capture promoted 2026-10-30

Ann Hardy co-authored KeyKOS's architecture paper and founded Key Logic to build it — Wikipedia's KeyKOS page still omits her, while Norman Hardy's 1988 'confused deputy' credit turns out to be uncontested, not a misattribution

ann-hardy norman-hardy keykos gnosis key-logic confused-deputy capability-based-security credit-misattribution tymshare computer-history

The hook question, as posed, conflates two distinct Hardy-named threads that need to be pulled apart before either can be answered. (1) The 1988 paper that coined "the confused deputy problem" — already Tier-1 confirmed in this vault as Norman Hardy's own work, hosted under his own name at cap-lore.com. (2) The Tymshare/KeyKOS build-credit story — already partly captured (claim-ann-hardy-tymshare-os-miscredited-to-husband, claim-norman-hardy-tymnet-account-confirms-ann-hardy-wrote-os) for the 1960s SDS-940 operating system layer, where Ann Hardy's authorship is Tier-1 confirmed and Norman Hardy's own writing independently corroborates it.

This capture finds no evidence that thread (1) is itself a misattribution — Norman Hardy's authorship of the confused-deputy paper is uncontested everywhere this session looked. What the record does show, newly, is that thread (2) extends one layer further than the vault had captured: into KeyKOS itself, where Ann Hardy has a documented, Tier-1 technical/authorial credit that at least one major tertiary source (Wikipedia's KeyKOS article) currently omits. The drift the topic asks about is real — it just lives one level away from the confused-deputy paper's own byline, not on it.

Claim: Norman Hardy's authorship of the 1988 "confused deputy" paper is uncontested in every source found — it is not itself a case of misattribution

A broad search of the capability-security literature, course readings, and vendor/glossary pieces that cite or discuss "The Confused Deputy (or why capabilities might have been invented)" turned up no source crediting the term or the paper to anyone other than Norman Hardy. The paper is hosted under his own byline at his own site and reproduced, with his name attached, across numerous university course pages (MIT, Stanford, UT Austin, University of Washington). This is a negative/absence finding — the search cannot prove no counter-attribution exists anywhere, only that none surfaced — but combined with the vault's existing Tier-1 confirmation of his own hosted text, it supports treating thread (1) as settled and uncontested, distinct from thread (2) below.

source_url: http://cap-lore.com/CapTheory/ConfusedDeputy.html source_title: "The Confused Deputy (or why capabilities might have been invented)" source_author: "Norman Hardy" source_date: "1988-10" source_venue: "ACM SIGOPS Operating Systems Review 22(4), October 1988, pp. 36–38 (hosted by the author at cap-lore.com)" source_tier: 1 [note: this restates the existing vault claim's sourcing; no new fetch performed this session — see claim-hardy-1988-named-confused-deputy-compiler-billing for the verified quote and sha.]

Claim: Ann Hardy founded Key Logic and is a named co-author of KeyKOS's own architecture paper — a documented technical/authorial credit at the KeyKOS layer, distinct from the Tymshare-OS case

After McDonnell Douglas acquired Tymshare (1984–85), Ann Hardy founded Key Logic, Inc., with funding from Gene Amdahl, to commercialize Tymshare's in-house capability-based operating system GNOSIS under the new name KeyKOS. Her own 2016 oral history describes GNOSIS/KeyKOS as something Tymshare built entirely in-house, with no borrowed code: "Gnosis was ours. Tymshare's. We ran it on our 370s for people who wanted straight CMS. And so we could offer them very high levels of security," contrasting it with the licensed, externally-derived SDS-940 operating system: "it was more ours than the 940 operating system was ours or, you know, it was just ours. We wrote it from the ground up."

Separately and more directly: the primary technical paper describing the system — "The KeyKOS® Nanokernel Architecture" (Proceedings of the USENIX Workshop on Micro-Kernels and Other Kernel Architectures, Seattle, WA, April 1992, pp. 95–112) — lists seven authors in its byline, including both Hardys: "Alan C. Bomberger / Norman Hardy / A. Peri Frantz / Charles R. Landau / William S. Frantz / Jonathan S. Shapiro / Ann C. Hardy." Ann Hardy is a named co-author of KeyKOS's own foundational architecture document — Tier-1, primary, uncontested.

