Chris Dzombak

New IDs resulted in many “new” posts in Ann Arbor FOIA RSS feed

Subscribers to my RSS feed (or JSON feed) of Ann Arbor FOIA requests will have noticed that yesterday their feed readers showed a large number of “new” posts, which were not actually new and dated from weeks ago.

RSS feeds require each entry to have a unique ID, which is how your feed reader tracks which entries you’ve marked as “read.” The code that generates the RSS feed uses a FOIA request’s ID as its entry ID when generating the RSS feed. (This allows the feed generator to be essentially a pure function; the server runs it periodically to generate the feed file but otherwise needs to maintain no state.)

Per the City’s FOIA Requests page, on March 16th “the FOIA numbering system changed.” There’s no explanation of why the system changed, but all previous FOIA requests were assigned new IDs. Therefore, as far as your feed reader was concerned, all previous requests were completely new (and therefore unread) entries.

This shouldn’t happen again, until and unless the City reassigns new IDs to all its old FOIA requests again.