Last weekend, Jennifer and I both joined Twitter. For my part, a large use case for it is sharing links and random thoughts that aren’t substantial enough to warrant a full blog post. Previously, these would likely be shared via my Jabber status, instant messgaes, and/or e-mail. Twitter will provide a platform where I can share interesting things I come across in a broadcast fashion.

What is Twitter?

Twitter is a micro-blogging platform — a service for posting short messages (no longer than 140 characters). These messages (called “tweets”) comprise your stream. Other Twitter users can follow your stream and see your messages in their timeline (the stream of messages from everyone you follow).

How do I see the messages?

Now that I have completed some development work on the web site, the main page of my blog now has a Latest Tweets bar that shows my most recent Twitter postings. Additionally, there are now two versions of the RSS feed for my blog: one of them is the standard feed that just has blog posts, and the other includes Twitter messages interspersed with the blog posts. So, if you are not on Twitter but want to see these messages, either of these methods will deliver the content to you.

Alternatively, you can subscribe to the RSS feed of my Twitter messages directly from my Twitter profile to have a feed which only contains tweets. And finally, if you’re on Twitter, the best way to get them is to follow me.

Miscellaneous details

And now for some boring minutae which you may nonetheless find interesting.

Twitter post duration on blog

The logic for determining how long a Twitter post will stay visible on my blog is relatively simple. The most recent four posts are visible. Additionally, any posts before those which are marked as “sticky” (I put a “/s” at the end) which no more than 36 hours old will be displayed.

The RSS feed tracks a few more messages, and sticky posts are visible for 72 hours (unless I somehow post way too much and break it).

Update frequency

I re-pull Twitter data every 5 minutes, using a local cache the rest of the time.

Implementation

The Twitter client code is implemented using Drakma, CL-JSON, and some other helper libraries.

Future features

Currently, I display the tweet text unprocessed other than turning certain things in the links. Twitter, however, uses URL shorening services such as TinyURL to shorten URLs and allow more text in the 140 character message limit. I would like to implement support for resolving such redirects before displaying them.

I also have ideas in mind for automatically tweeting when a new blog post is published (and filtering these messages out in the elehack.net display).