Research minute: The practical aspects of evaluating video quality in real time

Whenever we present our work on evaluating users’ quality of experience (QoE) with online streamed video (like watching YouTube), people ask us, “So, how would this work in a real-world system?”

We’ve come up with some neat algorithms for determining how a user would evaluate video quality, based solely on measurements we can easily obtain like bandwidth and received packets, but our work has largely remained in the proof-of-concept, measurement and analysis tool realm up to this point. As a former engineer, though, I’m interested in building things. My vision for this project has always been to build a functioning system, one that can predict and evaluate video QoE in real time. So a couple of years ago, we set out to answer some of those questions.

This week, I’ll be presenting the first results of that study, “Systems Considerations in Real Time Video QoE Assessment”, at Globecom—specifically, at the Workshop on Quality of Experience for Multimedia Communications. In this study, we attempted to answer the following questions: How frequently can we generate video QoE ratings with some degree of accuracy? How often should we sample the measurement data? How do we weigh the need to consolidate data collection (arguing for fewer, less frequent data points) with the need to monitor video quality in real time (arguing for more frequent data points)? What are the timing requirements for such a system, both in training the system and in assigning ratings to videos?

To answer these questions, we used the data we collected in the summer of 2010 for this paper, developed a mechanism to play back the data in pseudo-real time, and then sliced and diced the data in various ways. We played with the sampling rate: how many seconds should pass between measurements? We played with the amount of data to process at once: would ten seconds be enough to give us an accurate video QoE rating? would a minute be too long? Which video measurements should we use: all of them, some of them, one of them at a time? While trying all of these combinations, we kept our eye on the clock, literally: if this system is going to deliver results in real time, then we need to make sure that “teaching” the system how to evaluate videos does not take too long—otherwise, our system is not very adaptable, and thus not very useful.

In the middle of the study, we realized that our assumptions about the video delivery system itself might also impact how the system is designed. For instance, our system could look like Netflix: Netflix controls which videos are available to viewers, and the content is fairly stable (new videos are added on predictable schedules). This looks very different from something like YouTube, where the available content changes rapidly because people are constantly uploading videos.  In the former case, we can “teach” our system using videos that people will be watching. In the latter case, there are no guarantees that we have videos to teach our system that look anything like what people will be watching. So we considered both of these scenarios as well.

So what did we learn from this study?

• Taking data samples frequently and processing smaller amounts yields the best results, most of the time. Except for the shortest video in our study (a 2 minute clip of dialog from a movie), we were best able to predict video quality by sampling data every second and evaluating the data in 20-second chunks. (For the shortest video, going 50-60 seconds between evaluations worked better. We think this is because the scene changes in this clip happened about every 40-50 seconds.)
• More is not better, when it comes to what type of data to use. While we had 4 different categories of data available from the videos—bandwidth, frame rate, received packets, and the number of times the clip buffered—we found that concentrating on just bandwidth and received packets gave us the most accurate picture (no pun intended) of video quality, in general. Again, our shortest video was an outlier: here, frame rate did a better job of judging video quality.
• The system can operate in real time, because video ratings assessment happens in less than a second, and at most it takes 10 minutes to train the system, which is done off-line anyway. The “less than 10 minutes to train” is key, because this means we can continually re-train our system as new videos come online, if we choose to do so.

Clearly this paper doesn’t definitively answer the question of how such a system would work, but it’s a step on the right path. We are actively considering some related questions, specifically what other infrastructure pieces would be required to support collecting, analyzing, processing, and feeding back such measurements into the system, so that the system could fix itself when video quality goes south. There’s also the question of how reliant our data is on the particular videos selected. We actually found support for “one configuration to rule them all”: a sampling rate, evaluation time, and set of measurements that worked universally well for all videos and for each of the video system scenarios, which is promising in terms of the generalizability of our solution, but further study with additional videos would definitely help.

Acknowledgements: Two of my research students, Tung Phan ’13 and Robert Guo ’13, did the initial studies and analysis of the data, in 2011. At that point, we actually got stuck and put the project aside for a bit. The insights we gained from what didn’t work informed the approach in this paper, and definitely made this paper possible! The infrastructure for collecting and analyzing the data, and the data used in the paper, came out of work in the summer of 2010 by Guo, Anya Johnson ’12, Andy Bouchard ’12, and Sara Cantor ’11.

