I’ve been cleaning up the code the last few days. I finally got around to switching to tweet entities. They’ve been around for a while, but every time people complained about a problem Twitter HQ said that developers should code for their occasional disappearance. I don’t have time to code for a major data component being there part of the time. The complaints have died down, so I guess it is stable now. I don’t disagree with the Twitter model of getting improvements out early, but I stay away from the bleeding edge.
One of the nice things about entities is that they include the disambiguated versions of shortened URLs. That saves a ton of processing time for each user of the API, and even more bandwidth for the target of popular URLs. Unfortunately, they only contain the original value of URLs that are shortened by Twitter itself through its t.co domain.
I’m also making sure that every piece of text is pulled out into config files. I want this code to work for any language.