I emailed Pownce’s support team on July 10, notifying them that their Atom feeds contained malformed XML (specifically, unescaped ampersands in attribute values). This prevents many feed clients, including Tumblr’s feed importer, from parsing them. I described how to fix it and offered my help in doing so.
It’s still not fixed, and people keep emailing Tumblr support (…me) because their Pownce feeds aren’t working.
I finally got a response today, 20 days later, from Ariel: “We’ll look into it!”
Come on, developers… XML should always and only be generated by proper DOM creation and output methods. With them, escaping is never even an issue because you don’t need to do it: you just call node->setAttribute() or createTextNode() with raw strings.
It’s like using parameterized SQL queries: so many bugs are avoided that it’s stupid not to use them.