A Letter to the Editor

January 11, 2010
Alan Perlis

Alan J. Perlis

I had planned to write a post later this spring on the collisions between what engineers sometimes perceive as practical and what turns out in practice to be useful.  It’s a complex issue and there are examples that cut both ways, suggesting that a deeper understanding of both the underlying technology and the social “soup” where innovators thrive are needed to avoid some famous traps.  I mentioned this briefly in my discussions of the impact of social fragmentation on innovation and the pitfalls of ignoring social contexts.

Then the January 2010 issue of Communications of the ACM crossed my desk.  As I skimmed the contents, I was surprised to see my name in the headline of the Editor’s  Letter, an attack by the Editor-in-Chief Moshe Vardi on a thirty-year-old paper [ Social Processes ] that I wrote with computing legend Alan J. Perlis and my colleague Richard J. Lipton (author of the popular Godel’s Lost Letter blog and subject of Dancing with the Stars ).  The paper itself was controversial in its day and addresses exactly the WWC questions that I plan to write about.  It is extraordinary for an Editor of a professional journal to use his position to make derogatory comments about articles, especially to  further his own views.  Mr.Vardi’s letter demanded a response.  Lipton and I will jointly publish a longer and more technical essay on this subject at some point in the future, but today we are jointly publishing the following Letter to the Editor. The letter will also be sent to the Communications of the ACM.

In his  Editor’s Letter in the January 2010 issue of CACM entitled “More Debate Please”,  Moshe Vardi makes a plea for controversial topics in these pages, citing a desire to “let truth emerge from vigorous debate.”  It is a sentiment that we support as well. But we question Mr. Vardi’s judgment in using his editorial position to mount an attack on colleagues who were neither forewarned nor given an opportunity to respond.  Mr.  Vardi’s target was  our 1979 critique of formal program verification entitled  “Social Processes and Proofs of Theorems and Programs,”  It was co-authored with the late Alan Perlis, one of the originators of the field and a lifelong advocate for the kind of open discussion that the Editor advocates.  We can only hope that future contributors have higher standards for debate than does the current Editor, because his out-of-context references to the 1979 debate over the practical efficacy of formal verification, his ex-cathedra determination that the article was “misguided” and his ill-informed view of the decision to publish it have no power to illuminate  a serious subject.

We do not care to respond to Mr. Vardi’s substantial mischaracterizations and misstatements at this time, but we do think it is fair to point out that  the publication of “Social Processes and Proofs of Theorems and Programs,” was not a singular event that might be classified as either misguided or not.  “Social Processes” was a refereed article.  A preliminary version was accepted  by a highly selective conference program committee in 1976 and its presentation was attended by virtually every living contributor to the field.  It was then submitted to this journal and reviewed by anonymous referees. Its publication was followed by many months of public presentations and workshops, letters to the Editor, written reinforcements and rebuttals, and — several years later — a special issue of this journal devoted to the topic.  Mr. Vardi faults the editorial board for not publishing an opposing “counterpoint” article, a suggestion that — although it has all the “fair and balanced” trappings — would have been hard to reconcile with the confidentiality usually accorded to contributed articles that are sent to referees for review. The irony is not be lost on us  that we were offered no such opportunity prior to publication of his letter.

The article itself has been reprinted dozens of times and has appeared in several anthologies in the philosophy of mathematics.  Its publication and the ensuing debate have been the subject of social science research (Donald MacKenzie’s 2001 book[1] “Mechanizing Proof” remains the definitive sociological and historical analysis of both the paper and its implications for the field). If our arguments seem off the mark to Mr. Vardi, then perhaps the right course of action is to resurrect the social process that led to the article’s publication in the first place and jump into the fray. Until that time, the correct editorial position for CACM and its Editor is to let both the paper and the written record that surrounds it speak for themselves.  It strikes us as inappropriate, after thirty years of silence,  to use the cover of an Editorship to  attack unsuspecting passersby, especially while touting the moral virtues of free and vigorous debate.