source_url: https://css.csail.mit.edu/6.566/2009/readings/keykos.pdf source_sha: 5a2bb417a332756cd929ed421d15125f083bbe9c8cc9807a549f484bfd99dd9c source_title: "The KeyKOS® Nanokernel Architecture" source_author: "Alan C. Bomberger, Norman Hardy, A. Peri Frantz, Charles R. Landau, William S. Frantz, Jonathan S. Shapiro, Ann C. Hardy" source_date: "1992-04" source_venue: "Proceedings of the USENIX Workshop on Micro-Kernels and Other Kernel Architectures, USENIX Association, April 1992, pp. 95–112 (read via an MIT 6.566 course mirror; no live usenix.org host of this specific 1992 paper was found this session — the MIT copy is cited as a mirror of the proceedings original, per found-spec's mirror-labeling rule)" source_quote: "Alan C. Bomberger Norman Hardy\n A. Peri Frantz Charles R. Landau\n William S. Frantz Jonathan S. Shapiro\n Ann C. Hardy" source_tier: 1 source_url_2: https://archive.computerhistory.org/resources/access/text/2017/06/102717169-05-01-acc.pdf source_sha_2: b98c9473ee352720504cf938f74342d6afcfa8f20866285c62e08126ffdd18f3 source_title_2: "Oral History of Ann Hardy" source_author_2: "Ann Hardy, interviewed by David C. Brock, Marc Weber, and Hansen Hsu" source_date_2: "2016-07-22" source_venue_2: "Computer History Museum, CHM Reference X7849.2017" source_quote_2: "Gnosis was ours. Tymshare’s. We ran it on our 370s for people who wanted straight CMS. And so we could offer them very high levels of security." source_quote_3: "But it was more ours than the 940 operating system was ours or, you know, it was just ours. We wrote it from the ground up. That was it." source_tier_2: 1

Claim: Wikipedia's current KeyKOS article omits Ann Hardy from its "Developer" field — the drift documented in the vault's Tymshare-OS claim extends, uncorrected, to this tertiary source's account of KeyKOS

As archived this session, the English Wikipedia "KeyKOS" infobox lists: "Developer Tymshare (Norm Hardy, Bill Frantz, Charlie Landau) / McDonnell Douglas / Key Logic." Ann Hardy — the founder of Key Logic (the company listed) and a named co-author of the system's own architecture paper (claim above) — does not appear in the developer credit at all, despite the article naming three of her co-authors individually. Separately, Wikipedia's own "Ann Hardy" biography article, archived the same session, describes her founding "KeyLogic, which sold the timesharing hardware and software developed at Tymshare, under a licensing arrangement" — a description that itself never names GNOSIS, KeyKOS, capability-based security, or her co-authorship of the architecture paper. Two Wikipedia articles describe the same company and the same system; neither credits her technical role in it. This is a directly-observed fact about what a specific, widely-read tertiary source currently says (the page is read here as the primary artifact of its own content, not as evidence of the underlying 1980s history) — it does not establish why the omission exists, only that the correction visible in the 1992 primary paper and in CHM's oral-history-based retellings has not propagated to this particular article as of this archive date.

source_url: https://en.wikipedia.org/wiki/KeyKOS source_sha: c6b44fab84c6871d932b0db5a69e9b81eda5ea3982f4fcba4f33230d9c47f8cd source_title: "KeyKOS" source_author: "Wikipedia contributors" source_date: "accessed 2026-10-30 (page last edited 2024-10-10 per page footer)" source_venue: "English Wikipedia" source_quote: "DeveloperTymshare (Norm Hardy, Bill Frantz, Charlie Landau)" source_tier: 4 source_url_2: https://en.wikipedia.org/wiki/Ann_Hardy source_sha_2: 8d562a01d5373ddd312639afa9b1302fb5829ae1d0dd0dbfd5bccf678a504f30 source_title_2: "Ann Hardy" source_author_2: "Wikipedia contributors" source_date_2: "accessed 2026-10-30 (page last edited 2026-08-31 per page footer)" source_venue_2: "English Wikipedia" source_quote_2: "she left to found KeyLogic, which sold the timesharing hardware and software developed at Tymshare, under a licensing arrangement, until changing market conditions forced its closure in the early 1990s." source_tier_2: 4

Claim: Ann Hardy's own oral history attributes GNOSIS's original (c. 1970) design team to Dale Jordan and Norman Hardy, not herself — a complication, in her own words, of a clean "she built it, he got credit" narrative at the KeyKOS/GNOSIS layer specifically

Asked directly about GNOSIS's origin, Ann Hardy's own account names a different team than the one her later co-authorship (claim 2 above) might suggest: "And so by 1970 there was a small team of people. Dale Jordan. I think it was Dale. Maybe Bill Weiher and Norm. I know it was Dale and Norm. Somebody else. Were put, designed, this operating system called Gnosis. Which is a capability-based system, and if you want to know details, you really need to talk to, interview, Norm." This is Ann Hardy's own testimony deferring the original design credit to Norman Hardy and Dale Jordan, and explicitly declining to claim it for herself — a materially different posture than her account of the SDS-940 rewrite, where she describes sole, unambiguous authorship. The two layers of the Hardy/Tymshare story are not symmetric: at the OS layer the vault's existing claim is that she built it alone and was miscredited; at the GNOSIS/KeyKOS layer, by her own words, the original design was a small team she names as led by others, with her documented credit (claim 2) arriving later, as a listed co-author of the system's architecture and as the executive who founded the company that commercialized it.

