I've been searching on and off for quite some time for an audio player for Linux that suits my needs, and have thus far been unsuccessful. I find this somewhat surprising, since I don't think that my needs are particularly demanding:

• Full library browsing, with browse by artist and album funcionality
• Live find-as-you-type searching
• Ability to handle large music libraries (tens of thousands of songs)
• Finally, the focus of this post, shuffle-by-album. In case it's not clear, that means the ability to listen to all of the songs of an album in order, and then choose another album at random to play afterwards. This is particularly useful to me since I prefer listening to full albums, but I have enough albums that if I don't play them at random I'll end up listening to only a small subset of them.

Basically, what I'm looking for is iTunes, but not necessarily with its full set of features, though the interface is pretty much exactly what I'm looking for. Since iTunes is only available on Windows and MacOS, I basically just looked for an iTunes clone, and found Rhythmbox. Rhythmbox is great, for the most part, and works just the way I want it to (and I can even set it to watch the music library folder for new music automatically, which iTunes won't do). Lately, however, I've become frustrated with its lack of album shuffle, so I went looking for other alternatives.

Amarok is another very popular media player for Linux, so I gave that a shot as well. I realized pretty quickly that it wasn't going to work for me. My initial impression was that the interface is a total mess, and while I might have eventually become accustomed to it after a lot of use, I didn't want to bother:

More importantly, though, it seems to be more of a playlist based media player (like Winamp) than a library based media player like iTunes, Rhythmbox, or Windows Media Player, which is what I prefer.

I tried another program, Banshee. I had used Banshee for a while a few years ago, but stopped because it was slow or buggy. The latest version was much better, and it even had the option to shuffle by song, artist, or album. Sort of. The GUI options for shuffle by song and shuffle by album are both greyed out, and not selectable. Wondering why, I did a few quick searches and all that I could find was a single thread on the Banshee forums asking the same question, with zero responses. Digging further, I went into the Banshee development IRC channel and asked about it:

Me: i've got a quick question re: shuffle
Me: does banshee *actually* have shuffle by artist/album
Some dude: AFAIK "nearly" for a very very long time
Me: so, then, why are the options listed on the shuffle menu and grayed out
Me: instead of just not there at all?
Me: because, frankly, that's obnoxious and counterintuitive
Me: i just tried switching to banshee from rhythmbox specifically because i was tired of not having shuffle by album
Some dude: bug 348582
Some dude: needs fixing. i don't know why nobody's just done it though.

The bug in question: Random by Album. It was created nearly 3 years ago. After about 8 months, Banshee was updated to include a GUI widget to select shuffle type, but nothing in the backend of the application was actually added to provide the real functionality. Ever. The glory of open source is that I can go and look at the source code myself. So I did. The only piece of backend code that I was able to find that even referenced an ability to shuffle by album or artist was the following:

public enum PlaybackShuffleMode
{
  Linear,
  Song,
  Artist,
  Album
}

That's it. Why on earth would they actually bother to put UI elements in their software for features that don't actually exist yet? As far as I can tell, nobody is even working on it. I don't think I've ever seen that in a release version of a widely-used application.

This brings me to what I suppose is the point of all this, which is what I see as a flaw in the open source development model: frequent lack of control, organization, focus, drive, whatever you want to call it. I can promise you that this would never happen in an application from a major software company. However, the open source model clearly does work for quite a lot of other applications; what's the difference here? I think that it has to do with the number of competing offerings in a particular area, and it seems to have an effect opposite that in commercial software development. Having a large number of commercial software offerings targeted to the same audience tends to result in better applications; more competition drives further innovation to sell more software. This obviously isn't the case for most open source software; too many competing offerings dilutes available developer time, talent, and effort, and shrinks the userbases for each which results in less feedback and input. Since the developers aren't making money off of the software, they won't be as driven to compete. There are fewer developers to implement new features, to write bug fixes and to improve usability. Fewer developers means that the personal preferences of a few of them are much more likely to have a significant impact on the software, instead of focusing on what the users really want or need. As a result, we've got the situation that I saw with Linux media players. They've all got the general idea right, but the development is so spread out that the overall quality is less. In my personal experience, the software is less reliable and harder to use on average than iTunes or Windows Media Player.

Of course, like I said, some open source software is very successful, and works incredibly well. I thought about every single piece of software that I use on a regular basis, and almost all of the ones that I didn't particularly like have a number of competitors, whereas the great ones all seem to either stand alone or have one or two alternatives. Banshee, Amarok, Rhythmbox, Quod Libet, and a number of others are one example of the former, and there are a number of the latter: gcc, make, emacs/vim, pidgin, Firefox, the Linux kernel itself, etc.

I wonder how different Rhythmbox and Banshee would be if the developers had worked together to create just one application? Obviously at this point it wouldn't be feasible for them to combine into one application, but things might have gone differently if they had started together from the outset. I don't see anything in particular that distinghuishes them from eachother, why not have twice as many people working together?