[1] Donald MacKenzie, Mechanizing Proof: Computing, Risk, and Trust, MIT Press 2001, The Massachusetts Institute of Technology, Cambridge, MA


10 Responses to “A Letter to the Editor”

  1. Moshe Vardi Says:

    It seems that both Rich DeMillo and Richard Lipton feel slighted by my January 2010 Communications of ACM editorial. I had no intention of slighting them or the paper in question, and I apologize for unintentionally causing them to feel this way.

    Let me address the substantive points in their letter:

    1. I am accused of using my “editorial position to mount an attack on colleagues who were neither forewarned nor given an opportunity to respond.”

    The paper in question is over 30 years old. History it is said, “judges and rejudges.” I hardly view my offering some comments on such a historically important paper as a “personal attack” on its authors. Personally, if someone saw the need to disagree with a paper of mine 30 years after its publication, I’d feel complimented! Most papers are long forgotten after 30 years.

    2. I am accused of not offering Messrs. Demillo and Lipton in question an opportunity to respond prior to publication of editorial letter.

    As Editor-in-Chief of Communications I write bimonthly editorial letters, in which I often express opinions on controversial matters. The proper way to disagree with my editorials, and many people often disagree with my editorials, is to leave comments online or to submit a letter-to-the-editor. This is the standard operating procedure in all publications I am aware of.

    3. It seems that Messrs. DeMillo and Lipton were offended by my usage of the word “misguided”.

    One should read, however, the full context of that word: “With hindsight of 30 years, it seems that DeMillo, Lipton, and Perlis’ article has proven to be rather misguided. In fact, it is interesting to read it now and see how arguments that seemed so compelling in 1979 seem so off the mark today.”

    In the paragraph that preceded that comment I referred to two Turing Awards that were given for works in formal verification. For lack of space, I did not include references to two to ACM Kanellakis Awards and two ACM Software System Awards for works in formal verification. It is in this context that I expressed an opinion that the 1979 article, which implied the futility of formal verification, was “misguided”, with “hindsight of 30 years”, in spite of “its compelling arguments”.

    4. Messrs. DeMillo and Mr. Lipton disagree with my opinion that “the editors of Communications in 1979 did err in publishing an article that can fairly be described as tendentious without publishing a counterpoint article in the same issue.”

    I am actually well aware of the process that led to the publication of the 1979 article. I stand behind my opinion about the lack of counterpoint article. Messrs. DeMillo and Lipton are entitled to a different opinion. We may need to agree to disagree on this one. I do not see why this is an issue that deserves such a strongly worded response, when I expressed strong support for the editorial decision to publish the article!

    5. In a blog comment, Mr. DeMillo said: “On the other hand, singling out three authors out of the thousands who have published in CACM sure looks like an attack.”

    The subject of my editorial was “More Debate, Please! The article in question is one of the most controversial and influential articles ever published in Communications. I read it as a graduate student and was deeply impacted by it. I singled this article out because it was the perfect example to make the point of my editorial.

    6. I’d welcome a new article from Mr. Demillo and Mr. Lipton examining the issues covered in their 1979 article. Of course, I’d seek to publish a counterpoint article in the same issue!

    Moshe Vardi

  2. richde Says:

    I certainly did not take it as a personal attack, and said so in my letter: “Mr. Vardi’s target was our 1979 critique of formal program verification…”

    Lipton and I will respond more fully, but the one thing that remains certain is that CACM has now taken an editorial position that a previously published paper is “misguided.” I am personally grateful for the new publicity, but you have to admit that it’s an extraordinary thing to do. I’ve searched my CACM archives, and I can find nothing that even comes close.

    Without getting into the merits of your argument, I think it is unwise for an editor to do what what you have done. If Professor Vardi (as opposed to Editor Vardi) wanted to resurrect this argument, then there are dozens of other venues that would have been more appropriate.

  3. Moshe Vardi Says:

    Are you saying that no editor-in-chief can ever comment on a paper published 30 years earlier? Even the Supreme Court occasionally reverses its opinion. I never heard of “Stare Decisis” being applied to editorial matters across such a time span. (In contrast, I did commit to respecting all prior editorial decisions in regard to pending submissions of Communications.)

