Opinion
Computing Applications Letters to the Editor

How to Claim Your Fair Share in Academic Publishing

Posted
  1. Introduction
  2. What an Algorithm Is, and Is Not
  3. Accounting for Death in Guns
  4. Author's Response
  5. Scenario Approach for Ethical Dilemmas
  6. References
  7. Footnotes
Letters to the Editor

In his editor’s letter "To Boycott or Not to Boycott" (Mar. 2013), Moshe Y. Vardi said the traditional author partnership with commercial publishers has turned into an abusive relationship. It is time computer scientists broke off that relationship, though not by boycotting, politicking, moralizing, or shedding tears, but rather by asking a fair price for the work we do as authors and as reviewers. With every niche activity organized as a profit center today, core contributions like authoring and reviewing should no longer be provided for free to for-profit publishers. The fees we ask for our contribution to publishing should increase with the price we pay for the respective journal. Moreover, the fees should go to the authors’ institutions’ libraries, not to the authors and reviewers directly. We must collectively think in terms of market forces, not giveaways.

Andreas Siebert, Landshut, Germany

Moshe Y. Vardi said, "I believe in keeping science separate from politics" (Mar. 2013) but may well have proclaimed, long after Napoleonic infantry tactics became impractical, "I believe we must fire only from the field, in square or in line."

Is not the form of battle determined by the nature of the obstacles, as well as the nature of those who carry it forward and of the opponent’s objectives?

That we may pressure them to reform their business practices, Vardi urged us to confront publishers "the old-fashioned way, by out-publishing them." I agree. Out-publish them… with robust vitality.

However, when attacked indirectly, asymmetrically, as when publishers lobby the sources of research funds, CS authors and editors alike must respond in kind. Two elements of such a response are those Vardi found disquieting: a boycott and influencing one’s peers. If among their opponents are colluding corporatized research publishers, then authors must expect to have their actions, along with their strength and resolve, tested on many fronts.

Editors of influential publications cannot waiver at first fire. Rather than ask, "Where shall we draw the line?," we should expect thoughtful, effective, farsighted leadership.

If authors wish the battle to be returned (and remain with) content, their response must be quick and unified. However, they might also find publishers are not the sole impediment and so must also weigh their readers and their products.

CS authors’ first few steps toward discovery are short, starting with a look in a mirror to give themselves understanding by first donning their institutional regalia.

Nick Ragouzis, San Francisco, CA

Back to Top

What an Algorithm Is, and Is Not

Addressing the question "What Is an Algorithm?" in his Editor’s Letter (Mar. 2012), Moshe Y. Vardi wrote that there is no consensus, suggesting the two abstract approaches of Yuri Gurevich1 and Yiannis N. Moscovachis2 were both correct; algorithms are the abstract state machines of Gurevich and the recursors of Moscovachis. This is (as Moscovachis said) like trying to explain that the number 2 is the set {Ø, {Ø}}, which is likely to leave an outsider less, not more, enlightened.

A proper definition would allow computer scientists to talk to the public about algorithms. The pragmatic motivation is to justify the treatment we informally apply to algorithms: construct, explain, and exchange; admire for elegance; debate whether one is the same as another; marvel that the one for the program Eliza is so simple; patent or sell them; or preclude protections based on principled argument. The theoretical motivation is to probe relationships to programs and abstract machines and promote discussion of semantics and implementation.

Now consider this definition: An algorithm is an abstract deterministic control structure accomplishing a given task under given conditions, expressed in a finite, imperative form. A critical look would reveal no formalities but plenty to consider, along with some surprises.

The definition includes compass-and-straight-edge algorithms (such as to bisect an angle) and other nonelectronic but straightforward sets of instructions. Excluded are recipes, which are not deterministic except in perversely rigorous cases. Also excluded are games, which, construed as a single agent’s instructions, are not imperative control structures that accomplish a given task (with certainty) and which, construed as interchanges of moves, are not deterministic.

Perhaps most controversial, it also excludes recursive definitions, which are not imperative but declarative. A definition of the factorial function with a base case and a recursive case does not undertake a computation of a factorial or provide directions to do so. An algorithm must "do" not "be." (Gurevich noted that one possible algorithm generally available through such a definition starts with "Apply the equations…"; that is, the algorithm tells us the obvious thing to do with the definition.) The imperative requirement yields the interesting result that computers, at the hardware level, do not execute algorithms. Current flowing through circuits does not express an imperative. Algorithms are a human construct.

So, do all programs implement algorithms? People are the interpreters of the elements of the definition, responsible for measuring the concept against them. But how do we map the path from the human view to the mathematical view? Gurevich and Moscovachis each suggested the rich connotations of the concept of the algorithm cannot be captured fully through a recurrence relation or other set-theoretic object. We can honor their contributions by pursuing the questions raised by the definition I have outlined here or through a competing informal view.

