Monday, February 1, 2010

Some Observations on Research and the Academy

The above are the first batch of CDMA chips from Qualcomm that I used in 1991 to sell CDMA worldwide. I carried them from country to country to validate that my former Student Advisor at MIT from almost three decades earlier had really gotten CDMA operational. You see Irwinn was a one man leader, meaning he did the work and then gave direction to others who folllowed in his path because he was right. If you were in the communications space you knew Irwin. He was one of the many who were contributors, individual leaders, and change makers. This was the world of thee 1950s and 1960s. We seem to be in a different world of group contribution.

Along those lines, I was rereading a presentation Professor Robert Gallager gave on the celebration of the 70th Birthday of Professor Tom Kailath.

Gallager was on my Doctoral Committee and asked one of those life changing questions. You see I had just finished the answer book for Professor Harry VanTree Part 2, on Phase Locked Loops, which meant that I could calculate any Wiener-Hopf filter for any spectrum for any nonlinear phased locked loop, or whatever. Gallager asked me to explain the phase locked loop, with no equations! You see for an MIT student at the doctoral level that was impossible, that is, until I tried. It became a life changing epiphany. It changed the equation pusher into a Feynman, one who intuits the answer.

Now Gallager presents these following suggestions which I felt were worth commenting upon. To be fair, Gallager remarked that they were half serious but in humor there is often truth. So here goes:

1. Universities and research organizations should hire new faculty/staff on the basis of their best 1 or 2 papers.

Probably a great idea if anyone ever wrote a paper by themselves anymore. One of the problems is that authors go onto paper more than letters. I see in engineering almost a dozen authors per paper. In reality I know that at most two people can write a paper. At most! Usually even then one writes it and the second edits and contributes. Since we are trying to find the best thinker and innovator we cannot do that well when almost all papers have a group participation. Why has this happened, and more importantly why has this been allowed to happen? Faculty manage students, and the student is to write the paper.

What I notice today in may engineering professional journals such as the IEEE is the trick of having multiple authors along with some senior faculty, whose presence will insure publication. Thus there is some club of authors, and reviewers, many cases being one and the same, who attach themselves to papers to insure publication. Thus it not only adds the the students publications but also to the fame of the professor. This is the deadly embrace. From this we can never tell who did what and even more so whether this had any fair peer review. The other side of the coin is the new paper from an author not part of the club is assured rejection. Thus anyone who is aware of these tricks will discount the papers as being nothing more than mass self promotion.

I never wrote a paper with a senior faculty member. In fact I would have found that I had failed if I needed them, for I was testing my own intellectual powers. Now we see that a paper needs that senior author so that the "old boy network" and girl also works to get it published. Thus Gallager Rule 1 is gone and I think for a long while into the future.

2. The research component of tenure should also be the best 1 or 2 papers.

Again I agree and for the same reasons I gave above.

3. Since everyone can put their papers on the web (and reference other such papers), journals should publish only papers of real archival interest.

This is a real ground changer. It really begs the question of what function journals serve. They do peer review, but in my over 100 papers and dozen or so books that there was never a time when a reviewer added a scintilla of value. In fact there was a time an editor of a paper going into a book made changes he felt were critical and then wanted co-authorship. In fact he had no idea what he was talking about and then even threatened some actions.

So why not just use the web and its organic feedback. Get the comments from people who are willing to give their names and bona fides. It would be useful to have such a process so that anyone commenting is themselves subject to pari passu analysis and criticism. Such a process is iterative.

I can see on-line interactive "journals" where people can submit and then there is a trail of comments and changes as may be required. Such would be a Wiki like mechanism.

Frankly, I believe that there should be a new paradigm for "publishing" and the web has all the facilities for it. Papers of common interest can be placed on a "common interest" web site, and then they can be reviewed and edited real time. The focus should be on individual contributors, perhaps at most two, and then tracking of these can be done as a matter of course. Perhaps a limited accredited access may be required to keep out the Internet Trash, but the role of reviewers and editors can become a dynamic process, evolutionary in nature. The memorialization of a work can also readily be attained this way. I believe that Journals are often staid and many are barriers to getting results out. I almost never "read" a paper article, the journals are there for the ads, so I just read on line. I have been doing this for well over ten years.

Thus journals should and must evolve. It would be interesting to see who the "first mover" will be.

4. Conferences should try for more interaction rather than more parallel sessions.

Conferences have become means for people to get credit for more publications and to run around and chat. A hundred years or so ago when Einstein and his colleagues went to conferences it was something composed of series discussions of competent people and not a collection of disconnected collectors of conference credit.

Thus it is not the issue of interaction, it is the issue of anyone wanting to go, does, so all one has is a crowd of people adding to their CV. Pity

In reality the truth is that it is what "you" do not what your name is attached to. My concern is that the Academy has become an extension of those pervasive bumper stickers, "My Child is on the Honor Roll". You see anyone who shows up gets a bumper sticker! That is not the way excellence is developed.