depth first search

As you can probably tell, I haven’t been doing much more than micro-blogging. Two weeks in Thailand followed by all the catchup here at school has stood between me and putting together new posts. Besides some Google-China conspiracy theories, I don’t really have much lined up in the future. I’m hoping to continue to blog about my research, but that seems to be dominated by my efforts to replicate certain existing algorithms on the way to generating my own. It’s a slow process that I’ll write about when I’m finished, but not anything I’m prepared to comment on now.

On the technical front, I’m now an iPod Touch owner, I recently upgraded my work computer to 64-bit Ubuntu, and I bought a new router. The iPod touch has been heavily reviewed elsewhere, the 64-bit Ubuntu upgrade was surprisingly uneventful, and the router just works (probably because it was not made by D-Link).

So that leaves Thailand pictures, which I’ll go ahead and outsource to someone who has both a better camera and a better eye than I do. Check it out.

Oh boy. After a couple of wasted hours of futzing I was finally able to fix the column width issue. All was not lost, as a basic understanding of CSS/HTML is a good thing to have in the toolkit these days. A few debugging tips:

1. Use the Firebug Firefox extension.
2. Look for the largest containing box with the issue.
3. Modify IDs not Classes.
4. Sometimes the obvious keyword is not the right keyword (e.g. max-width versus width).
5. Change one thing at a time.
6. Be sure to remember how to reverse any change.
7. Ditch old IE workarounds on principle.

I should probably figure out why WP-Syntax boxes force the left column of this theme to the left. I’m hoping to crowd source this problem to the spam bots. Any Akismet captured commenter that knows how to fix my CSS will get unblocked for life. Free link dumps!

Sigh.

I’ve been playing around with a new theme by Onehertz. Since I was accumulating all kinds of side-matter, I felt the need to move to a three column format. This particular choice seemed the most customizable and has numerous options under the hood (so many that the theme has its own wiki).

The best part, however, are the little user widgets embedded in the theme. Try clicking the little icons in the top right corner of each post or clicking on sidebar headings. Enjoy!

When I was growing up my parents went through some lengths to make sure that I was well-rounded. The term well-rounded seemed to signify a kind of suburban trivium of sports, music, and various other school activities, though upon repetition, lost any kind of original meaning and because something more like an abstract concept — one that was easier to identify in the negative than as any sort of positive goal.

As I’ve grown older I’ve certainly dropped a lot of hobbies and have been slow in collecting new ones. My interests, perhaps naturally, have narrowed. I have a singular professional goal and a small number of things that I do for fun. However, working to become an expert at a small number of important things is a skill that, how should I say this … I am still trying to develop. I don’t think this is a product of my upbringing, at least not in any superficial sense.

One of the things I, sort of perversely and (a little) narcissistically, continue to do for fun is write blog entries on the web for the benefit of spam bots and perhaps a handful of friends and family. Part of the problem with my blogging is that my entries span a wide gamut from obscure non-general-audience computer science stuff to food/photo blogging for everyone. In between you’ll find the odd book related entry and many entries on politics.

Well-rounded blogs are not as desirable as well-rounded children.

So as an effort to practice trying to sustain a narrow focus, I’m going to cut it out with one category of entries — the entries on politics. This has numerous benefits beyond just improving my own ability to focus. For one, whenever I look back at some politically themed entry I wrote in the past, I feel a small amount of shame. They just aren’t really that good. If I have a point, it’s usually something that seems clever until I come back to it a week later.

[Aside: I'm beginning to believe that part of the problem with political discourse in this country is that there is entirely too much of it. I'm sure that some of my political beliefs are not really the product of rigorous reasoning and more the result of simple repetition.]

Another reason is that there are plenty of good political blogs out there, most attached to money making outfits with some caché and bullpens full of professional supporting staff (or so I imagine). This isn’t really the same wild west environment that prevailed when I started putting words online circa 2000 or so. The wheat has already been separated from the chaff. I know which category I fall into.

So from now on I’ll relegate my left-leaning political sensibilities to shares in the sidebar and the odd Quote of the Day/Twitter. I expect the book/food/photo/CS blogging to continue.

