The Community for Technology Leaders


Pages: pp. 5-6

Taming Chaos

I often find myself disagreeing with Neville Holmes's opinions in his articles. The disagreement, I think, stems from a simple difference of personality. Neville seems to prefer structure and organized taxonomies. He often advocates that we need more precise, prescriptive definitions for our technical terms. As for myself, I'm much more of a laissez-faire kind of guy—let a thousand flowers bloom and so forth. Knowing that we maintain two opposing world views, I just casually disagree with much of what Holmes writes but look for the points that do interest me.

So when he claimed in the The Profession column in Computer's September issue ("The Internet, the Web, and the Chaos," pp. 108, 106-107) that we need to tame the chaos that is the World Wide Web by cleaning up the terminology and imposing a clear structure on the Web, I said to myself, "Ah, that's just Neville being Neville." What got me, though, is that in the very same column he pines for the good old open market bazaar of butchers, bakers, and candlestick makers over the corporate supermarkets that we have today. It left me asking, "Well, which way do you really want it?"

Michael Hobbs;

Neville Holmes replies:

The structural organization of butchers, bakers, and candlestick makers is a very good model for what I feel should be happening with the Web.

You see, if I go to a supermarket because I need something, and the butcher and delicatessen happen to be closed, I have two problems.

The first is that I have to spend a lot of time finding where what I want might be. For example, the kind of soy milk I prefer is not kept with other milks, or even with other soy milks, and the staff usually doesn't know where it is, and there is no map available. And different supermarkets keep it, if they do at all, in different places.



The second is that the supermarket often doesn't stock the product I want or it is temporarily out of stock. Thus, if I want to buy, say, filets mignon, I have to make sure I get to the local butcher before he shuts down for the day. If I do get there in time, he asks me how thick I want the filets, and he cuts them on the spot.

In the same way, I'd like the Web to be organized into separate components rather than one big kludge. It's a joy to be able to get such a great variety of data from the Web, but it's a tragedy that it's so hard to get to or get back to just what you want, so often finding that your old links are broken.

The separate top-level components the Internet needs right now, given the international bickering over Internet control, is a divorcing of Internet governance from Web governance, in particular the setting up of separate domain name systems for each of the world's major writing systems.

Professional disputes

I empathize with Dimitris Kalles's disputes, if I may call them that, with the referees of professional journals and conferences (The Profession, "Improving Professional Conduct in Publishing," Oct. 2005, pp. 116, 114-115). I would like to commend him for taking a proactive approach by trying to engage in a dialog with the referees and editors.

A paper I submitted a few years ago was rejected by a referee citing that an example mentioned in the paper was incorrect. It turns out that the example was given in almost every standard textbook at that time and still is. In other words, the referee was incorrect, and if the referee had been familiar with at least the standard textbooks, it might have resulted in a different outcome. Disenchanted, unlike Kalles, I opted out of academia where publishing is an implicit requirement in most cases.

Neither of the incidents Kalles mentions indicates any ignorance on the part of any of the referees. Hopefully, the review committees currently have fewer misinformed reviewers.

Raghavendra Rao Loka;

Should we fix it?

As the oldest living reviewer of computing stuff I feel impelled to reply to Dimitris Kalles's comments in "Improving Professional Conduct in Publishing" (The Profession, Oct. 2005, pp. 116, 114-115).

Kalles is concerned about bad reviews of conference papers, describes two personal cases, and suggests an elaborate and unworkable scheme to fix the review process. I accept his view that the process is far from perfect, but I suggest some good although probably equally unacceptable solutions.

The practice of keeping the names of reviewers secret dates back to the start of the Royal Society, which was rent with controversy and hoped, by this measure, to avoid fights and duels among its members in London pubs. I suggest that this policy is now so antiquated as to be useless and propose that authors be told the names of the peers who have reviewed their papers. Let any resulting fighting go on and improve the entertainment value of conference papers, publishing who said what to whom. I invite attention to the 46 years of Computing Reviews, which deals only with signed reviews and offers a little-used method of venting an author's steam.

It's my understanding that Kalles's original paper named the guilty parties in his two cases, but the names were deleted because that's Computer's editorial policy. Let's dump that secrecy policy too and take advantage of the publicity that libel suits might provide for Computer.

My other proposal strikes at the very concept of having proposed conference papers reviewed by anyone other than the chair of the session. It has always seemed to me that it is a cop-out for a session chair to refer submissions to others. Anyone who is considered worthy of chairing a session should be well enough qualified in the subject to be able to judge whether a submission is good enough or not by reading it once. Someone who can't do this should decline to serve as chair. Limiting the reviewing to the chair eliminates the question of secrecy and avoids wasting the reviewers' time.

The problem with all these suggestions, Kalles's as well as mine, is that they seem to be unnecessary. The current curious ways of reviewing and accepting conference papers, although imperfect, must be working well enough since conferences are multiplying at a rate far in excess of the growth rate in the number of computing practitioners. If the reviewing system didn't work, the number of conferences would decline.

If it ain't broke, don't fix it.

Eric A. Weiss;

Dimitris Kalles responds:

Tracing the origins of anonymous peer reviewing back to the need to avoid duels and fights is both educating and amusing. But, today, publications, tenure, reputations, research funds, and the like are all linked to the thriving "publish or perish" paradigm. I don't think that things have changed too much.

Scrapping reviewer anonymity would increase transparency, but I am concerned about whether we can try something so fundamentally revolutionary at any significant scale, beyond any individual editor.

From the logistics point of view, preparing an author's statement is a matter of crafting two to three well-thought paragraphs. The review transcript can be semiautomatically prepared and easily stored. Both require the author to do substantially more work and the editor just a bit more.

The transcript could be edited according to a more relaxed timetable and appear with the metadata that now accompanies all papers in most publishing houses. Since most publishing houses now offer online access, that should not be a problem.

Perhaps by taking such small steps to decrease reputation-related pressure for journals, editors, and authors, we would also walk along the path to eventually removing reviewer anonymity.

The publishing process ain't broke from a business perspective. It's the scientific and professional perspective that I'm concerned about.

Wrong Bridge

I presume that Bob Colwell meant the Tacoma Narrows Bridge rather than the Verrazano Narrows Bridge (At Random, "Complexity in Design," Oct. 2005, pp. 10-12) when he referred to the structure that was demolished by high winds. I was a math major at Washington State University when the Tacoma Narrows Bridge self-destructed (with a little help).

Pete Anderson;

Bob Colwell responds:

I've been confusing those two bridges ever since I was a child, but this is the first time I've managed to expose my intellectual dyslexia to 100,000 readers. Thanks to Pete Anderson and the other readers who wrote to me to point out the correction.

64 ms
(Ver 3.x)