The Topology of Bendless Three-Dimensional Orthogonal Graph Drawing, a new paper I wrote, is now online at the arxiv. This is basically the material from a series of blog posts I made on "xyz graphs", written up formally in more detail and formatted as a paper, although there are a few additional results not in the posts.

Together with my papers on flip graphs and simplicial arrangements, this marks the third time I've written a series of blog posts on a subject before realizing that my half-baked noodlings had somehow morphed into the content for a paper. Or maybe the fourth time, if you count the triangle center paper that grew out of some web animations I made.

Although I don't tend to get a lot of feedback on my more technical posts, something about the blogging process makes working this way feel more collaborative than when I'm working on a problem by myself without blogging about it. Just as I would when I'm discussing a problem at the whiteboard with someone in my office, I'm spending less time purely thinking about a problem and more time drawing pictures and trying to explain clearly what I'm thinking. And I think that it's helpful to interleave this sort of communication with the more traditional style of doing research while staring at walls or taking walks or taking showers or trying to fall asleep: it helps clarify the parts of the problem I'm working on that I already understand and focus my attention on the parts I don't. I know, I could just keep a notebook instead of posting things on blogs for the world to read, and sometimes I do, but that wouldn't supply the same kind of pressure to write in a way that would (I hope) be understandable to other people. And when I do eventually write a paper on something I've blogged about, the writing stage of the process is easier, because I have the posts and their illustrations to use as content.

I'm wondering that I don't see more of that sort of thing from other math or theoretical CS bloggers, actually — the category theorists have an active collection of blogs with this sort of content, but it seems to be less common behavior among the rest of us. I suppose it's natural to worry about being scooped; I haven't yet found that to be a problem with my own blog posts, but, to be honest, I haven't been posting the ideas that I felt from the start had a strong chance of turning into papers, only the ones that started out feeling too small and then grew. And anyway, if one's blog posts succeed in getting someone else interested in the same subject, I'd think the more likely outcome would be a chance at collaboration and co-authorship.


It occurs to me that, in a lot of my own posts when I'm doing that sort of thing, it helps to have a sense that there is an audience out there. So I figured I'd note that I do read most of your more technical posts, and generally feel like I understand them reasonably well -- I just don't comment usually because I don't have the background to add something else to that understanding.
Thanks, it does help to know that. Though I do try not to pick the subjects of my posts for what I think my supposed audience might like — that's more on the basis of what I find interesting for myself.
For me, the main issue is that of being scooped I guess, along with the general feeling that if I write something stupid as a blog post, the level of shame is much higher than if I were to write something on a board and have someone point out my idiocies. --Suresh
Tenure helps a lot for learning not to worry as much about occasionally looking stupid, I think.
do I have to wait till I get tenure? I was hoping finishing the PhD would suffice.

I used to post more in that vein, though never a lot. I wish I did more, because I greatly appreciate it when you and others do it. Even if, or especially if, we aren't working in the same area.

A couple of things in my case are working against it --

1) I try not to post about my "main" research work. I don't want the pressure of having others judge my progress or lack thereof from postings on my LJ...and besides, I typically already have collaborators and discussions ongoing about that. That being said, I sometimes have ideas that are related to things I'm doing but aren't the "main" thing -- like you point out, writing about them is a great way to work out the research itch in a way that could potentially lead to fruitful discussion with others. Whether that's a paper, good comments, or whatever.

2) Right now I'm entering my 5th year and trying to focus on work related to my thesis. So I spend less time thinking of interesting things that might fit in the scope of 1). There is at least one thing I have notes for a blog post about, but in that case I keep thinking I want to do just a "little more" work before writing. (Repeat ad infinitum).

3) While I was working at Microsoft Research, I actually wasn't entirely sure whether these "side ideas" fell under the scope of the work I was doing there or not. So I tended not to write about them to be on the safe side. Now I'm at Berkeley again and feel less concerned about this.


I try not to post about my "main" research work.

I tend not to, either, at least until I have a preprint available. When I start posting about something that I haven't written up as a paper, it's usually more because I think it's too removed from my research or too small to make a real paper, so I might as well put it on my blog instead of just throwing it away. But then some of those posts have started taking over and demanding to be taken more seriously, and once I start posting on a subject it's easier to continue.

While I was working at Microsoft Research, I actually wasn't entirely sure whether these "side ideas" fell under the scope of the work I was doing there or not.

Yeah, better safe. I haven't talked at all here about some consulting work I was doing this summer, even though I hope to eventually get at least one paper out of it.