Caught a reference to this at inessential (via this).

At first blush it seemed like a really good idea, especially when I consider not just my blogs but also Furl, del.icio.us, and all of the other associated dren I leave trailing along behind me. But the commentary regarding the potential usage of OPML for the same purpose is pretty compelling.

The major argument posed by the NeutrinoXML developers is that OPML isn’t specifically person-oriented, isn’t guaranteed to contain person-feed data. I would argue, however, that that’s irrelevant. If a provider of a person-OPML file asserts that it contains only the feed-data surrounding a person then that should be good enough for a potential subscriber and the benefits of sticking with existing formats are obvious.

Seems to me that what’s missing from this model (in so far as it could be easy for non-geeks) is an OPML element that could instruct an aggregator that this OPML is inteded for subscription, so that a simple click would subscribe you to the content of the outline.