Employee furloughs at Arizona State University

I've been crazy busy with a new position here at work, so I haven't been able to indulge myself in the luxury of ranting and raving on the internet (except for 140 characters at a time, on twitter). But the article linked from the title above caught my eye, and then the following quote from that article made me laugh out loud:

Faculty members will take furloughs on days they don’t teach class...

I wonder if this means that faculty can refrain from grading on those days and hand things back late. Seriously, about the only thing this frees faculty from, I suppose, is going to meetings. And, considering that meetings are scheduled in the cracks between the various members' classes, that's probably not even true.

Coming back up for air

I'm always impressed by people who can go to conferences and blog about their experiences during the conference. I typically find myself struggling to fit all of the conference-related activities into each day. And then there are the conferences with long travel times (20+ hours to Montevideo, Uruguay, for example) and jet lag to contend with, during which I consume mass quantities of coffee in a sometimes vain attempt to stave off narcolepsy (despite the fact that I'm not much of a coffee drinker). Maybe next time I'll take a bunch of decongestant, instead.

Presentations, be they oral or poster, are always curious things. Oral presentations are often a complete waste of time. My understanding is that, once upon a time, oral presentations at scientific meetings involved people reading their papers to the audience. While, on the surface, this might sound dreadful, upon further reflection I think this is a better approach. For one thing, this would greatly reduce the number of oral presentations. For another, it could (note the conditional) engender more discussion among the participants. Right now, presenters try to make their points in a few slides and questions are basically at the end. There's no time for serious questions, clarifications, or the like, so questions tend to be of the "got ya!" or the "let me tell you about what I've done that may relate to this" varieties. And those are the good talks -- the bad ones are those that are incomprehensible or those in which the presenter hasn't confined him or herself to a few slides and everyone is at the mercy of the session Chair's ability to cut the speaker off so that they can make the next coffee break in time.

Posters are sometimes better, but they unfortunately try to serve two purposes. One purpose is to provide a forum for more embryonic work, often that of students. The other is as a "booby prize" for submissions for which oral presentation wasn't justifiable but which were still above the "ramblings of a crackpot" cutoff used for rejections. So, one's first task is to find those posters that one is interested in and then locate the author (who may or may not be there at the moment). If you're lucky and do both of those, odds are the author is in the middle of talking to someone else and then you have to decide if you want to listen to the last five minutes of that conversation, trying to understand what they're talking about, before you finally ask the author to essentially repeat the whole conversation with you.

So, what's good about conferences? They're basically the only place I can go where I can meet and talk to people who are interested in the same things I'm interested in, research-wise. I can have research for breakfast, lunch, dinner, and late-night drinks. I can sometimes make contact with future collaborators. I can get ideas for the next thing I'd like to do. I can be re-energized to continue to find ways to carve out space for research in a life with plenty of family, administrative, teaching, and service demands.

I'd like to run a conference in which oral presentations are done in a more old-fashioned way, where papers are read and discussed in a seminar format. That way, we could all go through each other's research with a fine-tooth comb. What a frightening prospect! Anyway, back to seeing if I can bring the number of unread emails in my inbox below 50.

Measuring science

This post was inspired by an excellent one by GrrlScientist, linked from the title above. She starts off discussing journal impact factors, which are a measure of the average number of times a paper in a journal is cited by others. Then there's what is essentially a personal impact factor, which is the number of times a particular researcher's papers are cited. These have problems, which the H-index is meant to address. Briefly, a person has an H-index of h if he or she has at least h papers cited at least h times. So, if I have 100 papers, each cited once, then I have an H-index of 1. If 99 are cited once and one is cited 43,000 times, my H-index is still 1. If 95 are cited once and the remaining 5 are each cited at least 5 times, then I have an H-index of 5. And so on.

So, first of all, there is the question of gaming the system. It's unlikely that I can convince 43,000 of my colleagues to cite one of my papers (but, if you'd like, pick one from my CV on my UW home page and cite away). But if I'm only shooting for, say, an H-factor of 20 or so, then that might be doable. Supposedly, people do try to game the system by doing things like citation swapping, though this seems to me to represent time better spend being a more productive researcher (rather than just trying to look more productive or impactful).

