I had written a long post about getting a few apps on the iTunes App Store, and I might post more about it later, but in the mean time, I'm happy to write that I have two new apps available for download:

• New Moon Countdown - Team Edward Edition
• New Moon Countdown - Team Jacob editions

More about the apps can be found at the support page.

To market New Moon Countdown, I've decided to give a book from the Twilight saga to a women and children's center, school, or library in the Seattle metro area for every 100 downloads of the app through iTunes before July 31^th. These books have brought a lot of joy to young women and I hope that this will get books into the hands of people who may not have been able to read them otherwise.

I haven't written any documents in L^AT_EX since Dr. Taylor switched to an XML submission format. (Or perhaps it was during Dr. Sebern's Formal Methods class.) I started the paper portion of my thesis this weekend, and decided to delve into the wonderful world of T_EX once again.

There are several ways to set up a productive L^AT_EX environment on your Mac, but the combination that seems to work best for me is teTeX installed from MacPorts and TextMate as my editor. I've also found Skim which will automatically refresh PDFs when they are generated by pdflatex.

Alternative, there is a MacTeX, however, the download is nearly 750MB, and includes many packages which already come with OS X.

Here's a collection of sites I found useful when setting up my L^AT_EX workflow:

• Setting up LaTeX on Mac OS X
• LaTeX on Mac OS X
• LaTeX: Live PDF preview with TextMate and PDFView

If you are writing a paper, a book, or any significantly long document (there are even packages for resum?s and presentations), I'd recommend looking into T_EX or L^AT_EX. They are a refreshing change from the WYSIWIG editors we've all become so accustom to.

About 16 months ago I scratched an itch and wrote Spangle, an app that checks for updates to the applications you have on your Mac. It's worked well enough, and its largely stagnated since its initial 0.1 release.

Back in May, Pierre Rudloff sent me an encouraging email and a link to a French translation he had written for Spangle. I also fixed an anchoring issue with the icon in the upper left of the main window.

The most major change is that Spangle has been recompiled with llvm specifically targetting Leopard. This shouldn't affect much (except for people still using Tiger), but I was interested in seeing how seemless a transition to llvm would be.

So the 0.2 release is out and the appcast has been updated. Feel free to contact me with any comments, suggestions, or bugs.

I want to thank the Builder team for indirectly allowing me to have a call free weekend out of town. I'm sure other XML Ruby library teams could have done it (except one, that is); but Builder was the first alternative I came across.

I have a production Rails app that I recently migrated to new hosts. The setup isn't very complicated, and for XML generation, I chose libxml-ruby, which is built on top of libxml. What seemed like a good choice at the time turned into a perfect storm of after hours calls. Between a bug in Apache's mod_proxy_blancer, what seems like double free()s in libxml-ruby, and FastCGI not properly timing out workers, I found myself constantly monitoring and killing processes.

libxml-ruby seemed like it would be the easiest component to replace, and in my search for a new XML building package, I stumbled across Builder. It worked like a charm. I was able to strip out all of my libxml-ruby code and replace it with equivalent Builder code in about an hour. All my tests passed, QA approved it, it baked in production for half a day before I left for the weekend, and I had a call free weekend not worrying about my app's stability.

This morning I checked the logs: clean as a whistle. The last error was from just before I deployed. So, hats off to the Builder team. You happened to be fresh in my mind in front REXML.

I've spent nearly every hour of every night for the past 3 weeks, as well as 2 entire weekends grinding through the death march that was this ray and path tracer.

Its incomplete: the kd-tree implementation doesn't work properly and it generates odd sinusoidal noise when path tracing. For simple ray tracing cases, though, it looks really good.

The implementation was done in Objective C++ (an un-holy mashup of Objective C and C++). I enjoy programming in both languages, and there are features of each that I would like to see in the other, but using them both at the same time introduces complexity that projects with short deadlines don't generally need. On the other hand, I really like being able to intermix fast C++ code I have been using since college with elegant Cocoa. As long as you know that you have to give your Objective C++ files a .mm extension, it really isn't that bad.

Anyway, the point was that this project is in the can. I may have to revisit it later this semester, but for now, I can catch my breath.

Update: here's a shot of the path tracer with its ridiculous wave pool interference.

I've come to the conclusion that I don't like webapps. Some of them are cute or convenient, but in whole, they all have the same flaws: they require internet connectivity, they all reinvent common desktop application paradigms, and they don't communicate with any of the other apps that I use.

This may seem shocking, since I write webapps, a couple of which I think are pretty good. Many of the apps I work on exploit their web-by roots (they are traditional, page based sites which happen to be driven by some content management application), and all of the webapps have an API, so I (or someone more ambitious then myself), can write a decent client app which uses the web.

The majority of my favorite web sites don't make me visit them. I use del.icio.us through Cocoalicous or the del.icio.us Firefox extension. I use Twitterific to get to Twitter. I use Mail to get my Gmail over IMAP, the Package Tracker widget to get stats from USPS.com, UPS.com, and DHL.com, and for the most part, NetNewsWire to read periodical websites. Even Fat Charley has an RSS feed.

This is the direction I'd like to see web development take: Client/Server for the masses, with rich apps on the client and synchronization/storage on the server. Client apps can work offline (if nothing else, to show you the last state of the server), they use native widgets, but are designed to fit within their domain (a mail app looks different then a package tracker), and they all work together with my address book, calendar, and file system.

Not only that, but people are free to use the clients they want. I like Twitterific, but maybe iTweetr is more your style. There's more mail clients then can even be mentioned. And typically, Mac users get a different selection then Windows or Linux users, with apps on each platform integrating into each environment.

Decent client apps put even the best web apps to shame, and great client apps are in a completely different league.

Don't get me wrong. HTML over HTTP is great when I'm at an unfamiliar computer and need to check my mail. But why would I want to go to a website to read other websites (web-based RSS reader).

A year ago or so I was working on just that, a web-based RSS reader along with Mike Joyce and Nick Pazoles. We were nearly read for a public beta and the app was completely usable for a daily reader. The interface was web-2.0 to the hilt: it kept track of your read status, and even had some social hooks. But I lost interest as I realized I didn't want a web app. I wanted a real reader with synchronization. (which, thanks to NewsGator, I finally have).

I like great web app when I need them, but not for general use. If you're writing a web app, give me an API, some hooks, or some kind of way to get the data programmatically. I'll be very appreciative.

In the yesteryear of mid-2000 I bought, what was at the time, a big TV. Its a behemoth 32" CRT which weighs more than I do on most days. The power supply isn't properly shielded, providing a nice sine wave pattern in dark colors towards the left side of the display, the channel changes randomly after a few seconds after its been turned on, and the menu's don't display properly if the input signal isn't clean, but overall the picture is good, it was a decent deal, its the only TV in the house, and Kortney and I have no intention of getting a new TV any time soon.

We also don't have cable, or satellite, and in many cases, reception of any kind, without playing games with the antenna, which on the second floor (TV is on the first), so that we can get better overall reception. We've been toying with the idea of subscribing to a service; there are days when I miss Cartoon Network and Comedy Central, and Shark Week potentially makes the monthly fees worth it. The biggest win would be that we'd finally have a clear picture on our TV and we wouldn't have to monkey around with an antenna anymore. But that's all we really want - a clear picture - not 150 channels to veg out in front of. So instead, we've decided to get an ATSC tuner.

An ATSC tuner, what's that?! Most stations (at least here in the valley), have already started their digital broadcasts, which use a broadcast format called ATSC (classic television broadcasts in america were NTSC). The big difference between the two standards is that NTSC provides an analog picture and audio signal, while ATSC is digital. With NTSC if you got a weak signal, you got a weak picture. With ATSC you either get a picture, which is crystal clear, or you get nothing. There is never a degraded picture like with old style broadcasts.

In just over a year, all stations will be required to turn off their analog NTSC broadcasts and broadcast solely using NTSC. If you don't have cable or satellite, and rely on over-the-air broadcasts, you'll stop getting a signal, unless you buy an ATSC tuner, which will decode the new digital signals for you.

Unfortunately, ATSC tuners are both hard to come by, don't seem to be very good, and can be relatively expensive. The RjTECH RJ-1000ATSC is only $80, but the reviews are less than favorable. Samsung also makes the DTB-H260F but prices start drifting towards $200.

All is not bleak however. The government has heard the people's cry for free, over-the-air entertainment, and has started a program to help people ease into digital television. The department of commerce has created a program to give people a $40 coupon towards the purchase of up to two bottom of the barrel ATSC tuners, which will be available in a month or so. Since the MSRP of the boxes is estimated to be around $50, that means you could potentially get a ATSC tuner for $10!

Now, don't go thinking you can walk into your favorite consumer electronics store and plunk down your $40 coupon and get the Samsung tuner for $120. The eligible tuners can't have USB or Firewire outputs, or digital coax or optical audio output. But you will get standard features like program information (channel number and name, program name, time slot, and rating). If you're lucky, you might even find one with S-Video output. Some of the models that will be available shortly are the RCA DTA800, the LG/Zenith DTT900, or several from Magnavox/Philco. An article just popped up on Slashdot as well.

I like how several commenters on Engadget think that only your grandmother will need these converters. Note to the intarwebs: not everyone cares about (or had a TV which can accept) HDMI, component inputs, or getting a new TV every 36 months. I'm looking forward to finally getting decent reception at a great, non-recurring price!

In episode 10 of The Talk Show, John Gruber hits the nail on the head, but this time it isn't apple related. He's getting caught up on Lost and has gotten to the second season and has the same reaction I had... not as good as the first season.

Speaking of John Gruber, he posted an link to D/Objective-C bindings on the Linked List today. I was introduced to D several months ago by a friend and was always interested in implementing something written in it but didn't really have a itch that Ruby or Objective-C couldn't scratch. Now it seems like I can have my cake and eat it too because Michel Fortin has released D/Objective-C bindings. Looks excellent: this means you can write Mac apps in D with full access to the libraries and APIs you'd typically be required to use Objective-C to access.

For those not in the know, D is an language similar to C which adds a full object oriented framework, C++/Java/C#-esque syntax, closures, namespaces, and garbage collection among other things. One of the best parts of the language is that it compiles down into a binary object which can be called by C libraries. D could potentially replace Objective-C as the Ruby of the compiled-language world.

There are two kinds of people in the world:

1. People who like their Mac.
2. People who haven't used one.

Mark Cuban uses a phrase I've been telling people who haven't "switched" for years: once you go Mac, you never look back.