#AcWriMo wrap-up

#AcWriMo comes to a close today, so it seems fitting to wrap up the month of crazy writing goals with an accounting of how well I did (or didn’t) meet those goals.

As a reminder, here are the three goals I set for myself for the month:

1. Finish the grant narrative, and send it out to colleagues for comments, by 16 November. (The grant narrative is the “meat” of the application, and is the technical explanation of the work and deliverables.)
2. Finish the supplementary documents (RUI statement, budget outline, budget justification), and send them to our amazing grants person, by 28 November. (That’s actually kind of a “drop-dead” date—I’d like to have the budget numbers to our grants person a bit earlier so he has time to assemble all the budget stuff on his end and is not scrambling.)
3. Write my conference talk. I’m giving a talk at a conference (workshop) in early December, but I’m taking my family with me (and my mom too—so excited!), so I would rather spend the conference week hanging out with my family and not stuck in my hotel room frantically assembling slides. So, by 30 November, I will have my talk completely done: slides finished, AND a few run-throughs of the presentation under my belt.

So, how’d I do?

Goal #1: Nailed. I was most proud of myself for this, particularly since my self-imposed deadline was two days after fall term classes ended (i.e., the busiest time of the term). As I mentioned in the post linked above, this was a huge win for me because it was the first time I allowed myself to send out something less than “perfect” for review.  I’ve already gotten some feedback and am waiting for some more, which should come in next week.

Goal #2: Met, a tad late. I conveniently forgot how long end-of-term grading takes. Particularly when you have 34 projects and 65 final exams to grade in a short time span! These goals all fell in the second half of the month, during the grading crunch. I was off by a couple days with my targets for each of them, but I made great progress in the past couple of days to catch up. Right now the only supplementary document left is the RUI statement, and that draft will either be finished later this afternoon or Monday, still in plenty of time to meet the deadline (Dec 17!). And I actually got ahead and started working on some other supplementary docs; those should also be finished this afternoon.

Goal #3: Partially met. I have a complete draft of the talk; all the slides that need to be there are there. I need to add/modify some visuals, do some run-throughs, and edit edit edit. It’s not done, but it’s in awfully good shape, one of the better talk drafts I’ve done.

I am thrilled with my success in this project. I didn’t fully meet my goals, but I came very close, and that makes me very happy. I also learned a few things about myself along the way:

1. My perfectionism goes into overdrive when I’m working on less-familiar tasks. At this point in my career, I can knock out a conference paper with little angst, because I’ve done it a zillion times. I’m reasonably able to do the same with journal articles now that I have a few under my belt. But grant budgets and RUI statements? That’s unfamiliar territory, and that’s where I floundered.  Having the AcWriMo framework actually helped here—the goals and public accountability kept me from spiraling too deeply into perfectionism and helped me move forward, albeit more slowly than I’d hoped.
2. I love external validation. Ok, this wasn’t exactly news to me. While I have a crazy amount of self-motivation, the fact is I love the external motivation I get by making my goals public, whether that’s announcing my intentions to a group of friends over coffee (and updating them on my progress weekly) or to the world via a hashtag. I feel good when I meet an internally-imposed goal, but I feel great when I can announce to the world that I’ve met a goal.
3. Slow and steady wins the race. My grant deadline is 18 days away, and I am not panicked. I am in great shape. My drafts are solid, and I have a solid plan to make them even better and more polished, without killing myself in the process. I am not panicked and my drafts are solid not because I’ve spent large swaths of time on the grant, but rather because I’ve spent small amounts of time every day on the grant. (In fact, often I am most productive when I only have 30-60 minutes in a day—that’s when I’ve done some of my best work, because I am forced to focus.) And because even before AcWriMo, I spent small amounts of time every day for months working on the ideas, reading the literature, and preparing for this. This is true not just for this grant, but for all my work generally—slow and steady, a bit every day, is how I progress.