Though I may be unconvinced about the effects of such gaming, I see this as a fatal flaw of any attempt to extract a simple metric from the interrelationships among such publications. Just look at how much effort Google has expended on providing good search results. Since these results are presented in a sequence, presumably from most relevant (or "best") to least, they have been implicitly assigned a single measure. And there's a cottage industry surrounding pushing sites' ratings up that has nothing to do with their content. I'll come back to this idea of creating a one-dimensional ordering later.

To me, there's another problem with metrics such as this. Let's say that my H-index is 11, as computed using Google Scholar. Furthermore, let's assume that issues such as self-cites (citing one's own work) and co-cites (citing of one's work by collaborators; I'll revisit this topic, too) don't effect rankings (these may be invalid assumptions). There's still one problem: is an H-index of 11 good? Bad? Middling? If we read Wikipedia, we learn, "In physics, a moderately productive scientist should have an h equal to the number of years of service while biomedical scientists tend to have higher values."

But what about computer scientists? We could consult a listing like the CS Meta H index. We would then have to compare my H-index with other faculty at similar stages in their careers who are working at similar institutions and who have had roughly similar career paths. Unfortunately, that information isn't in the index. We need to know a lot about different universities, different CS departments, and individual faculty. Maybe it would just be easier to read one or two of my papers and judge for yourself.

Coming back to the subject of co-cites, this could be considered a sign of an attempt to game the system. On the other hand, it would make more sense for me to make gaming arrangements with colleagues with whom I have no direct professional connection. (Hmm. Three more strategically placed citations will get me to an H-index of 12; five more in just the right spots will get me to 13.) But what about people who collaborate widely? Their papers will have lots of co-cites, but their work will also be more broadly influential because of all that collaboration. So, when I've prepared materials for external review, I always separate out the co-cites. Make of them what you will.

The desire to create this scalar (one-dimensional) metric of scholarly is a natural one. When I look at the complex dynamical behavior of a neural network, one of the first things I want to do is extract a single measure to characterize that behavior, so that I can then more easily examine how behavior depends on various parameters. But I have a very carefully defined question in mind when I do this. When we measure science, what is our question? Are we asking if a particular scientist is "good"? What is good? Does it mean that the scientist's work has impact in the field? How can we really ascertain this without understanding the field and the scientist's contributions in that context?

Einstein had four papers that changed the field of physics forever. But that's just an H-index of 4. I was discussing this with one of my colleagues, however, and his opinion was that 4 was a reasonable assessment of Einstein, and that we should want to hire and promote scientists who are consistently productive, not ones who have one brilliant flash of insight and then nothing approaching that for the rest of their lives. But how can we tell the difference between consistent, quality productivity and a laser-like focus on getting out each least publishable unit? To me, the only solution is knowing the person; we can't reduce the behavior of that large a neural network to a single useful measure.

My Mathematics Genealogy

Follow the link above; unfortunately, my advisors and myself are not close enough to being mathematicians to produce a very full genealogy.

The last syllabus you'll ever need

Over at Halfway There, Zeno has a post about a one-size-fits-all syllabus. It certainly is the case that course syllabi have evolved from a means of clear communication with students to written contracts on the nature of courses. Perhaps we should take a page from software End User License writers, who typically tell us we have no rights and that they can change the contract anytime they want and that a post on some web page we'll never check shall be considered sufficient notice of such change.

Oh, well, I guess I'm not curmudgeonly enough yet to do this. But, if I'm told one more time that my use of the word "should" in a homework assignment can be reasonably interpreted as advice, rather than a requirement...

P.S. If you're a student reading this, I hope you realize it is in jest. I am well-known for my voluminous syllabi. The next best thing to a captive listening audience is a captive group of readers.

My alma mater

Yes, the UCLA police repeatedly tasered a UCLA student who forgot his student ID and was a little too slow leaving Powell Library. I used to think that cameras and video capability were superfluous on a cell phone, but this YouTube video has convinced me otherwise. The statement from Acting Chancellor Norman Abrams is noncomittal, as might be expected at this stage. But there should be a policy preventing the use of force, not to mention weapons, when police are not themselves threatened.

Watch the video. A nice touch is when an officer threatens to tase a bystander asking for his badge number.

I'll be sending a letter to Chancellor Abrams today. Other UCLA alumni might consider doing the same, or emailing him at chancellor@conet.ucla.edu.

Update: Here's a May 2005 GAO report on the use of tasers by law enforcement, including policies in selected departments. Referring to Figure 3 on page 8, I would classify the student's actions as "Resistant (passive)," which at almost all of the cited departments would not justify the use of a taser.

Update 2: Apparently, the UCPD's meritorious service award is called the Taser Award.

Sick day blogging

Upper respiratory infection #2 (since this is the Pacific Northwest, I suppose these should be numbered, rather than named) is blowing through the community and our household, just when I thought I was getting over the first one (I was at the "coughing so much I think I jarred my brain loose" stage). A perfect accompaniment to a dark, overcast, rainy day among a long sequence of such days. So, lacking the energy or mental firepower for complex activities, let's see what's happening in Academia, Blogland:

• In Knowing and Doing: When Opportunity Knocks, Eugene Wallingford writes about making CS assignments more interesting and relevant to the real world by being more open to opportunities to do cool, fun stuff in class.
• Michael Bérubé says "I got plenty of nothin'", but I respectfully disagree. What he has is a great post on the value of academic blogs for showing what academics do.
• Love blooms for FemaleCSGradStudent in a seemingly unlikely place.

My Erdös number is at most 5

Daniel Lemire sparked my interest in finding this out. The link above is to a page under the Erdös Number Project. Paul Erdös was a prolific mathematician, and an Erdös number is the number of "degrees of separation" one is away from him. As that site says, the vast majority of mathematicians have an Erdös number less than eight. So, one of his co-authors has an Erdös number of 1, a co-author of a co-author has an Erdös number of 2, etc. See this Wikipedia entry for more. You can use MathSciNet to help you computer your Erdös number, but as it focuses on mathematics publications, it only gives an upper bound (for instance, it gives me an Erdös number of 6, but one of the folks along the pathway, listed as a co-author of my PhD advisor, is also one of my co-authors). I can actually trace multiple paths of length 5 to Paul Erdös. And I don't consider myself a mathematician; I just hang out with the wrong people.

Vanity press

Those WMSCI ("World Multi-Conference on Systemics, Cybernetics and Informatics") people seem to be at it early this year, with their solicitations for submissions to their curious conference. Click the title above to learn more about how to save money, your reputation, and put your work in a more useful venue. If you want to feed your vanity, start a blog like everyone else!

Studicious

No, it has nothing to do with Congressional Pages. Via Brad DeLong's Semi-Daily Journal comes a early version of a web site (linked from the title above) that allows user to enter and share class notes. It's an interesting idea, and it makes me wonder about the possible uses of collective note-taking. A substitute for taking -- and rewriting, with refinements, while studying -- notes it is not.

Gaming impact factors

Via Stephen's Web and ACRLog comes a pointer to a Wall Street Journal article from the title above. The article reports that some journals are trying to boost their impact factors by asking (telling?) authors of papers to cite more articles from that journal. When a scientific journal editor says something like:

...when we edit a paper...we sometimes ask authors to ensure that the relevant literature is cited... I can state unequivocally that we do not attempt to manipulate the JTT's impact factor. For a start, I wouldn't know how to.

You know that something's up. If a journal editor doesn't understand impact factors well enough to manipulate them, then who does?

I've blogged on journal rankings before. Taken out of context, impact factors don't tell much. For example, the Journal article says that an impact factor below 2 is "low". But this depends on the size of the field -- fewer researchers mean fewer citations -- not quality.

Topics: academia, research.

Spamference, spamjournal, spamreviews

Everyone's favorite spamference, the World Multi-Conference on Systemics, Cybernetics and Informatics (WMSCI; I'll refrain from linking to them in the name of not promoting that which doesn't deserve promotion) has now posted their conference submissions on the web. So, for your amusement and edification, head on over to My Pet Goat, where Mrs. Cake has baked up some pointers to them.

Of course, any profitable economic niche will attract new business ventures. The IPSI BgD folks do the WMSCI'ers one better, with two generic domains (internetjournals.net and internetconferences.net), generic conferences (actually, VIP conferences, which are also "M+I+T++" conferences: "Multi-, Inter-, and Trans- disciplinary") held all around the world, and no less than two journals! Here, I'll concentrate on the journals, since I recently got some email about them. There are two: IPSI Transactions on Internet Research and IPSI Transactions on Advanced Research.

Among their submission rules are, "Each paper must include at least 10 references. Papers without at least 10 references will be returned to the sender." Why? I assume that's because this is how they get names of reviewers. They charge €500 to publish a paper, which as they say is fortunate for you, because, "...the publication fee is to be paid only if your paper is accepted - we do not charge for the reviewing process, which is a very costly process." How costly? Well, they need to send out 12 emails to solicit reviews. Here's what I assume a typical email looks like:

Dear Dr. Michael Stiber

We have found on the Internet that your area of research is similar
to a topic submitted in the paper for our Journal
Transactions on Internet Research.

We know that you are extremely busy, but please take time to do the
quick review of a attached paper and send us your comments on:

- What did you like in the paper
- What you did not like in the paper
- Scientific approach level in the paper
- Updated info in the paper, etc.

Any comments of yours will help the authors of the paper.
Please reply as soon as possible.

Our journal is aimed at supporting a truly multidisciplinary,
interdisciplinary and transdisciplinary research.
For details, please see the web (www.internetjournals.net).

We expect that your review has two parts.
First, improvement requests to the authors.

Second, your recommendation:
A. Unconditional accept
B. Conditional accept, with minor changes
C. Conditional accept with major changes
D. Reject

If you have any more questions, please contact me!
Please reply to this e-mail.

Best Regards,
Prof. Dr. Veljko Milutinovic, Fellow of IEEE

Then they attach a PDF of the paper. Unfortunately, they don't offer to pay me for my review, so I'm not sure what makes this review process so expensive. Could it be a lot of time spent finding reviewers? Well, no, as they themselves say, they grab previous journal and conference participants, folks cited in references, and Google the rest. Despite what they say, my research interests aren't similar to the attached paper, so they don't seem to do much more than type in a keyword or two, grab an email address, and shoot the request off.

Oh, all of their issues seem to be online in PDF form, but you can subscribe for €500 if you're an individual (€1000 for institutions; €500 for top 500 institutions).

I'm sure that there are people submitting to these things and reviewing their papers in good faith. After all, people get taken in by spam all the time. On the other hand, they apparently get 27 hits/day on their journal web site, so I guess that shuts me up.

Topics: academia, conferences, journals.

[+/-] show/hide the rest of this post

Ward Churchill report released

Click the link from the title above, if you are the kind of person who likes to find out what happens after all of the sensational headlines and hoopla have faded away.

Automated bogus research paper detection

By now, is there a person in the world who isn't familiar with the annual WMCSCI spamference and the bogus paper submitted and accepted by them in 2005? Well, technology has caught up, and now Nagib Callaos and the other folks who run this and similar conferences can click on the link from the title above to get information on automatic detection of bogus papers without all that messy and bothersome reviewing.

The best job in America

You may argue about the exact ranking, but not its presence in the top 10. It also ranks second in job growth rate. Of additional interest is the #7 ranking of computer IT analyst (ranked 7th also in job growth rate). Oh, and college professor is #2.

Drinking my own lemonade

In a previous post on 12/5/2005, I commented on one impact of lower CS enrollment: smaller classes. I was wondering what I would do differently in my Winter quarter class, given that there would only be a few students in it. Now that the quarter is over, I thought I'd comment on what actually happened.

