Thanks for posting this Jakob. I was just reading RFC 5005 on the
train yesterday (literally) and the parallels between it and OAI-PMH
struck me as well. It's not quite clear to me how deleted records
would be handled with an atom archive feed. But I guess one could
assume if the identifier is no longer present it has been deleted it.
But that would require pulling the entire archive... I'm not really
sure how much deletes are really used in OAI-PMH repositories anyhow.
Stuart Weibel has written  about the subject of blog archiving in
the past. And I remember hearing Jon Udell and Dan Chudnov talk about
it . Who knows what technorati, bloglines and googlereader are
doing in this area. I guess the reality is that blogs are on the web
and as such will be archived by InternetArchive . But perhaps that
doesn't really fit quite right? That's my feeling.
I think your general point is correct. Libraries need to be
integrating themselves into the web these days rather than expecting
the web to integrate into them.
Oh, and would it be alright to add your blog to
http://planet.code4lib.org -- we need more of an international
presence on there IMHO.