The Eighty-Twenty Solution

I had a bug report on NewsGator Outlook a while back concerning the fact that NewsGator doesn't accept or send cookies delivered with an RSS feed. It's just that to date, we've never had to, and off the top of my head, I can't think of a site that uses cookies on its RSS feed. LiveJournal used to use cookies for authentication, but IIRC, they're using Digest authentication these days.

So I'm having trouble figuring out a use case for cookies with RSS. About 18 months ago, Tim Bray had a post on the difficulty of counting subscribers to a feed, along with some pushback on the use of cookies for this purpose. Today, Brad Feld posted on Feedburner's new features, including basically what Tim was asking for. Of course, this places the reliance for counting on a central site, in the hands of a 3rd party, no less. Still, there are certainly other ways of counting subscribers. Another argument would be for saving user preferences, but I'm not sure of what concrete preferences you'd be saving (or maybe that's none of my business). It's pretty well accepted that server side state is a burden on a web architecture, and I'm willing to accept that cookies are a valid way to push state management out to the client, but I just haven't seen it done with RSS. And the bottom line is that I don't know of any aggregators that would support this. Furthermore, feeds using cookies without additional authentication requirements would pose a big problem for centralized services like NewsGator Online and Bloglines; now any feed delivering cookies would either sacrifice personalization features or force the centralized aggregators to treat each instance of that feed separately.

Lots of people have made excellent posts based on The Long Tail, including Dare Obasanjo and Adam Smith. One quote drew my interest:

"What percentage of the top 10,000 titles in any online media store (Netflix, iTunes, Amazon, or any other) will rent or sell at least once a month?"...Most people guess 20 percent, and for good reason: We've been trained to think that way. The 80-20 rule, also known as Pareto's principle (after Vilfredo Pareto, an Italian economist who devised the concept in 1906), is all around us. Only 20 percent of major studio films will be hits. Same for TV shows, games, and mass-market books - 20 percent all. The odds are even worse for major-label CDs, where fewer than 10 percent are profitable, according to the Recording Industry Association of America.

Really, what you're seeing is a restatement of what Joel Spolsky mentioned in Bloatware and the 80/20 Myth; responding to the contention that 80% of the users use only 20% of a program's features: "Everybody uses a different set of features.".