watch this  

Great article on managing your $HOME directory with source control (in this case, Subversion). It may be a bit of overkill, but it gave me some good ideas.

I have too many computers and it’s a hassle keeping my personal configurations (usually in the form of dot files) synched up. Whenever I setup someplace new, I always grab my .bashrc, .vimrc and a couple others… but they quickly get out of synch. So, using the ideas from the O’Reilly article, I put my most commonly used files in CVS on an easily accessible server. I chose CVS because it is basically ubiquitous. Now, I can “checkout” my $HOME anytime I need to setup in a new place and I can “update” whenever I need the latest.

A couple of things came to mind reading Downloading comics: threat or menace? over at Boing Boing.

One, some comics - namely Queen & Country - I prefer to read in the TPB. I think that reading an entire story arc at once can really enhance the experience. If you don’t believe me: go buy the first Y: The Last Man trade.

Two, my local comic shop is great. Great people, great selection of new and old titles, really great prices. But, their subscriptions aren’t worth a damn… I mean, if I don’t come in for a few weeks, none of my titles have been pulled or worse some have and some have not. So, I started thinking: there has to be a Technical solution for this. I’ll spare you the details, but I quickly went from envisioning ways of programmatically managing my subscriptions to designing my own Online Comic Shop. I can’t imagine that running a comic book shop is the most lucrative business. Online competition - be it from a “Amazon” Comics or P2P - is only going to hurt these local shops… but they have to compete more effectively…

Which brings me to my last point: sometimes online is the only place to get a comic… There is no way around it. For example, I missed Issue #4 of Powers Vol. 2 and now it’s out-of-print. Where’s the next place I looked? Ebay.

I generally agree with the notion that downloading may (eventually) drive business. In fact, I got hooked on Fables after downloading the (here it is again) out-of-print first issue. Great idea, Vertigo! Next?

The 2004 OnJava Reader Survery results are in. No big surpises:

• Java programmers use Oracle. It follows that .NET programmers use SqlServer…
• Java programmers develop on their Windows workstations, but don’t trust Windows for Production.
• Some Java programmers still don’t use an IDE and more power to us.
• Ant is still the most important Java development tool.

Howard hit several nails on the head in this post. However, I think he held back on one of the major reasons why people choose in-house development: hubris.

I have used the Tapestry framework for my last two major projects. It has made an incalculable difference. By using an Open Source framework, I was able to focus on meeting my requirements while, at the same time, having access to the deepest levels of the code if need be. Earlier in my career I created and used my own J2EE framework, no more. I encourage developers to use a framework, any framework. The benefits (shorter development time, less “plumbing code”, genericity, community support) certainly outweigh the costs (steep learning curve, not being able to do it “your way”).

A wikit? I am so underwhelmed.

I was really expecting something impressive and almost entirely composed of eye-candy… I’m still not sure where Dashboard is going to fit-in outside the novelty realm. I mean, little informational pop-ups are great, but they seem to be a dime-a-dozen. Is the contest winner just the first in a long line of too-small versions of standard apps? I guess we’ll see.

I am totally stoked for the Mac version of Doom 3 in February. There’s a good, in-depth preview over at InsideMacGames. Doom 3 will be the first game I’ve bought in years. I just don’t play many computer games any more, but I have followed Doom since the original shareware version. Configuring my old 20Mhz laptop with 4MB RAM to run Doom was the impetus behind my studying Computer Science!

Even though I don’t play games on the computer much any more, there will always be a place on my disk for one game...

I think I have finally settled on a backup strategy for my Mac. You’d think it would be easy—given Mac users’ fondness for backups. But, it wasn’t. It was quite difficult to find a quality backup solution that provided the features I required:

1. A bootable clone of an entire disk
2. Rolling, incremental backups
3. Some sort of scheduling functionality.

I have an external Firewire disk that I have divided into two partitions: one for a bootable clone and one for 3-days worth of rolling backups of my /Users directories.

There is certainly no shortage of articles and forum posts regarding Mac backup software, but - as is often the case - it really boils down to personal preference. After a bunch of research, I narrowed my final candidates to the following:

• SuperDuper!
• Impression
• RsyncX

While initially I liked the simplicity and interface provided by Impression, I quickly discarded it. It just didn’t do enough. I am already familiar with rsync ; I use it extensively at work to backup source code across multiple machines. So, I wanted to try SuperDuper first. What impressed me most about SuperDuper and what I think is the best thing it has going for it is packaging. From the downloadable disk image to the interface—this application is good-looking, intuitive and easy-to-use. It is very straightforward to create a bootable clone and the app gives plenty of easy-to-understand feedback. Most importantly, it seems to be the only backup software with useful documentation! Note to other backup software developers: include a manually with a “quick start” or “Making your first backup” section!

