[SIGCIS-Members] My CACM column on "The Tears of Donald Knuth"

Bjorn Westergard bjornw at gmail.com
Thu Jan 15 08:19:43 PST 2015


"It’s not clear me that that gaining competence in some historical area of
computer science (or, more broadly, technology) is inherently harder than
those other things, just that there is no structured way to do so or
rational reason to attempt the feat."

If historical actor A claims their approach to packet-switching was an
obvious or natural continuation of the approach of historical actor B,
should the historian relay that opinion and leave it at that? It seems like
the historian is obligated to see the development as a contemporary
practitioner standing at some remove from that problem might have seen it
if they're to weigh in on questions of priority, causality (re: influence),
peer esteem, etc.

I write software for a living, and sometimes imagine how a social scientist
would write about my peer group. If you were to read our emails/chat logs
with no background in software engineering, or listen to recordings of our
conversations, you might be lead to believe that a disagreement about,
e.g., which library to use for time conversion is a matter of weighing
fairly objective technical criteria in particular circumstances with
imperfect knowledge.

But if you were familiar with the technical domain, I think you'd realize
that that argument in particular, and many others besides, are in fact
"about" petty personal conflicts, defending the importance of one's
particular competences over the claims to superiority of others (FOSS v.
Microsoft), job security through obfuscation, etc. It would be hard to see
the "social" if you're not immersed in the "technical".


On Thu, Jan 15, 2015 at 12:25 AM, Thomas Haigh <thaigh at computer.org> wrote:

