From: Bernie Cosell Newsgroups: alt.folklore.computers Subject: Early BBN computer history (was Re: The very first software company?) Date: Mon, 11 Sep 2000 10:50:45 -0400 Organization: Fantasy Farm Fibers Lines: 120 Message-ID: <0slprs89ud41oe83eg2dtgkjd3c351kfco@news.supernews.net> References: <6yem2v54lh.fsf@localhost.localdomain> <8pb206$ipa$1@flood.weeg.uiowa.edu> <39B9DA6F.5AAC7243@cmc.com> X-Complaints-To: newsabuse@supernews.com X-Newsreader: Forte Agent 1.8/32.548 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news.ip-plus.net!news.tesion.net!news.belwue.de!informatik.tu-muenchen.de!news.informatik.uni-muenchen.de!uni-erlangen.de!newsfeeds.belnet.be!news.belnet.be!newsfeed.online.be!nntp-relay.ihug.net!ihug.co.nz!sn-xit-02!supernews.com!sn-inject-01!news.supernews.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:63714 Lars Poulsen wrote: } Bolt, Beranek and Newman was a consulting firm, specializing } in acoustical engineering. (My church once hired a BB&N staff } engineer to analyze the acoustics of our sanctuary in preparation } for a major re-layout of our organ.) They got sucked into } computer work because the NSA (No Such Agency) hired them to } help analyze some muffled tapes in various foreign languages, } and they thought Digital Signal Processing might have some } application to that. This is a garbled-enough rendition of BBN's early entry into the world of computers that I think a small correction/elaboration is in order. This is very sketchy and leaves a *LOT* out, but you'll get the idea.... Yes, BBN was started as an acoustics/consulting firm. Early in the 1960s, through some path I have no clue about, BBN got a contract from the NIH to investigate automating a hospital. They acquired the _first_ DEC computer to actually leave the mill -- PDP-1b, serial #2, and on it they wrote what is arguably the first operational timesharing system ("Little Hospital") -- there were a couple of similar projects afoot at that time [notably one at Dartmouth] but even if not "first" (whatever that means), they were certainly right at the very-leading-edge of all that. Not long after, BBN got a contract to do AI research with ONR [Office of Naval Research]. This project started using that same PDP-1 --- BBN had acquired another PDP-1 for the Hospital project [PDP-1d, serial #45] and wrote two more operating systems for the NIH ['Exec 3' was that last], and the AI folk 'inherited' the PDP-1b. Both projects had the same basic problem: since there was no suitable -system- software at the time, not only did you have to hire appropriate experts in the field [e.g, Dan Bobrow, Warren Teitleman, etc] but also *system* folk to _build_ something for them to work with/on. In the early days of the AI projects at BBN, Dan Murphy [of 'TECO' fame, among many other contributions] was their systems linchpin and at that time, and for a while after, BBN was *THE* place to go to do AI if you didn't want to work for a university. That group, the AI group, continued to flourish for a long time and quickly outgrew their PDP-1 and BBN got an SDS 940. Building on some of the clever ideas from Butler Lampson's TSS 1.85 on the 940, BBN designed/built some fancy hardware for the DEC PDP-10 and wrote a new operating system for it: TENEX [that DEC would later acquire and turn into 'TOPS-20']. This OS was primarily used by the AI folk [and indeed, the dialect of Lisp they used, 'BBNLisp' [and later 'InterLisp' when the core group moved from BBN to PARC] was probably the premier Lisp-dialect for AI-folk around the ARPAnet. Meanwhile, things were still a-boiling over in 'hospital land'. The project was a joint venture with the Mass General Hospital and there were a bunch of TTY lines strung from BBN to MassGeneral connecting TTYs [mod 33s] on various floors and stations back to the PDP-1d. And the same dichotomy: there was a systems crew, working on the timesharing system and an applications crew building hospital/medical apps *on* that TS. When the division director of the 'hospital' part of BBN moved on, Frank Heart [then at MIT's Lincoln Labs] came to BBN with several members of his 'team' from Lincoln and took over running the project. So at that point, BBN has two wholly separate computer groups, building their own, incompatible, idiosyncratic timesharing systems with applications groups associated with each. Oh, and no 'NSA' in sight..:o) Lisp is a *BIG* program, and does best with a pretty fancy operating system supporting it, and so the AI group generally worked on "big" stuff [and indeed, TENEX was quite an impressive piece of work!]. Meanwhile, the Hospital project petered out, the actual medical-application folk got a PDP-10 [running TOPS-10... AFAIK BBN's first not-written-in-house timesharing system!] and moved their focus there. The PDP-1d was kind of abandoned [and indeed, its systems folk needed something new to do]. Just then, JCR Licklider went to ARPA and with Larry Roberts [Larry, Lick and Frank Heart all knew one another -- Licklider was at BBN before he went to ARPA] and put together the RFP for the 'ARPAnet'. And so the hospital-systems-folk moved on and the PDP-1 got new life [a cross-assembler for the DDP-516 was written for it and the original IMP code was largely written on that PDP-1. That PDP-1 was also the very first NCC/NOC, again --- too many systems hackers with not enough to do and the PDP-1 quickly had a network interface and was processing IMP status reports...] Eventually, the IMP development was