Most wiki-based documentation I have worked with has sooner or later turned into a tangled mess of outdated information. Some pages are maintained actively while others aren’t updated anymore and you never know how reliable the information still is. For a while I’ve been thinking about a simple technical solution to this problem. Here it is.

My general assumption is that new information is more reliable than old information. That means, a page that has been changed yesterday has a higher chance of being relevant than a page last changed five years ago. The idea isn’t particularly new: It’s a bit similar to what Eclipse Mylyn does for Java developers.

A wiki is a graph consisting of pages connected via links. Using age information, you could highlight the subgraph that is most active in respect to changes. Links that lead to newer pages are displayed in a lighter color while links to older pages are displayed in darker colors. Page titles could be color-coded in the same way. Users always know if they’re in the active center of a wiki or in a rarely updated, possibly outdated area. You can always push pages into the center by making a trivial edit.

Taking the idea one step further, with a little bit more effort a rating system could be implemented. The system could take implicit parameters like age information or page impressions and explicit parameters like user ratings into account. For each page, a score could be calculated that is used for color-coding of links and page titles.

I read a lot of RSS feeds to keep myself up to date. Perhaps too many, almost two hundred articles appear in my Google Reader each day. One of the most busy ones is the DZone front page feed. Not all of it is interesting, so some filtering would be useful.

My first idea was to subscribe to the individual tag feeds (“java”, “opinion” etc.) but that leads to duplicates and false positives: An article may be tagged with multiple tags, so when subscribing to “opinion” for example, you’d also get opinions on .NET stuff. Not exactly what I had in mind.

The solution is simple: Download the feed, weed out articles I’m not interested in (ie. blacklist everything tagged with “.net” or “php”) and subscribe to the filtered feed. Pretty simple to do with a few lines of code, but you need your own server.

That’s where Yahoo! Pipes comes to the rescue. Yahoo! Pipes is an comfortable mashup tool loosely modeled after Unix pipes. It comes with a nice graphical online editor and is pretty easy to use. Within minutes, I created a simple pipe that reads the DZone front page feed and filters entries based on category. In Google Reader, I subscribed to my pipe’s RSS output instead of the original front page feed. That’s it.

Spam filters are supposed to take care of unsolicited E-Mail and they save us lots and lots of time each day. But what about all that stuff from inside your organization that is only loosely relevant to your work or not at all? Crowded inboxes use up time and energy and turn reading your E-Mail into a tedious task.

A while ago I stumbled across an interesting Google Tech Talk video: Inbox Zero by Merlin Mann. Inbox Zero is a simple concept about how to process your incoming mail without getting overwhelmed. It’s a non-technical, simple process, similar in spirit to David Allen’s “Getting Things Done”.

The concept suggests five different ways to “process” an incoming message:

1. Delete (aka move to archive)
2. Delegate
3. Respond
4. Defer
5. Do

The talk itself takes about 30 minutes with another 30 minutes of Q&A. I also like the presentation technique used. The slides are clean and minimal and perfectly suited for the message they are supposed to convey.

Merlin Mann’s 43folders website contains a series of articles around the Inbox Zero concept that’s also worth checking out.