Even though AcWriMo ends today, I will continue working in an AcWriMo fashion moving forward. I’ll set ambitious goals for myself. I’ll commit to working slow and steady every day. I’ll continue to make myself publicly accountable to friends and strangers alike. And I’ll work on keeping my perfectionism in check. Thank you, AcWriMo, for a wonderful experience!

#AcWriMo mid-term (sort of) evaluation

We’re more than halfway into #AcWriMo, and I’m pleased to say that I’m making progress beyond my wildest expectations. I have already met my first goal (send draft of grant narrative out for review), and I’m working towards goals 2 and 3 (supplementary grant docs and conference talk, respectively). Today’s “writing” task is to sit down and, just as I did with the first goal, come up with micro-goals for the other goals.

I’ve been most impressed with my ability to stick to my self-imposed deadlines during this process. As a recovering perfectionist, I’ve often set deadlines for myself, only to ignore them because “this is not ready yet” (translation: it’s not perfect) or because “I’m too busy to meet that deadline” (translation: I don’t have time to make it perfect so I’ll procrastinate). This time, I’m making AND meeting them—I have yet to miss a deadline. (I’ve even resorted to working on Friday nights to meet my weekly goals, if I can’t get everything done during the work week.) The biggest test was meeting my Friday deadline. My draft wasn’t up to my usual standards, and I thought for a brief second of letting that deadline slide. Instead, I sent it out in its existing state, trusting that the feedback I’d get from a slightly more tweaked version of the draft would not be significantly more helpful than feedback I’d get from the draft in a “rawer” state. For those who know me well, this is a huge step forward for me!

I’m also impressed that I’ve been able to prioritize, even during the crazy busy time that is November. You see, our term ends in November, finals are going on as we speak, so I have an insane amount of grading and end of the term accounting to do. Also, I’ve had some other deadlines I’ve had to meet. AcWriMo has allowed me to prioritize and to “selfishly” claim some time for my research and writing, in a time when it would normally fall by the wayside. And I’ve been honest with my students: I’ve mentioned that I’ve had to meet some research deadlines, so grading would be a bit slower than normal. For the most part, they’ve understood and been supportive.

One more thing: I’ve been able to stay on track, stay sane, slay my procrastination demons, etc…all while lacking a key stress regulator in my life: exercise, and specifically running. I’ve been recovering from bronchitis, and just started exercising again last week. But I’ve also been injured, and haven’t run since August. Normally this would drive me stark raving mad, so I’m really pleased with myself that I’ve kept my stress in check and worked so productively without these very important mental/physical health elements in place.

The second half of AcWriMo will be even more challenging than the first half. Final grades are due next Wednesday, and I have 2 projects and 65 exams to grade between now and then. Recommendation letter season is starting, and I have a long list of students for whom I’m writing letters this year. Plus there is the Thanksgiving holiday this week. Keeping my focus, sticking to my goals (micro and macro), and remembering that “done is better than perfect” will surely get me through these challenges and onto the finish line!

Yes, I am insane: My AcWriMo goals post

(tap tap) Er, is this thing on?

Hello. You may remember me. I used to blog in this space….well, maybe not regularly, but certainly more regularly than of late. I’m still here, really. I’ve composed a number of blog posts….in my head, driving to and from work, which seems to be the only free time I have lately. It’s that whole transferring the posts from my head to the screen that’s proved difficult lately.

In short, this term is kicking. my. ass. in a serious way, and continues to do so.

So why on earth would I decide that this is a stellar time to participate in AcWriMo? That month-long festival of crazy writing goals for academics similar to NaNoWriMo? That takes time, and energy, and planning, none of which I seem to have lately. And I’m waaaaay far behind on everything else (see especially: grading of daily assignments for my 2 classes, sleep), so really, shouldn’t I focus instead on the immediate?

Actually, it’s precisely for these reasons that I’m participating in AcWriMo. I’m one of those annoying people who preach the value of prioritizing research time, even (especially) when the rest of life gets crazy. “Just 30 minutes a day!” I say. “Do some pomodoros!” I say. “It’ll make you a better teacher, and you can write articles and run experiments and analyze data that way!” I say.

Ask me when was the last time I worked on my research for even 10 minutes.