• For some reason, having daily Twitter digest updates really annoys me. Who wants to read a blog that consists of digests of low content Twitter messages? I’d turn off the feature permanently, but my OCD demands that I archive all my online prognostication in one place. So I’ve switched to weekly digests. But you don’t care, since based on statistical analysis of my logs, your’re most likely a spam bot.
• The quantity of potential paper targets for my research has exploded, and I’m stuck finishing class projects.
• Check out The Parodists Paroxysm and Whimwit.
• Compiz in Ubuntu 9.04 apparently does not like Intel graphics cards. Lame.

I recently redesigned the header for this blog, using this neat header creation tool on a photo I found while surfing Wikimedia Commons. In previous iterations of the header, I tried to “design” my own based somewhat on the abstract notion of depth-first search. My first effort was marginal at best:

During a lull one particular Saturday, I decided to attempt a redesign:

The best that can be said about either of these headers is that they didn’t blink. I’m happy to admit that anything dealing directly with aesthetics requires skills that I simply do not have. I know what I like, certainly, but I’m not deluded into thinking that my own attempts had any merit, I always just find it frustratingly difficult to create something that looks okay.

With this latest effort

I was able to find something that I think passes muster. The original photo is interesting, and the header is just-too-overcropped, making it web 1.5 hip. So I guess the takeaway is that, instead of rolling your own design, wait for somebody to create a web script that does it for you auto-magically.

I’ve been toying with the idea of trying to consolidate my online presence here. This would require folding in my professional splashpage into this blog, always a dicey proposition, but, let’s be honest, at least this isn’t Myspace.

I’ve been contemplating a couple of issues. One, have I sufficiently sanitized my site for a professional audience? I’ve periodically rebooted my online presence in the past in order to do so, and am wondering whether I need to do so again. I suspect not, since this last reboot was entirely aimed at creating a kind of outlet that would serve both my friends and family without offending any colleagues or students. That’s not to say that there aren’t posts that I’d delete if I remembered them – in particular posts of a political nature that deviate from my own (outwardly crafted and purposefully boring) center left political views.

I’ve read plenty of academic blogs with names attached, and think that I fall sort of in the middle of the pack with regards to tone and content. Indeed, some academics who don’t blog are often more likely to end up putting up ridiculous things on their own professional websites. If anything, my most important concern right now is whether my domain name (depthfirstsearch.net) is too tacky for professional usage.

The second issue is more technical. The blog used to be installed in the root directory of depthfirstsearch.net. Awhile ago I switched the install to the /blog subdirectory, thinking at the time that I needed a splash page and a wiki and all kinds of other stuff that I never really used. I have, however, been saved in the past on programming projects where a needed feature was easily added to existing code because previously unneeded layers of abstraction were included in the design. I’m thinking that keeping this blog installed in  /blog is the kind of choice that will allow me to move between options for an online web presence without breaking anyone’s permanent links.

The problem is that I’m not sure how to manage the page people first see when they visit depthfirstsearch.net. Should they see my professional splash page? Should they be redirected to the blog? Even if I had a splash page, is there a way to make it a link to a static Wordpress page so that I can use Wordpress as my one stop shop for updating my web presence?

I’ve relaxed the comment settings. Some posts have been getting more comments than usual (the usual being none), and I don’t want my own lazy moderation to impede any potential discussions.

That and Akismet is actually a pretty capable comment filter. I’d be hypocritical not to trust in the power of artificial intelligence.

I’ve added a new algorithm to my continuing series of Python implementations. A rather simple Metropolis-Hastings algorithm is ready for you perusal. All implementations in the series are now linked in the sidebar. Happy hacking!

I’ve also updated the site license. You are now free to copy and share anything written by me on this site (including the example code) provided you abide by the Creative Commons Attribution-Noncommercial-Share Alike 3.0 United States License.

Now for some figures. A histogram of the resulting sample sequence (starting from x = 10 and running for 2400 steps):

A time series plot of the resulting samples:

As McKay points out, the random walk behavior is undesirable and requires that a large number of samples be taken for such a small state space.

UPDATE: My plots and code have an error. I don’t actually reject the rejected samples. I suspect this is why the left and right most buckets have more samples than they should.