At any rate, if you're a good user of open source software, you should have one thing to say to me right now: "Submit a patch." And you're right; this is the real solution for all of the above problems, and is the only way that I'm able to be using all of this free software right now. So I think I'm going to. I've never actually contributed to an open source software project before, and I think its about time that I do. It'll probably take me a little while to find time in my schedule to work on it, and there'll some potentially significant ramp-up time to get a general sense of how Banshee's codebase is organized. It's also written in C#, which I've never used before, but picking up another new language is never that difficult and is almost always useful. Call me out on this in a few months if I haven't done anything.

What's wrong with this picture?

Sparks is no more

This is a perfect example a problem with mainstream media which angers me more than anything else that I can think of:

"A recent study found that college students who mix alcohol and energy drinks engage in increased heavy episodic drinking and have twice as many episodes of weekly drunkenness. College students who reported consuming alcohol mixed with energy drinks also had significantly higher prevalence of alcohol-related consequences, like sexual assault and injury."

I would be willing to bet money that these people don't drink that much as a result of drinking alcoholic energy drinks like Sparks, but rather that they drink alcoholic energy drinks like Sparks because they are already heavy drinkers, and that not selling them prepackaged drinks like that won't change a damn thing.

Attention everyone involved in the reporting of statistical studies such as this one: correlation does not imply causation. Period. Please stop drawing bullshit conclusions from statistics that you are not qualified to understand or interpret in any meaningful way.

I've been meaning to post my thesis for a while, but hadn't because I've for some reason been paranoid about intellectual property issues. These worries, of course, were unfounded, so here you are:

Optimizing Live Virtual Machine Migrations using Content­-based Page Hashes
Virtualization systems such as VMware ESX Server use content-based page hashing to help identify duplicate memory pages within virtual machines. Such duplicate pages, once identified, can then be mapped copy-on-write to a single page in physical RAM, thus saving memory in a manner that is transparent to the virtual machine. The page hashes that are collected in this process can be further utilized. This thesis demonstrates how these page hashes can be used to reduce the time required to migrate running virtual machines from one physical machine to another. Normally, this is done by sending the virtual machine state and all memory contents from the source to destination machine. However, since some memory pages may already be present on the destination, it is possible to reduce the number of pages sent (and therefore total migration time and bandwidth used), by sending only a compact hash instead of the full contents for each page likely to be present on the destination machine. This thesis accomplishes this by creating a database of canonical or “standard” pages to determine which pages can be safely sent as only a hash. Tests of the system demonstrate up to 70% reduction in migration time in idealized workloads, 40% reduction in some realistic workloads, and minimal slowdown in some pathological cases.

The full document

Having concentrated in economics in undergraduate studies, I know enough about economics to know how much there is that I don't know.

In light of that, I wonder about what bright economists might feel about comments that bright (but not educated in economics) scientists and engineers such as myself make about the economy. Specifically, whether those feelings might be similar to the feelings that myself and others had towards Sarah Palin's comments about fruit fly research.

Shamelessly jacked from Preyas:

Page 56

Rules:
*grab the book nearest to you right now
*turn to page 56
*post the 5th sentence on that page along with these instructions

"Let us denote the dual tableau formed from P in this way by the symbol P^S."
The Art of Computer Programming, Volume 3 Sorting and Searching, Donald Knuth

My dad is making mead out of home-grown asian pears, and sent me this photograph of the brew process taking place on the kitchen counter:

zwieselite: if this were a movie
zwieselite: a matte black v-22 osprey would land on a remote tropical island
zwieselite: and one dude in a black suit with sunglasses
zwieselite: with 6 special ops dudes
zwieselite: would enter a tiki hut or whatever
zwieselite: and interrupt greenspan eating a bunch of grapes out of some filipino strippers' hands
zwieselite: and be like
zwieselite: "your presence is required in the lair of decision making"
davemallingsband: "i left that a long time ago. i can't go back"
zwieselite: "this isn't about you or me, alan"
zwieselite: er, sorry
zwieselite: "al"
davemallingsband: "get out of my thatched hut. i'm not the same greenspan anymore."
davemallingsband: and then two days later
davemallingsband: a phone call
...to be continued

(why I write computer programs instead of movie scripts)

Hubble Finds Unidentified Object in Space, Scientists Puzzled.

My guess? One of these suckers on the camera lens.

Ford is making a 65mpg diesel car, but not selling it in the US. Why? A number of reasons explained in the article, but I suspect the underlying cause is that Americans are scared of diesel; they associate it with big polluting trucks and like the idea of the fancy new technology of hybrids despite the fact that there are many diesel cars in Europe that get better mileage than available hybrids for less money. My hope is that Mercedes and Volkswagen bringing diesel cars back to the US will do something to undo this, but I think it's unlikely that it will do that much. People seem to have an obsession with all-new technologies over tech that's already here and has been improved over the years.

I think the situation is the same with nuclear power. People are irrationally scared: nuclear power is dangerous, and we don't have any way to deal with the radioactive waste. Nuclear power is in fact less dangerous than most of what is currently used. Compare nuclear plant deaths to coal plant deaths, you might find that in the US there have been zero related to nuclear power (and then factor in the deaths related to coal mining vs. uranium/plutonium mining). Further, while it's true that we don't have a way to permanently get rid of radioactive waste, we can contain it, which is much more than can be said of the emissions from coal, oil, and gas plants.

But hey, what do I know. "Nuke-ular" and diesel are scary words, apparently.