The thing is, when I don’t make time for research (just like when I don’t make time for exercise or sleep), I get cranky, and I don’t work as effectively in general. Research really does ground my teaching. Plus, I kind of have this big grant deadline coming up in mid-December, and while I’ve been making slow and steady progress, my lack of attention to research lately has brought my progress to a halt. I want this to be a decent grant application, not a half-assed one, yet if I don’t jump-start my progress on it then it will be half-assed, and I’ll be disappointed in my effort.

So, I’ll be participating in AcWriMo 2012, tweeting about my progress periodically (with possibly the occasional blog post). Because part of AcWriMo is making your goals public, here are my goals for the month:

1. Finish the grant narrative, and send it out to colleagues for comments, by 16 November. (The grant narrative is the “meat” of the application, and is the technical explanation of the work and deliverables.)
2. Finish the supplementary documents (RUI statement, budget outline, budget justification), and send them to our amazing grants person, by 28 November. (That’s actually kind of a “drop-dead” date—I’d like to have the budget numbers to our grants person a bit earlier so he has time to assemble all the budget stuff on his end and is not scrambling.)
3. Write my conference talk. I’m giving a talk at a conference (workshop) in early December, but I’m taking my family with me (and my mom too—so excited!), so I would rather spend the conference week hanging out with my family and not stuck in my hotel room frantically assembling slides. So, by 30 November, I will have my talk completely done: slides finished, AND a few run-throughs of the presentation under my belt.

I’ve sketched out where I am and where I need to be and I’ve set some achievable micro-goals along the way (the first of which is due this Friday….yikes!), so I’m excited to get started.

If you’re participating as well, I look forward to working virtually with you this month! It should be an interesting ride.

You can’t go home again

Next week I’ll be in Chicago for the NCWIT summit*. Carleton’s an Academic Alliance member so I’ll be representing us there. I always look forward to the summit—this is my third one—but I’m especially excited about this one because Chicago is my grad school home.

I’ve been back to Chicago a number of times since I graduated, but I’ve never made it back to campus. My advisor left right before I finished, so that’s one big tie back to campus that’s not there anymore. But I still do have a number of ties there, and so this year I decided that I’d make the time to go back and visit my home of 5 1/2 years.

I’ve spent a good part of this week setting up meetings and letting people know I’ll be on campus. It’s awesome (and a bit weird) how many people remember me. Especially since it’s been…let’s say, a number of years…since I graduated. Of course I’m eager to discuss research with like-minded people, so I dutifully did some poking around on the various labs’ and groups’ sites.

In the process of poking around, I learned two things that really shouldn’t shock me anymore, but still did:

1. I have zero research interests in common with my old lab. Zero.
2. My research interests are much more aligned with the CS faculty than the ECE faculty.

Now, you are probably thinking “well, DUH! You are a computer scientist, after all!”  And yes, you’re absolutely right. I’ve identified as a computer scientist for at least 9 years now, and probably longer since the switch to CS really happened during my post-doc days. But part of me still identifies more as an electrical engineer. That was my undergrad identity. That was my grad school identity. That’s what I thought I was going to be when I grew up. Identities are hard to shake, apparently, even if they don’t quite fit anymore.

The thing is, shaking that identity, and taking some risks to do so, opened up a world of possibilities that wouldn’t have existed had I stayed the course. My postdoc, my current position, and all the research opportunities of the past…bunch of…years, none of that would be possible if I hadn’t decided to assume a new identity as a computer scientist. And of course, it was my time at my grad school alma mater that put me in the position in the first place to make that identity switch—where I gained the confidence in myself, and constructed a support group, and worked on the right research projects, to allow me to ultimately explore and eventually assume the computer scientist identity.

So I’ll visit my old lab and my thesis committee and reminisce a bit about my engineer-self. And I’ll make some new acquaintances as my computer scientist-self. And I’ll feel equally comfortable in both worlds, even if I can’t exactly talk research with my old lab anymore.

* If you are a reader of this blog and will be at the NCWIT summit next week, please introduce yourself and say hi!

 

The (research) definition of insanity?

This morning I had a deeply satisfying research session. I sketched out plans for a new testbed (related to the grant application I’m working on currently). I defined “roles” for each system within the testbed. I identified the main research questions, and even set next steps. I was feeling good.