Unfortunately, two things are lacking from SuperDuper: rolling backups and scheduling. I know that scheduling is scheduled for the next major release. The deal-breaker was lack of built-in support for rolling backups. SuperDuper comes with some feature they call “Safety Clones”. It basically checkpoints your system software and ignores your data, letting you roll-back in the case of an application- or system-software malfunction. That’s great—but it’s the exact opposite of what I need. I assume something could be hacked together using AppleScript, but I’m not sure it would be worth it.

So, I fell back the Ol’ Standby: rsync. I will just say this about RsyncX: it has the worst, most user-unfriendly interface I’ve every used. That being said, the Rolling Backup assistant was helpful and I was able to use my familiarity with the standard rsync to quickly build a comprehensive backup solution.

As far as scheduling is concerned: I just added an entry in ”/etc/daily.local” for my full and rolling backup commands and called it a done deal! This way, I can check “daily.out” every morning to verify things completed.

All in all, it was a bit more painful than I’d hoped. But, in the end, I settled on a strategy with which I am happy.

Quick note: be sure to uncheck “Ignore ownership on this volume” as it is the default after partitioning and/or erasing a disk with Disk Utility. Without this, a disk will not be bootable and rsync’s file modification check will (wastefully) assume the file has changed.

What a great way to start the New Year: I have been looking for The Conet Project for a long time! Of course, my problem was that I could not remember what it was called. I very clearly recalled reading something about recordings of WWII code stations on Slashdot years ago. So, it was just a matter of coming up with the right Google search phrase…

Anyway, the Conet Project is a 4-CD set of recordings from so-called Number Stations. I won’t rehash what has been extensively covered elsewhere:

• Excellent Washington Post article
• Dark Side of the Band at WIRED gives a good overview.
• Another good Salon article. Also explains where the title “conet” came from: apparently, it is the sign-off signal on one station.
• Pitchfork gave the recordings an 8.0 in this review.

There is an interesting side-story about how Wilco used a recording of a woman’s voice repeating “Yankee Hotel Foxtrot” in one of their songs (and as the title of the album) and was then sued by the Conet Project. This WIRED article has the details. By the way, “Yankee Hotel Foxtrot” is a phenomenal album… So, it’s a shame (or, perhaps, a blessing) that you can now download all four discs plus the 80-page booklet for free!

CocoaTextile is a simple text editor for the Mac that will convert plain-text to an XHTML fragment using either the Textile or Markdown engine. I know the name is a little misleading, but I had originally intended only to support Textile, until I realized it would only take minimal effort to support Markdown as well…

Anyway, this is Not Another Weblog Editor. It is intended to be a helpful tool for anyone - like myself - who uses Textile or Markdown, but also wants to see a preview without reloading their browser. Since CocoaTextile uses Cocoa’s text view component, all of the standard editing features are included “for free”. This includes, but is not limited to:

• Spell Checking
• Multi-level undo
• Drag & Drop

In the future, I’d like to override some of the built-in features to use the corresponding Textile or Markdown tag (e.g. selecting the Italic menu item would surround text with underscores). Also, since the preview pane uses WebKit, I plan on adding support for CSS. Documents can be saved as plaintext or HTML.

This humble project is great fun. I am continually impressed by the power and depth of Cocoa.

Download: CocoaTextile-0.01.dmg

The obligatory screenshot.

While working on adding forecast support to my Quicksilver plugin, I was investigating web services (in particular SOAP) support in Cocoa. From the Xcode documentation it initially looked like Applescript was the way to go. However, a little investigation turned up a WebServices framework inside the CoreServices framework! Strangely enough, these two (Part 1, Part 2) articles on Bioinformatics at the MacDevCenter on O’Reilly were absolute Gold Mines!

What really impressed me about the Apple’s covert support for webservices was the WSMakeStubs utility. Basically, you feed it a URL to a WSDL file and it generates the appropriate “wrapper” code in the language of your choice (Objective-C, AppleScript or C++). Very slick. It literally took a matter of minutes to add client-side webservices support to my app.

Before discovering the WebServices framework, I found AEXMLTutor . In a nutshell, you can enter some webservices AppleScript and it will give you back a ton of feedback… Much more than Script Editor.

Once I can parse the forecast data from the NOAA, I’ll update my Quicksilver plugin. Won’t be long!