At my job, I have to write parsers for data that comes from online information repositories. Each repository provides pretty much the same information, but each has its own quirky format. Until recently, I was writing parsers based on the domain name of the repository. The problem here is that sometimes a repository will start using a new format. Once this is detected (and it might take a while for us to realize that) I have to examine this new format and write a new parser.

Then, this morning, I noticed two repositories in a row in my queue that had the same IP address. D'oh! It turns out I've written a lot of duplicate parsers. Things that appeared different were not distinct, but only variable (e.g., an address that requires two lines instead of one). These differences between repositories made me write seperate parsers, but this IP realization has shed light on the issue.

As of tomorrow (when I should have the redundant parsers done away with and the core parsers centralized) there will be far less code to maintain, and it will be easier to deal with a change of formatting. See, if repositories A, B, and C are all the same IP address, then if that IP address's format changes, I only need to update its format. If just A changes its IP address, I can take care of A's format without worrying about B's and C's. Even better, if these repositories new IP address happens to already be covered, I don't need to anything.

Laziness, Impatience, Hubris. Amen.