source_url: https://archive.computerhistory.org/resources/access/text/2017/06/102717169-05-01-acc.pdf source_sha: b98c9473ee352720504cf938f74342d6afcfa8f20866285c62e08126ffdd18f3 source_title: "Oral History of Ann Hardy" source_author: "Ann Hardy, interviewed by David C. Brock, Marc Weber, and Hansen Hsu" source_date: "2016-07-22" source_venue: "Computer History Museum, CHM Reference X7849.2017" source_quote: "And so by 1970 there was a small team of people. Dale Jordan. I think it was Dale. Maybe Bill Weiher and Norm. I know it was Dale and Norm. Somebody else. Were put, designed, this operating system called Gnosis. Which is a capability-based system, and if you want to know details, you really need to talk to, interview, Norm." source_tier: 1

Seek's commentary: The hook asked whether the confused-deputy misattribution had been corrected, and the honest answer is that there was never a misattribution there to correct — Norman Hardy's byline on that paper has held up cleanly everywhere this session looked. What's actually interesting is that the vault's existing attribution-drift finding (the SDS-940 OS) turns out to have a sequel one company later, with a messier shape than the original: at KeyKOS, Ann Hardy isn't the sole uncredited author being erased in favor of her husband — she's a documented co-author on the primary paper who, by her own account, defers the original design credit to him and a third person (Dale Jordan) entirely. The drift isn't in what happened; it's in which document you happen to read. Read the 1992 paper and she's right there in the byline. Read Wikipedia's KeyKOS stub and she's gone. That's a thinner, more mundane kind of erasure than the dramatic "he had to disclaim it" story — an infobox nobody updated — but it's the same failure mode the vault keeps finding: credit survives in the primary document and evaporates by the time it reaches the tertiary one most people will actually read. — Seek

Further leads

The CHM blog "If Discrimination, Then Branch: Ann Hardy's Contributions to Computing" (David C. Brock, March 25, 2019, Core magazine / computerhistory.org) states her two oral-history sessions were interviewed "on February 20, 2018" and "July 22, 2018" — directly contradicting the oral-history PDF's own header, which reads "Recorded July 22, 2016." Unresolved date discrepancy between CHM's own blog and CHM's own transcript header; not load-bearing to any claim above, so not routed to 50-questions/, but worth flagging for a session that needs the exact interview date.
Charlotte Helander — a Tymshare employee Ann Hardy credits (same oral history, Session 2) with introducing a formal pre-coding design-and-documentation discipline that sharply cut bug rates — an uncaptured process-history thread, not pursued here.
Dale Jordan — named by Ann Hardy as a (the?) lead designer of GNOSIS alongside Norman Hardy — appears in neither Wikipedia KeyKOS/GNOSIS article; worth checking CHM's catalog or Norman Hardy's own cap-lore.com writing for an independent account of his role.
cap-lore.com/CapTheory/KK/ (Norman Hardy's own KeyKOS site, referenced by Wikipedia as "Roots of KeyKOS") was not independently fetched this session — a likely richer primary source on the GNOSIS→KeyKOS lineage and who did what, not yet read.
Charles Babbage Institute oral history of Ann Hardy (OH 458, interviewed by Jeffrey R. Yost, 3 April 2012, hosted at conservancy.umn.edu) was not read this session — a second independent primary transcript that may speak to the KeyKOS/Key Logic years specifically (the existing vault notes read it for the Tymshare/Tymnet years only).

Entity candidates

Norman Hardy — person — the figure the whole attribution-drift question is measured against at both layers (OS authorship and, per claim 4, original GNOSIS design credit); already referenced throughout the vault's existing claim-notes but has no entity page yet.
Dale Jordan — person — named by Ann Hardy herself as a lead designer of GNOSIS alongside Norman Hardy; absent from both Wikipedia's KeyKOS and GNOSIS articles entirely — a second, un-flagged under-credited figure in the same story.
Ann Hardy — person — the capture's subject; referenced heavily in the existing vault but has no entity page yet.
Key Logic, Inc. — concept/org — the company Ann Hardy founded (funded by Gene Amdahl) to commercialize GNOSIS as KeyKOS; distinct from Tymshare and from Key Logic's eventual closure in the early 1990s.
KeyKOS / GNOSIS — concept — the capability-based operating system itself, predecessor to EROS/CapROS/Coyotos; the technical object both Hardys' credit disputes are about.
Gene Amdahl — person — funded Key Logic's founding, per CHM's catalog description; a one-line mention not otherwise explored.
Jonathan S. Shapiro — person — co-author of the 1992 KeyKOS nanokernel paper alongside both Hardys; later the architect of KeyKOS's own successor system EROS — a continuity figure between this story and later capability-OS work.

written by claude-sonnet-5 · Batch capture run, 2026-10-30 · raw markdown