Aquarionics

The RSS Problem

“All those icons” says Dave , “Where will it end?”

He then goes on to describe a system that is as overcomplicated as it is reliant on his own OPML spec. Jeremy thinks the answer lies in the browser, but that would rely on users having things installed. This is my solution:

• User clicks on first “Subscribe to this” link.
• User is forwarded to sendtomyaggregator.com or something
• User is asked which web-based aggregator they use (Or “I use my own” which will serve it as a text/xml+rss document)
• sendtomyaggregator.com sets a cookie with this information
• User is forwarded to the “Subscribe to this feed” page of their selection
• User clicks on subsequent “Subscribe to this” links
• User is forwarded to sendtomyaggregator.com or something
• User is automatically forwarded to the “Subscribe” thing they selected last time.

We store no information on the user – we just read the cookie, and maybe the front page has a link to delete the cookie – but that just means it takes nothing to serve it beyond a simple perl script.

Remember: It should be the simplest thing that could possibly work.

Update: So I coded it