The Yahoo Problem – a draft
I the pdf below I have outlined two examples of the flow for a web application in the lines of “The solution to the Yahoo problem”? post by Dave Winer (I’ve called it subscription memory service).
I have made a couple of additions to the initial Winer’s proposal:
- A check is added to the registration page asking the user if he/she wants to make the subscriptions OPML publicly accessible. If it’s not then anyone accessing the opml has to provide both the userid and the password.
- In the confirmation page I have added the ability to specify one or more category / keywords / tags for the feed being subscribed (which should be optional).
This way the user’s subscriptions could be retrieved filtering for a specific category and the data base of tagged subscriptions would be a good start to identify the relevance of a feed about a specific topic (it would be a kind of peer reviewed topic system).
What I see as important as the reference software implementation is that the entry points (URLs, parameters) for the site be clearly defined in the specification for the service.
btw what this give more than the feed:// URI scheme is that it is not linked to a user’s pc setup but could work for web based aggregators, i could even subscribe to a feed while away from my pc (think browsing in an internet cafè) and have my aggregator get the subscription when i go back.