One of the “next steps” involves solving a rather tricky problem…one that sounded familiar. On a hunch, I went back into my research notebooks….and found notes on the same tricky problem, from at least three different occasions, dating back to the spring of 2010. A tricky problem that, obviously, still remains unsolved, despite my best efforts.

The famous quote by Albert Einstein states that “insanity is doing the same thing over and over again and expecting different results.” Clearly solving this problem is necessary to moving this particular line of research forward. Clearly I’ve not been able to jump this hurdle in the past. Does this mean I’m foolish for trying again?

On the surface, yes, I am crazy for repeating my past failures. But on the other hand, each one of my previous attempts taught me something about the problem, something which I applied the next time to the problem. And each previous attempt also taught me that I was not quite ready to solve the problem—I didn’t have enough information, didn’t know enough about how the system I’m designing and developing behaves. In the interim, while I was off solving other problems and tinkering with the system, I gained (or so I’d like to believe) valuable insights and knowledge that should get me closer to the solution.

And then there’s that little thing about how my biggest breakthroughs tend to happen when I revisit old failures…and finally see that little nugget I failed to notice before, the key to solving the problem.

So as usual, I’ll plunge ahead in my normal insane way, trust my instincts, and hope that the fourth time’s the charm.

Image credit: http://www.rudecactus.com/

Spring term projects

It’s Week 5 of Spring Term here at Carleton and I’m wondering where the hell the term went. I’ve been so busy I’ve barely had time to catch my breath, or blog—and this is supposedly my “light” term, since I’m only teaching one class. Ye gods.

So what have I been working on? As usual, too many things, but I’m enjoying all of them immensely. Here’s a snapshot of what’s keeping me out of (too much) trouble these days:

Teaching: Flipping for the flipped classroom

I’ve been reading about, and contemplating, the “flipped classroom” concept for quite some time. (The basic idea is that the “lecture” part moves outside of class—students do this part on their own—and classtime is spent on “homework”—more involved problems that get the students to grapple with and apply the material, under the guidance of the professor, in groups.) Since I’m only teaching Intro this term, and since I’ve taught Intro so many times that it’s second nature to me, I decided to experiment with flipping my Intro classroom this term.

I made some modifications to the model: I don’t post any video lectures, nor do I have an online discussion component. What I have done is assigned very targeted readings and exercises before each class. Most of the exercises the students do on their own. I frame these exercises as a self-check: “after doing today’s reading, you should be able to do the following”. The students hand in one exercise at the start of class. Sometimes it’s a programming exercise, sometimes it’s algorithm development. Based on the exercises that come in early enough and my past experience with certain tricky concepts, I will sometimes do a short lecture/example/summary at the start of class. The bulk of classtime is spent having the students work in pairs on programming problems. I usually have them work on a problem for 1-2 classes, and then hand in the finished product a few days later (so that they have time outside of class to finish up if need be).

The main thing I’ve learned from this experiment is that I talk way too much! The urge to lecture is strong. But I’ve also seen gains in learning and, more importantly, tremendous gains in confidence in my students using this model (particularly when I shut up and get out of their way). They’re working on what’s traditionally been an angst-ridden, tricky project involving datasets and lists this week, and I’ve been amazed at how smoothly it’s gone. They’ve even found conceptual issues with the textbook, on their own. I’ll admit, I’m flipping for the flipped classroom!

There are some changes I’m contemplating. I see myself posting some videos in the future (particularly around class development and recursion), and I may add a discussion board in the second half for students to post (and respond to) questions about the readings. But so far, I’m very pleased with the results. It will be interesting to see what the students say on their midterm evaluations about the model, to see if they are as enamored as I am.

Research: Writeups and grant apps and data, oh my!

My research life is where I feel torn in many different directions. On the one hand, I’ve got a tremendous backlog of data to analyze and stories to tell about that data. I figure I have 3 more papers worth of stuff there. So I’m trying to get those out into the review stream. One’s close, but I’m having problem with the framing. One needs a lot more data analysis before the story becomes clear. And one I haven’t even started yet.

