On The Matter of Constructing RSS Feeds

I believe this is working now, for real.
The situation is that RSS readers vary significantly in how they mark what is new and what isn’t. It turns out that the GUID, supposedly a “globally unique identifier”, when constructed as a url, can confuse newsreaders – even good ones like NNW. It is globally unique, and it is valid, but it’s not enough info to go on, so the reader errs on the side of not annoying a user with a “new item” mark when it might just be an edit to an existing story.
So, to fix this, I added a random number as an anchor to the guid url, like this:
http://department.college.edu/news-events/news/index.html#466
Because it’s random, it will update on every publish, and NNW likes this. It still validates. When you click the link it takes you to the top of the page since the anchor doesn’t exist.
Nice brain puzzle. Involved NNW forums, the RSS spec, much Googling on RSS in practice, figuring out random numbers in Perl, and making sure this worked for both single-page and normal publishes.
Please test! ;)
Signed,
Your Consultant