Robin K. Hill, Laramie, WY

Back to Top

Accounting for Death in Guns

Massacre by shooting is uncommon. A movie theater showing a Batman movie, Columbine High School, Fort Hood, Sandy Hook Elementary School, and Olso, Norway, together claimed 240 casualties, dead and wounded. There are undoubtedly more, and I encourage you to find every relevant mass murder committed with a gun over the past 15 years and tally the casualties. Now find the annual deaths in the U.S. attributed to, say, bicycling, choking by children 14 and younger, food poisoning, drowning, or motorcycling. The U.S. Centers for Disease Control and Prevention maintains metrics for most of them (http://remstate.com/blog/2013/04/gun-violence-facts-and-citations.html). Each year, any of these causes claims more lives than massacres involving guns. Tailoring specific solutions to these statistically insignificant occurrences is an ineffective expenditure of both resources and rights, regardless of how tragic the individual cases are. However, gun-related deaths in general are on par with total U.S. automobile-related deaths per year, ~30,000, suggesting we are able to make progress on the larger problem.

Suicide is a top cause of death in the U.S., with more than 38,000 per year, about half involving guns, accounting for two-thirds of annual gun-related deaths in the U.S. "Fixing" guns, as Jeff Johnson discussed in his Viewpoint "Can Computer Professionals and Digital Technology Engineers Help Reduce Gun Violence?" (Mar. 2013), will not fix suicide. There might be some improvement (guns are convenient and quick), but identifying and helping depressed people would be more effective.

The ~11,000 annual gun-related homicides in the U.S. (of ~16,000 homicides total) account for the final one-third of gun-related deaths. Note the CDC includes all homicides, criminal and noncriminal. The FBI provides numbers that help separate the "bad" incidents from the "good"; for example, it appears that ~5.6% of all justified homicides are noncriminal, and virtually all justifiable homicides turn out to involve guns. Justified shootings reflect a 60%/40% split between those committed by police and those by private citizens.

Good or bad, guns work as advertised—as lethal weapons.

Attempting to use technology to reduce gun lethality is destined to follow the same path of failure as digital rights management technology, with law-abiding gun owners risking failure of their weapons in potentially life-threatening situations, while criminals jailbreak theirs to sidestep "safety restrictions" intended to prevent crime.

I have left out the 606 annual accidental deaths (2010) in the U.S. involving guns. Amazingly, more people die in bicycle accidents per year (~700 in the U.S.) than by accidental shooting, speaking well for the state of gun safety today.

Consider the following three suggestions for reducing violent deaths in the U.S. (including those involving guns): help medical professionals detect and/or treat underlying causes (such as depression); apply that knowledge to keep guns away from at-risk people; and eliminate the rock star ideal (nonstop work, hitting perpetually unreasonable deadlines, monotonic achievement) from hightech culture to reduce stress-induced depression and suicide.

Travis Snoozy, Seattle, WA

Back to Top

Author’s Response

Comparing frequencies of intentional crimes and accidents is nonsensical. Injuries from fights in schools happen less often than injuries from playground accidents. Should we therefore allow fights? Bombings in the U.S. are rare compared to car accidents. Should we therefore accept bombings? We put effort where potential leverage is greatest, and we have more leverage over intentional crimes than over accidents. Keep guns away from "at-risk" people through universal background checks and waiting periods. Digital rights management technology has failed? That’s news to me.

Jeff Johnson, San Francisco, CA

Back to Top

Scenario Approach for Ethical Dilemmas

I wrote and won approval for the first ACM Professional Guidelines, which evolved into the ACM Code of Ethics in the 1970s. This led to my obtaining two National Science Foundation grants from the Office of Science and Society, Science Education Directorate, Ethics and Values in Science and Technology Program. The grants helped me hold two ACM ethics workshops in 1977 and 1987, respectively, resulting in two books, Ethical Conflicts in Computer Science and Technology, AFIPS Press, 1981, and Ethical Conflicts in Information and Computer Science, Technology, and Business, QED Information Sciences, 1990.

I strongly support Rachelle Hollander’s scenario approach to ethics explored in her Viewpoint "Ethics Viewpoints Efficacies" (Mar. 2013), finding it useful and revealing in the two ACM ethics workshops. I invited CS opinion leaders to discuss, evaluate, and vote "unethical" or "not unethical" on almost 100 ethical-conflict scenarios. The scenarios came from my earlier NSF-funded studies of cases of computer abuse and misuse. It was fascinating to see how ethical values changed from 1978 to 1988. Ponder how they have changed since.

Donn B. Parker, Los Altos, CA

Back to Top

Back to Top

Join the Discussion (0)

Become a Member or Sign In to Post a Comment

The Latest from CACM

Shape the Future of Computing

ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.

Get Involved

Communications of the ACM (CACM) is now a fully Open Access publication.

By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.

Learn More