[SIGCIS-Members] Donald Knuth
ktatarchenko at yahoo.fr
Thu May 15 09:42:28 PDT 2014
This is a fascinating exchange!
I am actually writing from the Stanford archives, where I am looking through McCarthy's and Knuth's materials.
Indeed, Knuth has been preoccupied with History for some time.
In 1979, he and Andrei Ershov organized a "pilgrimage" to Uzbekistan, the "cradle" of the algorithm. Of course, medieval mathematics had little importance per se, but the questions about the subject matter of computer science defined the agenda for the meeting.
My favorite quote from the proceedings is this one by Heinz Zemanek :
abstractions are as operational and goal oriented as the abstractions
which are used for and are running the computers of our own
century... I think that we computer scientists of the 20thcentury, in particular the algorithmic community, have quite a lot to
learn from Al-Khwarizmi’s method and success...Al-Khwarizmi
teaches the programmer humbleness a thousand years before Dijkstra
and reminds us that we are servants of our society just as he was a
servant of the caliph.
At this meeting history's function was to preparer a particular future for the discipline of computer science. But the location and timing also make it an important episode for understanding Cold War science.
There are many incredible images from the meeting in the Ershov archive at http://ershov.iis.nsk.su/archive/eaindex.asp?lang=2&gid=24
I did this research for my dissertation (chapter 4, part II "The Past Perfected" ) which is available on proquest. And I hope to develop it further in my book manuscript.
So I am truly looking forward to seeing the recording of this recent talk.
Le Jeudi 15 mai 2014 10h37, Paul N. Edwards <pne at umich.edu> a écrit :
I completely agree with Tom's analysis. A few additional questions re. Knuth’s abstract:
* Who is the “we” in Knuth’s “let’s”?
* In what way has history ever been “useful” to computer scientists? It would be interesting to hear Knuth’s views on this - useful in education? Personal interest? Celebration? Credit (reputation)?
* Finally, for what audience do people write internalist history today? Who publishes it, and who actually reads it? Could someone create some statistics, for example, on the number of downloads of Annals articles, divided along some rough criterion between internal and external history?
In my view, as in Tom’s, there’s still a great deal of internalist history going on. A lot of it is at the level of raw materials - oral histories collected by ACM (I did a couple of these myself a few years back), collecting of documents and other archival materials, etc.
The Computer History Museum — probably the single best-funded institution in our field — generally takes an internalist approach, though it does try to build a bridge to external history. It would be interesting to get some data on CHM visitors’ backgrounds. I imagine it’s unbelievably hard to create compelling museum exhibits on the history of key ideas in algorithms, computer architecture, software engineering, etc. Is CHM visited by computer scientists in search of their field’s past?
Knuth’s idea that “we only get a scorecard” is downright offensive. Aside from journalistic corporate histories, I can’t think of much professional history that fits this description — certainly few people on this mailing list write that kind of thing. This makes me suspect that Knuth does not have much of a sense of the difference between journalism and professional historiography, nor of this SIG as a community. Can we educate him?
On May 14, 2014, at 12:19 , Thomas Haigh <thaigh at computer.org> wrote:
Hmm. I can't find an online video for it anywhere, which is an increasingly common practice for invited talks. We might try to get Knuth to write up the talk for Annals if he wants to make his case to the history of computing community.
>Here is the abstract:
>Talk Let's Not Dumb Down the History of Computer Science
>Abstract For many years the history of computer science was presented in a way that was useful to computer scientists. But nowadays almost all technical content is excised; historians are concentrating rather on issues like how computer scientists have been able to get funding for their projects, and/or how much their work has influenced Wall Street. We no longer are told what ideas were actually discovered, nor how they were discovered, nor why they are great ideas. We only get a scorecard.
>Similar trends are occurring with respect to other sciences. Historians generally no prefer "external history" to "internal history", so that they can write stories that appeal to readers with almost no expertise.
>Historians of mathematics have thankfully been resisting such temptations. In this talk the speaker will explain why he is so grateful for the continued excellence of papers on mathematical history, and he will make a plea for historians of computer science to get back on track.
>Annals liked to feature a quote from Knuth in its promotional materials, and of course he was involved in documenting the very early history of systems software back at the Los Alamos conference in 1976. So I suspect he is unhappy with the demographic transition that has taken place in the history of computing from eminent computer pioneers to younger, Ph.D. historians as the most active producers of historical work. This has naturally been accompanied with a shift in methods, questions, framing, etc. from an internalist approach centered on questions of interest to computer scientists to an externalist approach centered on questions of interest to one or another tribe of Ph.D. historians.
>It’s natural for Knuth to regret this, but it’s not clear to me that this is a zero sum game. His abstract reads as if there is a fixed pool of historians whose attention has unfortunately been diverted from substance to fluff. It seems that the first generation of digital computing pioneers had an interest in technical history, stoked in part by the legacy of the ENIAC patent wars. Wilkes, Eckert, Mauchly, Zuse, Goldstine, Metropolis, the IBM team of Bashe, Pugh, et al, Malinovsky, Burks, Randell, Ware, Sammet, the LEO team, and many of their peers are/were active produces and consumers of technically and/or institutionally oriented history. All had begun to work in computing by 1957, which seems to be about when Knuth himself first programmed. While many more people entered computing after 1957 than before it seems that subsequent cohorts have been much less likely to develop an interest in history. There are exceptions of course. Dave Walden,
for example, is one of the most active members of the Annals board. Several ACM SIGs launched historical projects, following the three successful history events on the History of Programming Languages organized by SIGPLAN over the decades. ACM and IEEE CS both have history committees. Articles are written to celebrate the anniversaries of departments, technologies, etc. Just today we heard on this list of a project on the history of BSD. But overall it seems that the relative eclipse of technical, internal history of computer science and technology has a lot to do with a loss of interest in history among the people best equipped to write it. Neither have computer science departments embraced the history of computer science as an important area of teaching or research. As far as I know, no computer science program in the US has ever hired a faculty member specifically as a historian of computing – which is different from the history of law, medicine,
communications, and to some extent business where the disciplines in question have sometimes deliberately hired faculty members to teach and research history. Instead history has been an interest people have developed late in their careers, if at all.
>Knuth probably appreciates the efforts of De Mol, Bullynck, and their colleagues to establish the Commission for the History and Philosophy of Computing and the associated series of events over the past few years. This reflects an engagement with the history of mathematics, which as Knuth notes maintained a more traditional approach to the history of science. So it seems that there is scope for many historical traditions to thrive side by side.
>My other point is that “the history of computer science” is a problematic category in this respect. Much technical history of computing is on the 1940s and early 1950s, before the emergence of computer science. The bulk of recent effort has been on Turing. Even my current work with Priestley & Rope on the history of ENIAC, which goes deep into technical analysis of early code, architecture, flow diagramming techniques, concepts, etc. is really about the history of computing practice and computing technology rather than the history of computer science. Furthermore relatively little externalist work on the history of computing is about computer science. Mahoney wrote about computer science, and there has been recent work on the history of Algol from the efforts of the SOFT-EU project, history of software engineering, and history of formal methods. Also coverage of the history of ARPA funding in the books by Norberg & O’Neill and Abbate, and a
couple of articles on NSF support for computer science by Aspray. That’s a rather small proportion of everything written on the history of computing over the past twenty years. So in as much as the history of computer science is written at all, which is not nearly as much as it should be, the dominant approach is still internalist and technical.
>This might be an interesting topic for one of my “Historical Reflections” columns in Communications of the ACM. So if you send your thoughts to the list these could help to shape it.
>From: members-bounces at sigcis.org [mailto:members-bounces at sigcis.org] On Behalf Of Ceruzzi, Paul
>Sent: Wednesday, May 14, 2014 9:18 AM
>To: 'members at sigcis.org'
>Subject: [SIGCIS-Members] Donald Knuth
>I heard that Donald Knuth gave a Kailath Lecture at Stanford last week on "Let's not Dumb Down the History of Computer Science." Did any of you attend this lecture? Is there a transcript available? I've seen the abstract and it appears to be of great relevance.
>Paul E. Ceruzzi, Chairman
>Division of Space History, MRC 311
>National Air and Space Museum, Smithsonian Institution PO Box 37012 Washington, DC 20013-7012
>This email is relayed from members at sigcis.org, the email discussion list of SHOT SIGCIS. The list archives are at http://sigcis.org/pipermail/members/ and you can change your subscription options at http://sigcis.org/mailman/listinfo/members
>This email is relayed from members at sigcis.org, the email discussion list of SHOT SIGCIS. The list archives are at http://sigcis.org/pipermail/members/ and you can change your subscription options athttp://sigcis.org/mailman/listinfo/members
Paul N. Edwards
Professor of Information and History, University of Michigan
A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (MIT Press, 2010)
Terse replies are deliberate (and better than nothing)
University of Michigan School of Information
3439 North Quad
105 S. State Street
Ann Arbor, MI 48109-1285
(734) 764-2617 (office)
(206) 337-1523 (fax)
This email is relayed from members at sigcis.org, the email discussion list of SHOT SIGCIS. The list archives are at http://sigcis.org/pipermail/members/ and you can change your subscription options at http://sigcis.org/mailman/listinfo/members
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Members