    As EiC, I am committed to a scrupulous peer review process for submitted articles, but I have not taken a vow of silence, no does it make sense for me to do so. Furthermore, I give full permission to the EiC of Communications in 2040 to re-examine my editorial decisions.

  4. richde Says:

    No I am not saying that. The EIC is clearly free to comment as he/she sees fit. The vast majority of editors exercise the utmost discretion in using that authority however. I am saying it is unwise, particularly when, as in your case, the EIC has a dog in the hunt.

    A side issue: I might take issue with the extent to which your remarks actually were “comment” versus, say, “decree” since you offered no reasoning to shore up your opinion.

  5. Moshe Vardi Says:

    Just recall that analyzing your article was not the main point of my editorial. The main point was that, in my opinion, even with 30-years hindsight, the editors in 1979 did absolutely the right thing in publishing the article. As you may also recall, many people have argued over the years otherwise.

    Let’s do a Gedanken experiment. Suppose that your article had appeared in 1979 is IEEE Computer, rather than Communications of ACM. Would have also said then that its is inappropriate for me to comment on the article?

  6. richde Says:

    It is a shame that a few stray words and phrases blotted out the sun in what otherwise would have been a very welcome message.

    Re the Gedanken experiment, I’m not sure what the point would be. As EIC you have a special platform and speak for many interests. I prefer a more measured view of that responsibility.

  7. Moshe Vardi Says:

    It is obvious to me from the comments on Lipton’s blog that many readers did not view my editorial as an attack on your 1979 paper, but as a reflection on the importance of open debate. It is precisely because the paper was so influential that I chose it as an example. I honestly feel that you and Dick should be pleased that the paper is still trenchant, even if people disagree with parts of it.

  8. I note that it is possible for people to make notable strides towards impossible goals. That Turing Awards and others have been given to people for work in verification does not mean that it is a definitive method of program development or will be important in the future. A quick review of the Nobel prizes shows one in physics in 1912 for the invention of gas valves for use in buoys — a key issue today. 🙂 The Nobel Peace prizes have been awarded multiple times without Peace ever being achieved. Rutherford & Black were celebrated for the discovery of nitrogen, but they used it as proof of the theory of phlogiston. And there are many other examples that can be found, of notable scientists who received awards from their peers who all believed something was “true” that we now know was not — the 4 “humours” of the body, geocentricity, the “ether,” and so on.

    It is important for the science of computing that we continue to question accepted results in a measured, careful and reproducible manner. Results by distinguished, senior scientists should not be accepted simply because of who stated them and how many awards have been bestowed as a result; if anything, we should keep in mind (Arthur C.) Clarke’s First Law: “When a distinguished but elderly scientist states that something is possible he is almost certainly right. When he states that something is impossible, he is very probably wrong.” Science is based on theory that should always be subject to challenge or it is no longer science (think Trofim Lysenko and what happened to Soviet biosciences when challenge was no longer allowed, or the Inquisition for those who dared question Biblical inerrancy).

    In the case in point, I remember reading the original, the followups, and the many discussions that followed — both in my CS classes and in a philosophy seminar! I have yet to see a clear refutation of the conclusions of that first article that convinces me, as an informed reader. Perhaps I missed something published, or I misunderstood the original paper, so I may be “misguided.” But without something to sway my thinking, I still accept those results.

    As we consider our field as leaders and influencers (as both DeMillo and Vardi certainly are, and I may be to a lesser extent), we should always be willing to reevaluate results — our own and those of others. But we should carefully separate those evaluations into “personal opinion” and “professional conclusion” with the latter supported by evidence, peer review, and scientific procedures. We are each entitled to our opinions, but we should be cautious to label them as such — opinion and faith are not foundations for good science, and are more the province of politics and religion…which we mix with science at our peril.

    And to me, an editorial is always an expression of opinion. As are most blog posts. But that’s simply my opinion. 🙂

    • Moshe Vardi Says:

      I am pleased to see that the focus has shifted to the substance of my opinion, rather than to the legitimacy of my expression of it. That is the right focus, IMHO. I have invited Rich DeMillo and Dick Lipton to argue for their point of view on the pages of CACM.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: