Friday, July 29, 2011

Persistence vs. Publish Pressure (PPP)*

An anonymous commenter on a previous post asks, "How do you balance persistence and pressure to publish?".

The big question I have is - who is pressuring you to publish?

If it is yourself, then the way I view all publications/research is: 1) What is the big idea I want to do, and 2) How do I best tell the world about it?

Sometimes, if it's a big idea with lots of pieces, you publish as you go. This is another reason why it's good to diversify your publication venue. So, when you're just sketching ideas out, workshop. Maybe you have some preliminary results, low-tier conference. Maybe your research is rocking the house, top-tier conference / journal.

Some people say, "Bah, least publishable units, growl." But it's not about that necessarily. It's about telling a story, building on previous work, figuring out where you're going. Instead of waiting three years and squashing everything into one paper, you keep the ball rolling.

Sometimes you Don't Publish. And that's ok too. For example, in the middle of my PhD I spent about 6 months doing exactly nothing. Nada. I realized I was spinning to far out into the wrong direction. It was time to retool and rethink my plans.

Now, if someone else is pressuring you to publish, that's a whole other ball of wax. I think in that case it's a question of what relationship they have to you (dean, chair, advisor, colleague, student), and whether their request for you to publish makes sense. Are they pressuring you because they think your work is amazing and ready for the world to see? Are they pressuring you because they think you publishing now will help your career later? Is it to help their career? It helps to explore motives here.

If you need time to persist on a research thread, drop the self pressure to publish and find your center. If you have someone breathing down your neck to publish tell them to lay off for awhile while you get your groove back. The last thing you want to do is publish junk science just because you're being pressured to publish for the sake of publishing.

* For the geeks out there. (See also: this shirt.)

Tuesday, July 26, 2011

Reviewer armchair psychology

Did I mention July is the month for reviews this summer? I must have reviewed 25 this month (one for every hot, humid day!)

After I review papers, if I have time I enjoy doing armchair psychology on my fellow reviewers. Some conferences / journals let you see the reviews others have submitted, and some even allow you to change your score based on what you read. I'm not sure if this is a good thing or a bad thing, but it's interesting.

When there are 3-4 reviewers for a paper, the scores tend to regress to the mean. So on a 1-5 scale, the average score will be 3. There are also often repeats - so if I give it a '4', it's likely some one else will give it a 4 too. Really bad papers tend to have scores that cluster around 2, and really good papers cluster around 4.

So I'm always intrigued when I see the following:
Reviewer 1:  4
Reviewer 2:  5
Reviewer 3:  3
Reviewer 4:  1
As an nascent author, when you get a set of reviews back like the first one you tend to think, "Reviewer 4 is a jerk who Didn't Get It."

As a more seasoned author, you tend to think, "Oh no, what is my Fatal Flaw? (Reviewer 4 is a jerk who Didn't Get It.)"

And as a seasoned reviewer, you tend to think, "Who is Reviewer 4 and what is their beef?"

Occasionally Reviewer 4 has a valid point, and the other three reviewers really did miss something major. But more often than not Reviewer 4 is angry at the authors for taking too many liberties in their paper. Or for not citing Their Brilliant Work. Or it's the "Someone is WRONG on the internet" phenomenon.

In any case, when I'm an editor or paper chair I can ignore the outlier and life goes on. But when I'm a fellow reviewer I feel more vested in the outcome, particularly when I 'm Reviewer 2. I hate to see the possibility of good science getting squished because some reviewer was being thick, especially when it's someone else's science.

So sometimes, if a conference or journal offers a discussion period for reviewers, I occasionally have to confront Reviewer 4 head on, less they somehow manage to convince Reviewers 1 and 3 to change their scores.

Anyway, this is some of what goes on behind the scenes behind your favorite publication venue. As an author, try not to let the outliers get under your skin. If your other reviews are good, be persistent and try again somewhere else. There's an awful lot of randomness in this process.

Friday, July 15, 2011

How to get your paper accepted: Orshee

In today's installment of how to get your paper accepted, we shall discuss gender inclusive language.

Back in my days of blissful ignorance, I didn't notice gender use in language very much. "John Doe" and "He" were pretty much par for the course.

At some point, I was reading an article and it was positively littered with "him or her" "he or she" "his or hers", and I wanted to pull my hair (short or long) out. While I appreciated the sentiment it was completely distracting from the prose.

I once was given a Parenting 101 book, and it alternated between male and female examples per section (i.e., every few pages). I liked this approach a lot better, because it made for much easier reading while still being gender inclusive.

Gender exclusive language has no place in scientific writing, unless the author is describing a single case study (i.e., "When Patient M. first came to the hospital, he..."), a gendered-exclusive event (i.e., The Society for Women Engineers summer camp for fourth grade girls), or is somehow written in the third person from the perspective of one of the authors.

It's very easy to use anonymous, gender-neutral subjects in sentences to give examples of people. For example, "the student", "the user", "the agent", "the engineer", "the scientist", etc.

It takes practice to write in active voice while remaining gender neutral; sometimes the writing can get a bit bogged down when you start. Sometimes writing they or them can feel awkward. But like any sort of writing, practice makes perfect. After awhile it becomes second nature.

Unlike those days of blissful ignorance, as a reviewer I am now very distracted and occasionally annoyed by both gender exclusive language (of either gender), as well as by too many Orshees. In some particularly egregious cases of the former I have politely reminded the authors to be more sensitive to their use of language. I know it is often a result of English being a second language.

Google, however, really should know better. Check out this error message I just got in Chrome (emphasis mine):
In this case, the certificate has not been verified by a third party that your computer trusts. Anyone can create a certificate claiming to be whatever website they choose, which is why it must be verified by a trusted third party. Without that verification, the identity information in the certificate is meaningless. It is therefore not possible to verify that you are communicating with  XXX.YYY.ZZZ, instead of an attacker who generated his own certificate claiming to be XXX.YYY.ZZZ. You should not proceed past this point.
If I was a man I might be offended. I'm sure there are plenty of female hackers out there. (Heck, even that attack is poorly named - "man in the middle". I guess it's catchier than "person in the middle", but still).

Wednesday, July 13, 2011

How to get your paper accepted: Our results are very awesome!

In today's installment of how to get your paper accepted, I'd very, very much like to discuss intensifiers. And exclamations! So I will.

Scientific writing is first and foremost about clear, careful communication. You can have the most amazing results in the world, but if you can't clearly walk your reader through your science, you're going to run into problems. Furthermore, I said "careful" because in scientific writing it is also important to be humble, and not take your conclusions too far.*

By using intensifiers, which are adverbs that elevate the word following it, you not only run the risk of over-generalizing, but you also risk angering your reviewers/readers. It is highly unlikely most authors can make claims like: "Our work makes a very important contribution", "We present really groundbreaking work on embedded rubber ducks", or "This work is extremely revolutionary".

Exclamations, too, rarely have place in scientific prose. Sometimes if you are trying to write something that catches the reader's attention, exclamations may be appropriate. For example, if you were writing a technical article on cellular phone use in rural India and wished to point out some fact about how people are more likely to have phones than shoes, say, that might make sense.

An editorial or book review is a fine place to use qualifiers, and possibly also exclamations. These are publication venues that expect authors to state opinions and generalizations, as well as to catch a reader's attention.

But for your standard journal or conference article, keep the intensifiers (and exclamations!) at home. 

* This is important for many reasons, not least of which is making a generalization or prediction about the future that is entirely wrong and/or taken out of context, and having to relive it for decades. See also, "There is no reason for any individual to have a computer in their home."

Tuesday, July 12, 2011

How to get your paper accepted: Short paragraphs

July seems to be the month for reviews, so I thought I'd organize some of my observations on scientific writing into bite-sized advice posts.

1) If you want to get your paper accepted, please, for the love of all things, use short paragraphs.

I was reviewing a two-column ACM format paper recently, and a few paragraphs took up the entire left-side column and half of the right-side column. My eyes went blurry by the end, and frankly it negatively biased me against the authors.

If authors are concerned about space, they should either use less words or make their diagrams smaller. I'd much rather see smaller diagrams and more readable text than huge diagrams and squished prose.

Also - putting hundreds of lines of code into a paper is rarely necessary. (And XML is never necessary*). Use small chunks. Just the important idea behind the awesome algorithm. If the code paragraphs are taking up more than half a page, please consider an alternate presentation style. (See Justin Zobel for nice presentation ideas).

(*) <meta>I'm sure there's a good xkcd comic out there for this sentiment, though my Google fu is weak today.</meta>

Thursday, July 7, 2011

Having what it takes

There is only one trait you need in order to be a computer scientist: persistence. And I mean dogged persistence. Like you spend 10 hours on a problem persistent.

This is actually a learning exercise. After some time, you start to discern patterns in how things tend to break, and you see them in multiple places. Even as operating systems, programming languages, and applications change, you see these patterns of how to fix broken things, because you have experience under your belt.

The truth is, if you have no patience for such things, or if you want to quit after an hour of working on something, chances are you won't last long in this field.

I can tell you, though, that persistence pays off. After awhile, the mystery of machines begins to go away, you begin to see patterns, and your frustration dissipates.  And best of all, when you do have that breakthrough and fix the damned thing, you feel an overwhelming sense of accomplishment.

It feels like this is difficult to communicate to "kids today". I think it's hard for them to see the point of the struggle, when there are so many other fields that don't require nearly so much frustration. (A friend once described it as constantly banging your head against the wall and then feeling really good when you stop doing it, which perhaps is apt.)

Often people ask me why I became a computer scientist, and I always reply honestly - it never occurred to me to do anything else. (And I'm as stubborn as a mule, which helps. :-))

Wednesday, July 6, 2011

Burn after reading

I received an unencrypted email yesterday which had in the title "CONFIDENTIAL AND PRIVATE". It also had instructions that if I were to print out its PDF attachment, I must shred it immediately after reading it.

Clearly we Computer Scientists are doing a bad job with public outreach here. So, hey, chance to educate.

Email is hardly ever secure. I say 'hardly ever' because it is possible to encrypt email, and it is also possible to send email on secure, closed networks, free from the pull of the internet sea.

But most of the average email your average person is sending is being sent in the clear, unencrypted. This is a lot like walking down the street holding a big sign with the contents of your email. Which is recorded by a camera. And a lot of people can watch the video at any point in the future. Also, the video is archived in a library 4ever*.

The metaphor of a paper postal letter may have made sense about 15-20 years ago, but it's no longer valid. A letter sent by physical mail is much harder for lots and lots of people to read, unless someone tampers with the mail, makes a photocopy, etc. It also had ephemerality - you really could burn it after reading.

I pretty much operate under the assumption that any determined person can read my email and all unencrypted files on my computer. I also assume any emails I send could end up being forwarded to others, printed out, or posted on some blog somewhere.  Err on the side of caution, and all.

(*) Ok, except a library is a bad metaphor because it's not necessarily easy for people to find this video. (aka. "security through obscurity"). Nor is it necessarily around for ever, but it could be.