954.","supporting_citations":[{"cited_title":"Mr. Williams, University of Calgary, Alberta, Kanada","cited_arxiv_id":null,"evidence_quote":"Kämmerer and Kortum's 1955 published description of the Oprema as a program-controlled twin-relay machine."},{"cited_title":"About the Author Jürgen F","cited_arxiv_id":null,"evidence_quote":"Minutes of the 17 August 1946 meeting recording Kortum's early plan for a computer for optical calculations."},{"cited_title":"Zuse, Konrad Zuses Werk [The Works of Konrad Zuse], DVD V","cited_arxiv_id":null,"evidence_quote":"Patent for the 3-phase pulse system with current-free contact switching, the central technical mechanism."}],"review_version":1}