RSS is one of the cylinders firing the juggernaut that is Web 2.0. For the uninitiated, RSS is the technology which allows you to take a feed of news from a given website and display it in another place, usually in an ‘aggregator’ which displays a number of feeds in the same place. RSS has always been intended for delivering pieces of information which belong in chronological order, like blog posts, twitters or the latest items on the Digg homepage. It does an exceptionally good job of that, ensuring that I always have the latest posts sitting in my Google Reader account, and that I don’t miss anything I don’t want to.
The problem is, not all data on the web that we might want to syndicate conforms to the template which RSS is designed for of chronological snippets of information. For example, RSS would be inadequate for delivering a music chart to me. In a music chart, it is not the chronology which determines the order, but the popularity of a given song. RSS is therefore not the right tool for the job, as when delivered via RSS the information would make little sense, with a series of songs appearing in my aggregator in a seemingly random order.
This wouldn’t matter if snippets of information in chronological order were the only things which it would be useful to syndicate. I think it would be quite nice to input a feed URL into Google Reader and be able to see, at any time, music charts. Or maybe the new Digg Top Ten list (look on the right sidebar), which isn’t based on the chronology of the stories but on the activity surrounding them. Maybe RSS isn’t the best format to syndicate conventional web search results in. I genuinely think it would be useful to be able to syndicate things like this, and this is currently not possible to any meaningful extent under the current situation where the only technologies around are RSS and Atom.
Before I continue, I think its probably a good idea to say that I am by no means an expert in syndication technologies (and neither am I a developer of any quality), and so if I’ve made a technical mistake feel free to set me straight. Nevertheless, I’m going to get ideas above my station, and tentatively suggest what the answer might be.
It would seem impractical to have a multitude of new syndication standards, with one for charts, one for web searches and to have to develop a new standard everytime someone thinks of a different type of information they would like to push around. Instead, a standard XML document could be used, with a common set of tags such as <title> and <description>, and the ability to invent new ones (as one can currently with the RSS namespace idea). What would be new is that instructions on how the information contained in the tags should be used should be contained within the XML document itself, obviously in a machine readable format. This would be far more flexible, and enable all sorts of uses that I haven’t thought of.
This is just an idea that I’m putting out there; perhaps its not possible, or maybe it isn’t necessary. Someone might have even done it already. It just seems to me like a logical next step if we want to make what is already an incredibly useful idea even more useful. Thoughts?







