Ben's Stream

Create a multi-column Twitter client with 12 lines of HTML

A while ago I switched from Tweetdeck to TwitHive for all my desktop Twitter usage. It has a lot of advantages, like being completely web-based so you can take your lists/groups with you, and you don't have to install anything to use it. However, recently I've noticed more and more that TwitHive isn't picking up all @replies, which is sort of a deal breaker. This is apparently due to a bug in the Twitter search API, but knowing that doesn't really help me.

I've been using Dabr as my mobile Twitter client for a long time, and it's awesome. Rock solid, (mildly) customisable and constantly being improved. I often wish that Dabr would put together a multi-column web-based client aimed at desktop users. In the mean time, I've decided to hack (and really, I mean hack) something together. So, here it is in all its glory:

<html>
<head>
<title>Twitter</title>
<meta http-equiv="refresh" content="300">
</head>
<body>
         
         
         
         

</body>
</html>

The design and implementation took literally as long as typing the text... not a lot of thought or polish here folks! The main trick is to log into Dabr in one column then refresh the page. Obviously you'll need to change the hard-coded 'column' contents to be whatever you want. And you could probably do a lot to make it look nicer, if you could be bothered.

4 Dec 2009

The Best System

Chase Jarvis has recently driven home the old addage that the best camera is the one you have with you. For a long time I've been saying something similar about exercise: the best routine is the one you do regularly.

The underlying message is that its better to execute something imperfectly than to not execute at all. Or to state it more simply: delivery matters.

It's a constant frustration to me how many people don't give enough weight to delivery. They feel that it's just one of the many requirements of a project, along with proper architecture, project management, reporting, testing and documentation. While all of those things are important, and often aid the production process, they don't make a damn bit of difference if you don't deliver a system that meets the core business need.

For a lot of 'successful' projects these important, though secondary, aspects are missing. And because we agree that they are important, there is a tendancy to over-inflate their importance in an attempt to ensure that they are done in the future. While this is admirable, it's unrealistic. What it doesn't consider is how many projects had all these things in place, but never saw the light of day.

Often you need to take short term productivity hits to ensure long term viability. Focusing on delivery doesn't mean that you can't, or shouldn't, improve processes. But you have to do it whilst still delivering, even if it's at a slower rate.

A perfectly designed, archtected, managed and tested system is worse than useless to your clients if it doesn't help them meet their business needs.

So please keep striving for improvement, but remember: the best system is the one you actually deliver.