moved to one of the BBN-TENEX systems and the NCC functions were moved to a standalone H316 [there was a 'spare' 316 lying around after the development of the TIP and a little OS was cobbled up for it (BBN was good at that kind of thing..:o)) and *IT* started receiving and processing the IMP status reports] and both PDP-1s became just curios... sigh... And so you have the Hospital->ARPAnet and AI->TENEX groups at BBN. All really quite cutting edge. The two groups eventually mostly-merged, and then split again crosswise, as BBN spawned a commercial sibling ["BBN Computer Company, BBNCC"] all the while keeping the research-systems-folk in the original company... BBN did a fair bit of speech signal processing --- this as part of the 'AI->TENEX' group. There were two parts: first, part of the AI research at BBN involved a fair bit of natural language and speech recognition research [which required a nontrivial amount of signal processing on the voice-signals]. Also, back then the 'standard modem' was 2400baud. The gov't knew that analog-scrambling was not secure, and their most secure telephone was a box made by I-can't-remember-whom that digitized speech to 50Kb, then digital-encrypted it, sent that, decrypt/playback. That's OK for short-haul stuff, but the need for a 50Kb line was a real drawback, so BBN got involved in a LOT Of research to squeeze speech down to 2400baud [so it could be encrypted and thus used for a scrambled-phone over a mostly-vanilla phone line]. As a by-product of that, that group got VERY proficient at almost everything to do with signal processing, speech recovery, etc... And so when that infamous tape of Nixon's turned up with a 'gap', it was that group that got the actual tape and tried to analyze it to figure out what happened and see if they could recover any of the speech on the tape. [Yes, it was BBN's speech group that were the 'experts' who found the multiple stop/starts on the tape]. /Bernie\ -- Bernie Cosell Fantasy Farm Fibers bernie@fantasyfarm.com Pearisburg, VA --> Too many people, too few sheep <-- ###### From: jones@cs.uiowa.edu (Douglas W. Jones,201H MLH,3193350740,3193382879) Newsgroups: alt.folklore.computers Subject: Re: Early BBN computer history (was Re: The very first software company?) Date: 11 Sep 2000 19:19:30 GMT Organization: The University of Iowa Lines: 17 Message-ID: <8pjb82$cds$1@flood.weeg.uiowa.edu> References: <0slprs89ud41oe83eg2dtgkjd3c351kfco@news.supernews.net> NNTP-Posting-Host: pyrite.cs.uiowa.edu X-Trace: flood.weeg.uiowa.edu 968699970 12732 128.255.28.3 (11 Sep 2000 19:19:30 GMT) X-Complaints-To: usenet@news.uiowa.edu NNTP-Posting-Date: 11 Sep 2000 19:19:30 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news-ge.switch.ch!HSNX.atgi.net!newsfeed.mesh.ad.jp!sjc-peer.news.verio.net!ord-feed.news.verio.net!news.verio.net!news.uiowa.edu!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:63722 From article <0slprs89ud41oe83eg2dtgkjd3c351kfco@news.supernews.net>, by Bernie Cosell : > And so when that infamous tape of Nixon's turned up with a 'gap', it was > that group that got the actual tape and tried to analyze it to figure out > what happened and see if they could recover any of the speech on the tape. > [Yes, it was BBN's speech group that were the 'experts' who found the > multiple stop/starts on the tape]. The original tape also went to the Accoustics Research department at Bell Labs (Dept 1227) where I worked as an intern during the summers of 1973 and 1974. Max Matthews, I believe, was involved in the analysis of the gap (but not when I was there). Dept 1227 was alos big on speech processing and speech compression. Doug Jones jones@cs.uiowa.edu ###### From: bill_h Newsgroups: alt.folklore.computers Subject: Re: Early BBN computer history (was Re: The very first software company?) Date: Wed, 13 Sep 2000 23:13:40 -0700 Organization: Posted via Supernews, http://www.supernews.com Lines: 26 Message-ID: <39C06C94.410B@azstarnet.com> References: <0slprs89ud41oe83eg2dtgkjd3c351kfco@news.supernews.net> <8pjb82$cds$1@flood.weeg.uiowa.edu> Reply-To: bill_h@azstarnet.com X-Complaints-To: newsabuse@supernews.com X-Mailer: Mozilla 3.04Gold (Win16; I) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news.ip-plus.net!news.tesion.net!news.belwue.de!news.uni-stuttgart.de!logbridge.uoregon.edu!sn-xit-03!supernews.com!sn-inject-01!corp.supernews.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:63928 Douglas W. Jones,201H MLH,3193350740,3193382879 wrote: > > From article <0slprs89ud41oe83eg2dtgkjd3c351kfco@news.supernews.net>, > by Bernie Cosell : > > > And so when that infamous tape of Nixon's turned up with a 'gap', it was > > that group that got the actual tape and tried to analyze it to figure out > > what happened and see if they could recover any of the speech on the tape. > > [Yes, it was BBN's speech group that were the 'experts' who found the > > multiple stop/starts on the tape]. > > The original tape also went to the Accoustics Research department at Bell > Labs (Dept 1227) where I worked as an intern during the summers of 1973 > and 1974. Max Matthews, I believe, was involved in the analysis of the > gap (but not when I was there). Dept 1227 was alos big on speech > processing and speech compression. I recently heard something about a 'new' attempt to recover whatever was erased in that infamous 18 minutes. Something about advancing technology now having some chance of getting it back - and answering a lingering question or three ... Bill Tucson, AZ