The class in question was a combination of algorithms, discrete math, and object-oriented design and programming. I ended up with 8 students in the class. Instead of meeting in the regularly scheduled classroom, we met in my lab, which gave a cozier, more group-oriented atmosphere. I organized the class around a single project. I covered all of the material that I would normally have, but started each topic with a roughly 30 minute presentation given by each student.

Project

I initially left the project wide open to student input; I would have been perfectly happy with any project they proposed, as long as it exercised the various topics we needed to cover. Failing that, I proposed what eventually turned out to be a simplified version of del.icio.us, except using hash tables, AVL trees, red/black trees, and graphs -- instead of databases -- for the back end. The project eventually developed an architecture with a back-end server (the bulk of the project) responding to CGI application requests. It actually is quite a nice job; I'm just waiting for a couple students to finish a "story of the project" web page (which I indicated was beyond the scope of the course itself) before making the URL public (just for looking; I won't leave the server running).

Project management

I am by nature a pretty hands-off person. It was not my goal to micro-manage the project, and in fact I would have been happy to have the students do everything. However, I couldn't make that a requirement of the course, as it wasn't a project management course. Instead, I gave them a lot of leeway, and if they didn't take up the slack, I gradually tightened things up and prescribed what they would do. So, at first things were wide open, even including the project itself. Then, I specified the overall project goal. I later led an in-class design exercise, and broke the project down into subsystems and tasks and farmed them out to teams. So, step by step, I produced more specific directions on what would be done, by whom, and when. Eventually, we had all of the elements you'd expect of a project: milestones, interface documents, version control, team meetings, group meetings, weekly reports (with required reflection on successes, failures, what they were doing right and wrong, and what they'd do differently the following week), code reviews, etc.

Certainly, I could have done all of this from the start -- it's how programming assignments would normally be done in CS classes -- and more could have been accomplished. but I think there's value in students being given a chance to do something, see that one approach doesn't make much headway, and then see that an alternative given by me works. I think, as long as you don't allow things to get too chaotic or students to get discouraged, this is better than laying everything out at the start. We were still able to accomplish all of my essential goals, and that's all one can really hope for in one quarter (did I mention yet that the quarter system sucks?).

Presentations

I developed a rubric for grading student presentations and summarized it in my syllabus. These are undergraduate presentations, and so they of course can't be expected to be up to the level of presentations in a graduate seminar. Beyond the experience itself, it helped to give me clues regarding what topics to cover in more detail.

All in all, I was pretty happy with the class' results. I'm not sure I would recommend this approach in a class with more than eight students. I suppose there could be multiple, smaller groups, but how crazed would it be to have multiple, large (all quarter or semester long) projects ongoing? Generally, I don't like group projects (I do one in a class every so often to remind myself why I don't like them), but the length of this one and the number of students helped mitigate many of the negative issues.

Take home message: size does matter. Smaller classes are better.

Topics: academia, computer science.

[+/-] show/hide the rest of this post

CS enrollment continues decline

The title above links to data from the Computing Research Association on undergraduate computer science enrollment. The number of new students entering the CS pipeline continued its decline in Fall 2005 to about 8,000, from a high of around 16,000 in Fall 2000. Note that these figures are for US PhD granting departments only. This decline has had an impact on total enrollment since the 2001-02 academic year, and for the first time (during 2004-05) impacted the number of Bachelor's degrees granted (down 17% from its peak in 2003-04).

I take this as good news, not bad. Since the dot-com bust around 2000, the number of CS degrees granted rose from around 10,000 or so to 14,000, and is now only down to a bit less than 12,000. Continued decreases over the next couple years will start the pendulum swinging the other way, with companies increasing pay and benefits, and maybe even improving the work climate. The message will filter through to prospective students, and new enrollments will increase.

A problem of perception?

The title above links to a New York Times editorial about outsourcing, the computing profession, and employment. As anyone in education knows, the number of students interested in computing careers has dropped significantly, in reaction to the dot-com bust and fears of outsourcing. However, as in many previous cycles of engineering employment, things have now entered the stage of overreaction. The article rightly points out that the employment situation has improved to the extent that there are now more computing jobs than at the height of the bubble -- I've heard anecdotal stories from industry that are consistent with this. And fears of outsourcing have been overplayed.

