Oral history of British science
Furber, Steve (Part 3 of 5)
The British Library Board acknowledges the intellectual property rights of those named as contributors to this recording and the rights of those not identified.
Legal and ethical usage »
Interviewee's office, University of Manchester
Furber, Steve, 1953- (speaker, male)
Lean, Thomas (speaker, male)
Part 3 [1:51:56][Interview Three: 22 October 2012] Comments on early involvement of CPU with computers : SF link membership of Cambridge University Processor Group [CPUG], where Hermann Hauser and Chris Curry recruited staff for CPU; Clive Sinclair spinning out Science of Cambridge Company, with Curry, to explore computing; Science of Cambridge MK14 computer, prototyped by SF, improved on by Sophie Wilson by the 6502 processor based Hawk computer, later the Acorn System 1. [02:12] Discussion about System 1: limitations of SC/MP processor powered MK14 compared to the more useful System 1; description of System 1 and its use; anecdote about System 1's appearance in 'Blake's 7' television series. [05:55] Remarks on: CPUG, use of term 'hobbyists', activities; SF limited involvement in System 1 during development of Programmable Read Only Memory holding Wilson's operating system; first use of Acorn name with System 1; System 1's modular and expandable nature, exploited in rack based systems for specialist uses; anecdote about SF designing an analogue to digital card for Acorn. [10:15] Remarks on Atom Computer: American development of boxed consumer computers; roles of Curry, Wilson and designer Nick Toop; SF limited involvement; description of Atom's boxed complete design, anecdote about Motorola display controller chip causing problems tuning to a UK television; anecdote about Atom's wider appeal beyond skilled hobbyists; importance of Atom to Acorn; Econet network technology, encouraged by Andy Hopper; anecdote about expansion of Atom computer causing problems; status of networking at time, similarities of HDLC protocol to later Apple Talk; differences between Atom and System 1; [17:00] anecdote about SF writing 'Meteors' a version of 'Asteroids' game for the Atom, his only commercial software product; influence of electronics press 'Wireless World' and 'Electronics Today International' in keeping British hobbyists up to date with development from USA. [19:35] Comments on: SF feelings in late 1970s about impact of microprocessor and prospects for computing; description of designing a digital to analogue card for the System 1, prototyping equipment using Veriwire, use of bus to connect peripheral cards to processor, Acorn's Hauser, Chris Turner, Paul Pedder productionising SF circuit design, changes in circuit board production, low cost of development. [25:15] Story about origins of Acorn involvement with BBC: SF building new computer at home and involvement of new Acorn proton computer; anticipated demise of 8bit processors, leading to idea of Proton being a dual processor with 6502 and 16bit processor; Chris Curry becoming aware that BBC wanted to produce a computer; Acorn bidding 6502 part of Proton to BBC. [28:09] Story about about hectic week in 1981 preparing prototype computer for BBC: portrayal in 'Micromen'; Hauser's unconventional encouragement of Wilson and SF to complete prototype; wiring of design by Randomnich Bandagee; lengthy debugging of design, anecdote about Hauser's vital contribution; Alan Boothryod's industrial design of a case for new computer; presentation of computer to BBC, leading to BBC Micro. [31:48] Remarks on SF decision to move to Acorn after completion of research fellowship rather than follow Sean Fochs Williams suggestion to move to San Diego. [33:00] Comments on: SF early work as full time employee at Acorn fixing ULA gate arrays with Ferranti in Oldham; concept for Proton computer, SF work at home, advantages of combining main and graphics memory, anecdote about David Braban's 'Elite' game exploiting display in unexpected ways; Acorn premises off Market Square in Cambridge; SF working at home in house on Hollybush Way in Lynton, after marriage to a teacher in 1977; decision to use dual processor on Proton, exploited on through the Tube interface for a second processor on the BBC Micro, and on the Acorn Cambridge Work Station with a National Semiconductor 32016 processor. [40:50] Comments on: SF memories of hectic week preparing BBC Micro compared to 'Micromen', role of Randomnich Bandagee; description of debugging BBC Micro prototype using logic probes; SF surprise that BBC Micro prototype worked. [45:20] Description of wire-wrapping technique used to make prototype circuit boards, tacit knowledge of order to connect pins, use of technique to build large machines in University Computer Labs and parts of mainframe computers. [49:20] Remarks on BBC Micro and Acorn: SF memories of BBC demonstration; excitement at Acorn over BBC connection, importance of BBC to establishing Acorn as a trusted brand; sales of BBC Micro exceeding BBC's expectations; financial risks to Acorn, later sale to Olivetti; marginal design features of BBC Micro, such as quirk with 81NS95 chips; [54:37] problems with Ferranti ULAs [Uncommited Logic Array] used for video processing, eventually replaced with INMOS VLSI chip; small space inside Alan Boothroyd's computer case causing heating problems in linear power supplies demanded by BBC, until replaced by Aztech switching supply. [57:15] Description of ULA and their use: multi-purpose Ferranti chip whose purpose was defined late in manufacture; use in video and serial processor on BBC Micro, role of Computer Laboratory in development of video processor design, problems in video processor; importance in reducing number of chips in computer; SF first exposure to chip design, contrast with earlier hobbyist activities using chips but not designing them. [1:01:20] Discussion of design and production of ULAs: SF involvement in designing functionality of chip; University Computer Lab's Andy Hopper, Jeremy Dion and Peter Robinson's role designing final layer metalisation for BBC micro using Rubilith masks; Oldham based Ferranti producing ULAs; description of how ULA's NOR gates can be connected in different ways; hand design and prototyping techniques. [1:06:00] Remarks on informal contacts with Cambridge University Computer Lab: SF use of facilities and computers whilst a student; Andy Hopper forming Orbis company with Hauser, which merged with Acorn; SF contact with Jeremy Dion and Peter Robinson during ULA work. [1:08:20] Comments on Chris Curry: early work with Sinclair, less involved technically than Hauser, role in Atom development, SF relationship with Curry compared to Hauser. Comments on Hermann Hauser, background in research physics at Cavendish Lab giving him technical empathy, focused personality, good at recognising strong teams, family background in Austrian wine making, good relations with SF. [1:13:35] Comments on Andy Hopper: Polish background, business oriented academic, involvement with Hauser through Orbis; manager of Cambridge Olivetti research lab until closure by AT&T; good at combing university and entrepreneurial companies, such as Verata; role in Acorn, such as in use of Econet networking. [1:17:30] Comments on Sophie Wilson: called Roger during early 1980s; studied computing; CPUG member; anecdote about Wilson debugging SF's computer; drawn into Acorn during fruit machine work; software expertise complimenting SF hardware expertise at Acorn; perfect memory; constructive technical disagreements, comparison with value of Mike Muller's contrarian attitude; good interaction and mutual respect with SF. [1:23:30] Remarks on team at Acorn: dangers of becoming a clique; SF role as Deputy Manager Advanced Research and Development under Jim Mitchell in Palo Alto; SF managing Sophie Wilson. [1:23:30] Remarks on other Acorn colleagues and their later work for ARM: Chris Turner, Acorn first employee, practical engineer; Mike Muller and Tudor Brown, who became senior personnel in ARM; Alastair Thomas, who committed suicide; VLSI group's Robert Heaton; Jamie Urquet, later role at ARM, Pond venture capital, oversight of SF current research; former GEC chip designer Harry Oldham; Dave Howard; tool designer John Biggs; Lee Smith, Hugo Tyson, Brain Cockburn, John Thackary; strength of group overall. [1:30:10] Remarks on: SF role in Acorn as hardware technology leader; role of Sophie Wilson and Paul Bond on software; contrast of SF managerial role with that of Sophie Wilson; growth of Acorn; Acorn workforce size in early days. [1:33:00] Discussion about Acorn offices: basic offices at beginning; anecdote about noisy printer being kept in the toilet; description of later open plan offices in Cherry Hinton, anecdote about building's previous use in hard water softening, anecdote about people walking into glass doors, SF and Wilson office until move to 'Tin Palace' extension; anecdote about difficulties caused by identical telephone ringers. [1:38:30] Discussion on SF working day at Acorn: meetings and management; technical work at home; brainstorming meetings; SF work on first ARM chip in office; use of email in 1980s; use of Acorn computers within Acorn, such as design of ARM on BBC Micro; diverse days; travel to Acorn research in Palo Alto and suppliers, such as National Semiconductor research centre in Haifa, Israel; conferences, such as VLIS89, where he first heard Austek's Craig Mudge talk about Asynchronous Design; SF enjoyments in working at Acorn with strong team; [1:43:00] anecdote about applying VLSI design team's practices to improve circuit board production; team contributions to improving processes; anecdote about design reviews and confidence of staff; problems integrating new staff into established teams; Hauser's confidence in team protecting research team from redundancies during financial troubles. [1:47:30] Comments on freedom in work, with example of ARM: speculative work by SF and Sophie Wilson; backing of Hauser's and self confidence of company overcoming strangeness of venture; influence of Berkeley RISC ideas; visit National Semiconductor team in Haifa; SF expecting Acorn's ARM work to be overtaken by others before coming to fruition as Acorn should not have the resources to develop a microprocessor.
Interview with computer scientist, Steve Furber