Today, Microsoft unveiled their RSS strategy for Longhorn. After reading their documents online, here are my initial thoughts.
First, pushing the aggregator functionality into the OS itself is very interesting. Think about turning the aggregator engine underneath FeedDemon into a system-level API that many other applications can now use. Will FeedDemon ditch their own aggregator in favor of Microsoft's? Of course, Microsoft will use this across all of their own apps, and I wonder whether they'll ship a news aggregator app or integrate news aggregation inside Outlook. The danger here is that if MS owns the general aggregator, will they innovate and be receptive to innovation elsewhere? For instance, will they consider adding the XMPP Pubsub protocol?
Second, the notion that a user has one master aggregation list doesn't seem right to me. I want to subscribe to news feeds in my news aggregator, and subscribe to songcasts in iTunes... will the Longhorn aggregator allow apps to group collections of feeds arbitrarily, or will it always be based on media types?
Third, I think shoe-horning lists into RSS is OK, but in some ways, we're just talking variant flavors of XML, no? The danger here is that MS can offer all the CC-licensed specs they want, but if they keep their aggregator code proprietary, they can effectively control which RSS extensions get traction and which die... unless they detail how their engine is extension-agnostic.
Fourth, I think the multiple media approach is right on. I actually wanted this in email years go, but realize now that spam made that impossible. By integrating the enclosure support with their security model, MS is definitely thinking ahead here.
And I wouldn't be surprised to see Apple add a similar OS-level aggregation service in Leopard. This is the IE/Safari scenario all over again... Microsoft turned IE into a widget, so that third party apps could build web-enabled apps. It took Apple years to adds a browser widget to the OS, and when they finally did, saw an explosion of third-party web-based apps of their own. I'm sure they won't make that mistake again.
One question: how soon before the open-source community builds a general aggregator engine?
Recent Comments