Fundamentally, it doesn't make any sense to make a multi-decade career decision based on employment statistics from last year (not to mention as long ago as the dot-com bubble burst is now) -- unless you think that computing careers are going the way of blacksmithing. The real crisis is that the field is losing the battle to attract top students. Right off the bat, we lose half of the best students: women. This is not just a US phenomenon. We need to do a much better job of providing accurate information about the profession to K-12 students and teachers. Yes, as companies want more employees, salaries will increase. But, as I think we've learned, to have a profession that is perceived to be attractive only because of high wages is to have one with a very shaky interest base. We need to emphasize the non-monetary, psychological rewards of our profession. We need to work to nurture that kind of reward structure, both in higher education and in careers.

Topics: careers, computing.

What would the ideal computer science department look like?

Starting this fall, my campus will be admitting freshmen for the first time (up till now, we've been only upper division and graduate; don't ask why). As is the custom in Washington State, planning for this commenced approximately a year before the start date. I've been intimately involved in our CS lower division curriculum design, and am now turning my attention to what I will refer to as "educational environment". By this, I mean everything other than classes that contributes to learning and, just as importantly, to making the CS program an inviting, exciting, and comfortable place in which to learn -- that make a department a community.

As I've written before, our profession is in desperate need of new approaches. While part of that involves rethinking our curriculum and how we teach, I think a big part is also non-curricular: activities that students and faculty engage in outside of class, design of laboratory spaces, etc. For example, a typical computer lab is a room filled with desktop computers, all facing the same direction. In the past, this invited students to concentrate on working on the computer, I suppose. Also, it didn't require much design thought on the part of faculty. Nowadays, students have their own computers, so why spend time in a computer lab? The only reason would be if they're forced to, for instance because they don't have some required software. This is pretty much the polar opposite of an inviting workspace, and certainly not the way to encourage participation in the profession by students who are uncertain about their major. Ideally, a computer lab should be a place students want to spend time in, to think about their studies, work with other students, even relax and hang out.

The lab's just one part of the puzzle. There's also the overall environment, including organizations, events, traditions, etc. So, what I'm doing here is soliciting ideas: what do you think would make an ideal CS department? You don't need to assume anything, and you can address any aspect of the department (of course, I reserve the right to pick and choose from your advice). Please contact anyone you know who might have some advice to give and ask them to either comment to this post or email me. I'll even put my money where my mouth is, and give a gift certificate to Amazon.com or a membership to LibraryThing to anyone who submits a "surprisingly great" idea (either $10/1 year membership or $25/lifetime membership, depending on how great the idea is).

I'll summarize the suggestions in a later post, and also let you know what eventually happens (and, in future years, what the impact is, at least anecdotally). Besides all of the non-lab-related matters, I'm also looking for lab design ideas. Currently, we have one lab with around 30 Windows machines and one lab with around 15 desktop Linux machines and a 15-machine Linux cluster (so, basically a room that can hold 30 machines on tables set up in rows, but that only has 15 machines on tables).

Topics: academia, computer science.

[+/-] show/hide the rest of this post

A spam book?

I recently received an email from those fine folks who bring the "World Multi-Conference on Systemics, Cybernetics and Informatics" to your email inbox each year. They've been criticized for running their spamference without recourse to peer review of even the most rudimentary kind, at least in some cases. Now, it appears that they may be taking a page out of the Republicans' play book: soliciting "science" that supports their approach to conferences. Here's an excerpt from the email I received:

Based on your participation in conferences, we would like to consult your opinion and your possible contribution regarding the idea of collecting, in a multiple-author book or symposium proceedings, reflections and knowledge regarding conferences organization and quality standards/means. It will only take you about 30 seconds to give us your opinion and your potential support as a reviewer and/or paper contributor...

Apparently, the only qualifying factor needed to contribute to this work is participation in conferences. Science was never easier!

Topics: academia,conferences.