via Twitter, I was asked the above question.
It’s a good question, cutting to the core of my ambivalence over the religious wars between RSS, Atom, etc.
The flavor of XML a feed is published in shouldn’t matter.
Neither to the publisher nor the receiver.
Any parser able to handle multiple flavors should be able to parse all flavors equally fast. Some parsing engines are built for one flavor of XML or another – rather than abstracted to parse XML in general. Then again, it’s trivial to spit out one XML format as another, so, maybe format is a conversation between the user agent and the server.
Eh. (Get a smarter parser, jeeez.)
From my studying of both RSS and Atom, comparing them is like comparing the UIs of Windows and Macintosh. They do feel different. One puts window buttons over here, one puts them over there. One is this color, one is that color. One prefers the Control key, the other prefers the Command key. Some people prefer this one, others prefer that one. One says ‘potahtoe’. One, ‘potaytoe’.
From my I understanding, Atom was developed due to perceived deficiencies and ambiguity in the RSS 2.0 spec. Perhaps RSS 2.0 is guilty of being open to interpretation. I don’t know. I’ve found it to have logical places for everything I want to publish. Same for Atom.
Last I checked, Cullect was parsing somewhere north of 8100 feeds. Cullect doesn’t and shouldn’t care if a feed is Atom or RSS or RDF or filled with crazy namespaces. Cullect has 2 jobs when it comes to feeds; parse XML tags in a smart way, publish out useful feeds in whatever flavor the user agent requests.
The biggest issue I’ve found in parsing thousands of XML feeds is badly published XML. Feeds using the tags in bizarre ways. Feeds just not conforming to any spec. Feeds published in a way that just makes parsing hard.
Both RSS and Atom publishers are equally guilty. My Wacky-Feeds-That-Won’t-Parse list of contains just as many RSS feeds as Atom feeds.
A year ago, I wrote up my thoughts on publishing RSS 2.0 for easy publishing.