On the other hand, there’s this new research I want to get moving on, on self-healing networks. Even here, I’m pulled in 2 directions: extending my current work, to develop general system heuristics for quality of experience prediction; and extending the general idea of self-healing networks into the home networking space specifically. Thanks to our IT folks, I’ve been able to expand my testbed so that I can try a whole bunch of crazy things, but finding the time to design and carry out experiments is tricky. But I need the data from the experiments so I can write the necessary grants so that I can do more experiments.

When the number of hours I can spend on research each week is in the single digits, finding the right balance between writing and new research is a struggle. I’m still trying to figure out that balance.

Life: Keeping up with the kiddos

Someone told me that going from one kid to two is an exponential increase in work, and darned if that isn’t true. My son’s been home for 4 months now, and I’m still trying to get the hang of being a parent of 2. Not to mention all of the adjustment issues inherent with adopted kids and the daily assertions of independence from my almost-5 year old. One thing my lighter teaching load has afforded me is the opportunity for weekends mostly free of (school) work, so I am enjoying the extra family time. I just wish these kids didn’t exhaust me so thoroughly!

The big project on the home front has been the post-adoption stuff. Reports and social worker visits at 1 and 3 months, plus another one next month. Paperwork, paperwork, and more paperwork. We’re also starting the process of finalizing the adoption, which means….yep, more paperwork. But if all goes well, the adoption should be finalized this summer (and then we get to apply for citizenship and a passport and a SSN…the fun never ends. Yippie.). And in the end, it all means we have this wonderful, funny, happy, loving little boy in our family, so of course it’s all so very worth it.

So that’s a snapshot of life, spring-term style. What are you up to this spring?

Image source: http://fgd.trebec.org/posts/fun-and-games/

Five (academic) things I’m thankful for

It’s 2:30pm the day before Thanksgiving, and about 20 minutes ago my brain decided that we were done working for the day. So before I sign off for a long weekend of low-key relaxation (and football! and running! and Christmas decorating!) with my two favorite people in the world, I wanted to reflect on five academically-related things for which I am thankful this holiday season.

1. My students. I’ve spent the past few hours writing recommendation letters, and the exercise reminded me of just how amazing our students are. Sure, sometimes they drive you batty, but I can truly and honestly say that I love my students. They are so sharp, so smart, and so engaged (ok, maybe less so during weeks 9 and 10 of the term). It is a joy to teach them and to work with them, and I am truly grateful to have the opportunity to teach at a school like Carleton.
2. My colleagues. I am fortunate to work in a great department. We don’t always agree, but we listen and learn from each other. We work on making each others’ lives easier; we step in and pitch in. When I had to switch my upcoming leave from Spring to Winter at the last minute (see below), my chair and the rest of my department went into overdrive to help me figure out the logistics. I can go to any of them for advice or commiseration without fear of judgement. (Having tenure helps, but still.) They make my work life fun.
3. My research. As I’m sure I’ve mentioned her before, I am passionate about my research. I’ve worked on my current project for almost 10 years now, in various forms, and it still excites me. There are still so many questions left to answer! And I truly and honestly believe what I am doing has the potential to be life-changing (or at least paradigm-changing), which fuels my passion. I am thankful that my job gives me the space and the freedom to explore the questions that motivate me, without question and without oversight.
4. A term’s leave. I’ve had a pretty busy year, to put it mildly. My students started working in my lab this summer the day after Spring Term ended (before I had started grading my finals!), and we went pretty much right from the summer research time into a very busy fall term (with the dyad and various other service projects on tap). So it’s been about a year since I had a proper break. I was scheduled for a leave during Spring term, but due to some things going on in my personal life (more on that in a later post), I will be on leave next term instead. I’ll still be plenty busy, but this is the right time for a break in the routine and some time away from Carleton to refresh and rejuvenate before Spring term. I’m thankful that I have a job that’s flexible enough to allow me that much-needed time away.
5. My mentors, sponsors, and cheerleaders. My recent trip to Grace Hopper reminded me of how energizing and powerful an effective support network can be. I’m grateful for all those in my life, inside and outside of my institution, who listen without judgement, offer advice, open doors, and open my eyes. I have some great people in my network and I would definitely not be where I am today without their support, encouragement, and facilitation.

