It is great to hear from you but…

It is great to hear from people out there interested in OR, but there are a couple types of email that I don’t like to get. I think lots of faculty have the same peeves. Jeffrey Ullman of Stanford (founder of much of database theory) has a nice page on this, which I will essentially copy here:

One class of email can be summarized as follows: “You don’t know me, but I have the following credentials and I would like you to arrange for me to be admitted to Stanford, in preference to someone with equally good or even better credentials.” I’m not sure why so many people think I, or any other faculty member — at Stanford or elsewhere — would decide to favor one unknown person over another. Sometimes their argument is that they want to work in my area. They seem to fantasize that there is some competition among areas at Stanford, and that I would push for the admission of someone just because they claimed to want to work in an area that I favored. However, it doesn’t work that way at Stanford, or any other American university that I know about.

Admittedly, there are some countries where PhD students are in effect hired by a faculty member and selected by them, but that’s not how it’s done here. At Stanford, a committee examines all applications and picks the ones with the most promise for study in Computer Science. The process is honest and objective. It is impossible to influence it through individual faculty, regardless of whether or not they are on the committee. In fact, I know of some faculty who will count it against you if you send this sort of email.

Whenever I get a “please treat me specially” letter, I respond with the following:

Thank you for your interest in Stanford. All admissions decisions are made by a committee of faculty and graduate students, and there is no point contacting individual faculty in the hope of bringing your case to their attention. Questions regarding admissions should be sent to admissions @ cs.stanford.edu You may also find out more about our department from URL http://www-cs.stanford.edu

Added Nov., 2003: A more recent variation is people asking for “summer internships.” Unfortunately, the same principle applies. You can’t get a student job at a university without being a student at that university. Faculty have a responsibility to serve the needs of the students at the school that pays their salary.

The second sort of email I get a lot of sounds like: “I really like your book on [fill in the blank], and I’m learning a lot. But I just have a doubt about the solution to Exercise 4.5.6. Could you please tell me the answer?” In a slight variant, it’s not a book exercise, but another problem whose solution they would like to see. I’m not that stupid. The chances are 98% that this is a homework assignment, and I’m not going to do your work for you. I therefore have developed a policy of responding to questions about material in my books only to bona-fide instructors. I rarely hear back when these email writers get the following stock response:

Thank you for your note. When I get these sorts of questions, I like to know first what school you are attending, what class you are taking, and who the instructor is (email if possible please). I suggest that first you discuss the problem with your local instructor. If they can't help you, then please ask them to get in touch with me.

Exactly. The doctoral web page for Carnegie Mellon’s doctoral program is http://www.tepper.cmu.edu/phd . All information on applying to the program is available there.

OR and Air Security

Operations research has been getting a lot of press recently due to a study about the effectiveness of “no-fly” lists in preventing terrorism. Long-time researcher of airline safety, Arnie Barnett along with Harvey Mudd professor Susan Martonosi found, using OR models of course, that it is best to screen all passengers, rather than try to pre-screen so-called safe passengers. Here are some excerpts from the LA Times:

Operations research is a little-known but valuable tool for such things as scheduling airline flight crews, planning National Football League seasons and even designing waiting lines at Walt Disney World. And in a report released on Monday, it was used to assess the effectiveness of the nation’s security screening of airline passengers.

Using a mathematical model, Susan E. Martonosi, an assistant professor of mathematics at Harvey Mudd College in Claremont, and Arnold Barnett, a professor of management science at MIT, sought to explore the effectiveness of the “no-fly” lists in preventing terrorism. The conclusions they reached were less remarkable perhaps than the way they evaluated the program.

They found that improving the screening required of all passengers at security checkpoints would do more to enhance security than further refinements to the pre-screening of passengers by no-fly lists.

Here is a presentation on the topic.  Check out your December issue of Interfaces for the full paper, along with a number of other papers on homeland security.

Some sightings from INFORMS Pittsburgh

OK, a “real” blogger would have put together a dozen posts from the Pittsburgh INFORMS Meeting. But I guess I am not, since I am just plain exhausted from the meeting and can’t put together any extended words. Suffice it to say, I thought it was terrific. Some glitches naturally (sorry to those who were in Greentree with insufficient transportation: wish I had a mulligan on that!), but overall I was thrilled about the things we had control over: the plenaries, the receptions, the convention center setup, the scheduling. We had 3780 participants, an all-time record for OR conferences. Thanks to all who made that happen!
A few sightings on the web: the “Practical Communist” has some postings and some absolutely amazing pictures.

Procure IQ has a number of posts on the conference, including a summary of President Mark Daskin’s plenary.

I can’t read David’s Space, but he clearly was at the conference (before heading to Chicago) and also has some great pictures.

OR in Numb3rs

Numb3rs tonight had a plotline involving causing blackouts by selectively destroying substations. At one point, the head nerd looks at a pad of paper and says “These are Dantzig-Wolfe Decompositions; network interdiction strategies. That’s pretty high-end stuff.” A closeup of the pad reveals notes that look a lot like the things lots of us in OR do (albeit I use Dantzig-Wolfe for things like sports schedules). The episode is entitled “Blackout” and the scene occurs about half-way through (1:25 into part 3 of the innertube version). Thanks Gary Lorden for the shout-out to OR! And thanks Brian Borchers for emailing me about this.

Navigating an INFORMS Meeting

Well, the INFORMS Pittsburgh Meeting is about to begin. The weather looks like it will be fine (no hurricanes like in Miami a few years back!). It is cool tonight (Saturday) but should get a bit warmer for most of the meeting.

At the Doctoral Colloquium tonight, INFORMS President Mark Daskin made some good points about navigating an INFORMS Meeting. One point that may not be obvious to first-time attendees is that you are very welcome to leave a session in between talks (it is a bit ruder to leave in the middle of a talk, but that is certainly not uncommon). So if you like presentation 1 of a session, and presentations 2 and 3 of a session three doors down, feel free to leave after the first presentation (typically as presenter 2 fiddles with the technology) and change rooms. It is something everyone knows after a few conferences, but even first-time attendees should do this. A second point is that many people attend tutorials of areas they specialize in. The best use of tutorials is to learn something of an area that is not known to you. I admit I check out tutorials in my area (to make sure they refer to me in appropriately reverential tones), but I am really wasting my time: I should either be attending something new or partaking in social capital activities (like having a drink at the bar with friends old and new). My own talk at the colloquium was on social capital and OR, similar to my EURO talk.
For those attending, enjoy the conference! For those not attending, shame on you, and plan for Seattle next year!

Update November 5

Mark has kindly provided his full

Mark’s Two-Page Guide to Navigating the

INFORMS Meeting

With a focus on first-time attendees

Continue reading “Navigating an INFORMS Meeting”