> Those are good points raised by Janet and by several people in response to
> her. It’s true that I did not emphasize the inherent difficulty for trained
> historians of trying to grasp enough computer science to write technical
> histories. That is unquestionably another reason that such history is
> rarely written, and supports my belief that if such history is produced it
> will probably be by computer scientists.
>
>
>
> Pondering this, I think there are two reasons that I underweighted this
> point. One is that I wanted to emphasize that there are few incentives or
> opportunities for trained historians to take such an approach even if they
> have the skills to do so. Knuth’s “dumbing down” frame, which I challenged,
> implied that non-technical history was easier to write, less challenging
> for readers, and less valuable than deep technical history. Focusing on the
> difficulty historians would face in equipping themselves to write technical
> history might have confirmed, in the minds of CACM readers, the idea that
> historians write the dumb kinds of history of computing because they are
> unable to handle the smart kind.
>
>
>
> The other reason is that I think the issue is at root more
> social/structural than intellectual. I’m constantly impressed by the
> obscure competencies that scholars develop – knowledge of foreign or
> ancient languages, far-off time periods, the paints and materials used by
> Old Masters, the ability to interpret shards of excavated pottery, etc.
> Nabeel’s points seem to support this. It’s not clear me that that gaining
> competence in some historical area of computer science (or, more broadly,
> technology) is inherently harder than those other things, just that there
> is no structured way to do so or rational reason to attempt the feat.
>
>
>
> Shifting gears slightly, one of the thoughts I had when editing Mike
> Mahoney’s collected papers was that the history of computer science is
> challenging because the newness of the discipline means it arose after the
> history of science professionalized. We no longer value the production of
> vast, plodding, and comprehensive technical, institutional, or intellectual
> histories. If those are never produced in the first place it is hard to
> write the kind of history that would win favor. How are historians supposed
> to get a grounding on basic, prosaic questions like what were the key
> research areas in computer science each decade? Who were the key
> researchers? What did they discover/invent, and why did it matter?  Here is
> what I wrote in the introduction to *Histories of Computing* (Harvard,
> 2011):
>
>
>
> Kuhn’s great theme was the cognitive revolution in science. He insisted
> that the history found in scientific textbooks was distorted by presenting
> the work of earlier scientists within present day conceptual frameworks and
> disciplinary traditions. The task of the historian was to challenge this
> and reconstruct the alien subjective world in which their work actually
> took place. Mahoney approaches the history of computer science in a similar
> way, influenced also by his own work on the scientific revolution. He
> writes as if he has in mind an audience for whom the broad outline of early
> computer science, its theorists, ideas, and milestones, are already
> familiar. Much is alluded to rather than explained. Instead Mahoney works
> to challenge our presumed assumptions about how all these things fit
> together, disrupting and rearranging a received narrative that has, alas,
> never been written in the first place (though it was surely part of the
> folklore of the next generation of researchers). This is, of course, very
> much the way in which people write about early modern science, in which men
> such as Newton and Galileo have been studied for centuries and yet continue
> to provide material for generation after generation of historians. It also
> reflects a style of teaching found in graduate history of science seminars
> at Princeton, in which students are assigned substantial mounds of
> historical scientific papers and other primary sources for each week of
> discussion and expected to grapple for themselves with their complexities
> (or learn to skim, which Mahoney called “reading aggressively,” and bluff
> convincingly).
>
>
>
> His papers on the history of computer science are unmistakable
> challenging. If, as I have suggested, he wrote them for an audience that
> does not yet exist then we may be unable to fully evaluate them for several
> decades until a conventional history of computer science, whose assumptions
> Mahoney sought to challenge, has finally been assembled.
>
>
>
> Best wishes,
>
>
>
> Tom
>
>
>
> *From:* members-bounces at sigcis.org [mailto:members-bounces at sigcis.org] *On
> Behalf Of *Nabeel Siddiqui
> *Sent:* Wednesday, January 14, 2015 7:00 PM
> *To:* members
> *Subject:* Re: [SIGCIS-Members] My CACM column on "The Tears of Donald
> Knuth"
>
>
>
> Dear all,
>
>
>
> Janet beat me to it, but I had a similar reaction upon reading the
> article.  I believe that two additional factors will lead to the technical
> history of computer science diminishing even further.
>
>
>
> First, it is nearly impossible for a graduate student to currently spend
> time gaining a technical understanding of computer science and still stay
> competitive for jobs in a humanities field.  As Tom noted, there are no
> jobs available for a student that wanted to do this kind of work. I would
> take the technical history of computer science to be, at best, suited for
> an intellectual history position.  Intellectual history itself has eschewed
> this type of approach and now heavily focuses on social history. There,
> even a student trained in intellectual history would have a hard time
> gaining a foothold in the academy.  A student would likely have to learn
> the technical literature on their own unless they were specifically
> admitted to a program to study the history of computers. In my own
> interdisciplinary program of American Studies, I have had the opportunity
> to take a great deal of independent studies, but I have still had a
> difficult time taking a course outside of humanities departments.
>
>
>
> Two, while this listserv is an exception, few historians of science are
> trained in computer science in any significant way.  The few that are
> usually have their undergraduate degrees in computer science rather than
> vice versa.  While some of these people have been able to dable in both
> history and computer science, I would say it is getting rarer with newer
> students.  In fact, I have found that many students of computer science now
> take more courses in "marketable" skills like programming and networking.
> While these contain a technical component, I don't know how many in the
> future would have the training necessary to understand old technical
> documents that may contain approaches that are outdated or unnecessary for
> their own work.
>
>
>
> I would love to see some fellowships or opportunities for humanities
> graduate students to counter this.  In my own research on personal
> computers, I come through a great deal of articles dealing with technical
> topics in early hobbyist magazines.  Unfortunately, because these hobbyist
> magazines assume that their readers are other engineers, things like
> schematics, code, etc. is often lost on me.  I don't believe a historian
> necessarily needs to know this to gain the gist of the articles, but it
> would be helpful.  Business historians have been doing this for a while
> with workshops that teach economics and finance basics.  Digital humanities
> scholars have also been at the forefront of humanities oriented technical
> training with workshops on programming and statistics.
>
>
>
> Sincerely,
>
> Nabeel
>
>
>
> On Wed, Jan 14, 2015 at 6:37 PM, James Cortada <jcortada at umn.edu> wrote:
>
> Which all begs the question, have we now reached a critical mass in the
> teaching and study of computing's history that we need to develop a
> blueprint, format, curriculum, not sure of the term I am grasping for, of
> what an historian needs to know about information technology itself,
> perhaps borrowing from the course guidelines produced by such organizations
> as the Computer Society and the ACM?  Sounds like a great project to create
> such a core body of info for historians funded by the Sloan Foundation.
>
>
>
> Maybe the history committee at the Computer Society should consider taking
> this on with the collaboration of several IT historians.
>
>
>
> On Wed, Jan 14, 2015 at 5:14 PM, Brian Randell <
> brian.randell at newcastle.ac.uk> wrote:
>
> Hi Janet:
>
> Well said!
>
> However, as a "computer scientist not trained in history”, let me claim
> that at least for my main area of CS research, namely "system
> dependability”, a number of us put a lot of effort over a number of years
> into identifying and explaining the concepts involved - see
>
> Avizienis A, Laprie J-C, Randell B, Landwehr C. Basic concepts and
> taxonomy of dependable and secure computing. IEEE Transactions on
> Dependable and Secure Computing 2004, 1(1), 11-33.
>
> But I must admit that computer scientists were our intended audience, not
> historians :-)
>
> And though we got a lot of agreement in our immediate community, I cannot
> claim that we persuaded everybody to use our recommended set of concept
> definitions!
>
> Cheers
>
> Brian
>
>
>
>
> On 14 Jan 2015, at 17:54, Janet Abbate <abbate at vt.edu> wrote:
>
> > Tom does a good job of spelling out the problem of audience (or market,
> in institutional terms) for particular kinds of history. It's extremely
> encouraging that the column has received so much attention; perhaps it will
> even inspire some historical work by computer scientists.
> >
> > One important issue that Tom did not address is the difficulty for a
> historian not trained in CS to achieve a sold technical understanding of
> computer science concepts. Part of the problem is that in many subfields,
> computer scientists themselves have not made public an agreed-on set of
> major concepts that the historian could then set about trying to learn.
> This information is buried in the technical literature, and it's difficult
> for a non-specialist to tease out. We need better roadmaps of the field if
> we want better technical history.
> >
> > This makes me think that collaboration between historians and computer
> scientists will be necessary in order to move the history of computer
> science (rather than technology) forward. I hope to make some moves in that
> direction myself.
> >
> > best,
> > Janet
> >
> >
> > On Jan 12, 2015, at 2:57 47PM, Thomas Haigh wrote:
> >
> >> Hello Everyone,
> >>
> >> Thought you might be interested in my latest "Historical Reflections"
> column
> >> in Communications of the ACM, titled "The Tears of Donald Knuth: Has the
> >> History of Computing Taken a Tragic Turn?" (I thought about calling it
> "Flow
> >> My Tears, The Computer Scientist Said," in honor of a 1970s Philip K.
> Dick
> >> novel).
> >>
> http://cacm.acm.org/magazines/2015/1/181633-the-tears-of-donald-knuth/fullte
> >> xt
> >>
> >> The piece builds on SIGCIS list discussion from last summer, when the
> video
> >> of Knuth's talk, "Let's Not Dumb Down the History of Computer Science"
> was
> >> posted online (https://www.youtube.com/watch?v=gAXdDEQveKw). Knuth's
> talk
> >> was, in turn, centered on his distress at reading Martin
> Campbell-Kelly's
> >> paper "The History of the History of Software," which he saw as
> celebrating
> >> a regrettable shift away from technical history of computing.
> Campbell-Kelly
> >> recently published his own response to Knuth in IEEE Annals:
> >> http://www.computer.org/csdl/mags/an/2014/03/man2014030096.pdf.
> >>
> >> As a Ph.D. historian my natural sympathies are with the turn towards
> broader
> >> historical approaches praised by Campbell-Kelly, but I have also come to
> >> appreciate the value and rigor of more technical kinds of history
> (including
> >> early programming tools and techniques by both Knuth and
> Campbell-Kelly). So
> >> I attempted to be fair to both perspectives, while stressing that the
> >> history of computer science accounts for a quite small proportion of the
> >> work now being done on the history of computing. I also argue the
> realities
> >> of academic society mean that the kinds of history favored by Knuth will
> >> only flourish if computer scientists themselves make a significant
> >> investment in historical work. Finally I took the opportunity to cite
> some
> >> outstanding work on the history of computer science and to publicize
> some
> >> new initiatives and scholars in the area.
> >>
> >> Such is the allure of Knuth's name that the column was seized on by tech
> >> aggregation sites such as Slashdot, Hacker News, and Reddit when it
> first
> >> appeared online over the Christmas break. It's also been tweeted and
> >> blogged, and currently has about 75,000 downloads - more than the all
> the
> >> other articles in the January CACM combined. I can't say that the online
> >> discussions were particularly well informed, although they do include
> some
> >> thoughtful comments along with sentiments such as "Gawd academia
> disgusts
> >> me. These "historians" would be more useful to society if they dug cans
> and
> >> bottles out of the trash." Anything that makes more people think about
> the
> >> history of computing and its purpose is a good thing for the field.
> >>
> >> I'd like to thank those who commented on drafts of the column,
> including Len
> >> Shustek, Dave Walden, David Hemmendinger, Bill Aspray, Chuck House, and
> Paul
> >> Fishwick. Not all of them agreed with all of my points, but I did find
> their
> >> input very useful in sharpening the argument and presenting my thoughts
> in a
> >> way less likely to unnecessarily offend computer scientists.
> >>
> >> Best wishes for the New Year,
> >>
> >> Tom
> >>
> >> _______________________________________________
> >> This email is relayed from members at sigcis.org, the email discussion
> list of SHOT SIGCIS. Opinions expressed here are those of the member
> posting and are not reviewed, edited, or endorsed by 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. Opinions expressed here are those of the member
> posting and are not reviewed, edited, or endorsed by 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
>
> --
> School of Computing Science, Newcastle University, Newcastle upon Tyne,
> NE1 7RU, UK
> EMAIL = Brian.Randell at ncl.ac.uk   PHONE = +44 191 208 7923
> FAX = +44 191 208 8232  URL = http://www.cs.ncl.ac.uk/people/brian.randell
>
>
>
>
>
> _______________________________________________
> This email is relayed from members at sigcis.org, the email discussion list
> of SHOT SIGCIS. Opinions expressed here are those of the member posting and
> are not reviewed, edited, or endorsed by 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
>
>
>
>
>
> --
>
> James W. Cortada
>
> Senior Research Fellow
>
> Charles Babbage Institute
>
> University of Minnesota
>
> jcortada at umn.edu
>
> 608-274-6382
>
>
> _______________________________________________
> This email is relayed from members at sigcis.org, the email discussion list
> of SHOT SIGCIS. Opinions expressed here are those of the member posting and
> are not reviewed, edited, or endorsed by 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. Opinions expressed here are those of the member posting and
> are not reviewed, edited, or endorsed by 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...
URL: <http://lists.sigcis.org/pipermail/members-sigcis.org/attachments/20150115/2093fb50/attachment-0001.htm>


More information about the Members mailing list