I hope all of my US readers have a wonderful Thanksgiving holiday!

The 5 Stages of Conference Paper Writing

(with apologies to Elizabeth Kubler-Ross)

1. Denial.  This is going to be the best paper we’ve ever written! It’ll be paradigm-changing and highly cited! The data is so awesome it’s going to be a breeze to write! We totally have enough time to whip this into shape and send it out.

2. Anger.  What do you mean, we made a mistake in the experiments and now the data is not as conclusive? There’s no way we can spin this positively! This intro reads like it was written by drunken monkeys! We’re how many pages over the limit? And why is it that I just removed 2 whole paragraphs of text in LaTeX, but our page count just went up? Gah!

3. Bargaining. Ok, if we combine Figures 2 and 3 into one figure, we can gain back 2 lines of text. And if we cite Author X that’s kind of like citing Authors Y and Z, so we don’t actually have to include all three citations, so that gains us 6 lines of text. Which means now we’re…still 1 page over the limit. Let’s try cutting one sentence from each paragraph…Do we really need a conclusion? I mean, no one ever reads them anyway…

4. Depression. There is no. way. we’re going to make this page limit. We’ve already cut this paper to the bone and we’re still half a page over the limit. If we triage any more of this paper, we’re going to lose some content and context, and then the reviewers will ding us for not including the context and content.  And this intro still reads like it was written by drunken monkeys. We are the worst researchers ever.

5. Acceptance.  Wait, did we just make the page limit? Submit it now! No no no, whatever you do DON’T re-read it. I don’t care if it makes sense or not anymore, page limit met = done.

Writing backlog

Chances are good that if I ask you what comes to mind when I say “computer science”, “writing intensive” is probably not in the top 10.  Yet my job is very writing intensive, even when you take coding out of the equation.  On the teaching side, I write lectures (or “class plans”, depending on the day), handouts, assignments, exam problems, lab exercises, web posts, emails, reports, etc.  On the research side, I write grants, research papers, tech reports, howto documents for my research students, summaries of research articles, article reviews, and of course emails. Writing is the currency of research:  if you don’t write up and publish your results, it’s as if the experiments never happened, as far as the community is concerned.

This summer, I have a pretty severe backlog of writing projects.  I have 4 articles in various stages of completion:

• a journal article that was just accepted (finally!) that needs a couple of minor changes
• 2 conference papers outlining new experiments that we did last summer and winter, respectively.  One was sent out for review in the fall and rejected.  The other is at the almost completed draft stage and has not been sent out for review.
• 1 conference paper that’s purely in my head now.  We’ve got all of the data for it, but we never completed the analysis, so I have no idea if what we have is even publishable.

Normally I have writing projects in various stages, but this backlog is probably my worst in a while.  (And I haven’t even mentioned that I’ll be sketching out grant proposals this summer, too.)  Part of the issue is that I spent a lot of this year spinning my wheels, research-wise—I had students working for me all year, progress on some fronts was slower than anticipated, I had to redo some experiments and analyses, and I made some false starts on the 2 conference papers in progress.

Ideally I’d like to make progress on all of these this summer, while still leaving time to, oh, do actual new research, prep my fall classes, and pay some attention to my research students!

So, how do I prioritize?  Clearly getting the journal article finished and out is priority #1, and will also be the easiest to do (theoretically anyway).  And clearly finishing one of the drafts of the 2 papers-in-progress should be the next priority, so I can get that into the review pipeline.  But which one?  I will have to do some calculus to see which one I think is more likely to be accepted sooner rather than later.  In fact, you could argue that I should just get all the drafts-in-progress out the door and under review—clear the decks.

But.  Part of me thinks I should place some priority on that last paper, the one in my head.  I can and do get research done during the school year, but it’s often 30 or 60 minute blocks here and there.  I suspect to make traction on that paper, I’ll need some longer blocks of time to flesh things out, set up analyses, play with the data, etc.  But will doing so sap some much-needed time and energy from thinking about the more complete papers, and put me further behind on everything?

How do you prioritize your writing projects, or any other projects you juggle?