Blog aggregators based on the planet software are a staple of open source projects these days. Nonetheless, there are little gotcha’s here and there.
Thom May and I have run the PlanetApache blog aggregator for Apache Software Foundation committers for some time now. It looks like that planet is finally going to move onto official ASF infrastructure where it belongs (in my opinion). If you want a preview, point your aggregator at planet.apache.org/committers. This planet is using Sam Ruby’s refactored planet code, which is known as Venus.
Over in Python land, there are two different and non-overlapping planets: http://planet.python.org/, and http://www.planetpython.org/, as a reader of both of these planets, I would love to see them consolidated so I could stop seeing the articles that are on both planets.
In any case, planets are a great way to get a sense of what is happening in the various communities. If your project doesn’t have one, you should think about starting one. If you have one, make sure it is working super smoothly.
The solution is that planets should stop accepting non-Atom feeds. The whole point of requiring a proper guid in Atom (as in actually globally unique, not kinda-sorta that you see in RSS) is so that aggregators wouldn’t change it — otherwise the aggregator’s broken — and then feed readers would know you’ve already seen post ABC12345 and not show it to you again.
Of course, as you’ve no doubt experienced, there are any number of flaws with that. People still using RSS are not shunned in polite society. We should definitely take a harder line there. Feed readers are not doing proper dupe detection based on guid, possibly due to the first problem and the fact that so many feeds have a broken understanding of the words “globally unique”. The technology already exists to solve this problem. Encouraging mergers of quite validly separate sources for aggregation isn’t one of them, though. That’s just symptom patching, not problem solving.
I don’t think that having two Python planets is a good example for this. I’m fine with GUID based duplicate elimination. But from the project end, why are there two aggregators for Python?
Quality / acceptance, I would suspect, being on the “inside” of neither location. The list of people being aggregated at python.org is a subset of those at the “unofficial” planet python and is different again from, say, pythonlovers.com. This might just be a case of the Internet routing around perceived failure: if the python.org aggregator isn’t accepting every feed that some other group want (possibly for good reason, since there’s a bunch of arguable junk posted to unofficialplanetpython), then the other group start up their own.
It shouldn’t affect me or you because our feed readers will only show us each post once. That it does affect us shows that there are still bugs in the system.