Review: Harry Potter and the Order of the Phoenix

This book deals Harry a rather brutal comeuppance, which is much needed by this point in the series in order to flesh out his character. The previous book ended in a very dark way relative to what came before, but left Harry himself relatively unscathed. Here for the first time he makes serious mistakes with irreversible consequences. Through various circumstances he also loses (temporarily) a great deal of the privilege he has grown accustomed to. As usual the story is well crafted, although this volume is marred by what I expect is the most glaring plot hole in the entire series (the two-way mirror).

Review: Confessions

A bit harder to get through than I expected. This is my first direct exposure to the Church Fathers. The extended sections discussing philosophy were rather tedious but they were interleaved with passages of startling relevance and insight. On the whole I found it to be hugely valuable toward broadening my exposure to Christian history and thought.

It was entertaining to read Augustine's meditative interpretation of Genesis 1 soon after the recent media frenzy around Creationism. I'm not really sure what to make of patristic hermeneutics, but Augustine is no lightweight on the questions of Biblical authority and natural science.

Why RSS is the only good social network

I remarked a few days ago that the only good social network is RSS. It was kind of a mostly-false true-in-spirit kind of thought when I wrote it, but the more I think about it, I cannot think of a refutation. The thing I keep thinking is that for a social network to really be something, it has to be federated. All walled gardens eventually die. Something else comes along and people migrate away. Even when things are going well, there are multiple competing networks so there is fragmentation. You can choose to either interact with a subset of the people on the Internet, or you can deal with multiple networks and cross-posting.

A federated network, on the other hand, doesn't suffer from fragmentation, and has the potential to have much greater longevity. Commercial products can be built around a federated network, but the network itself does not depend on the viability of those products. Of course, a federated network is also immune from the woes of monetization. You will surely have hosted implementations that serve ads, but you always the option of using a paid service or self hosting, while still communicating with the same network.

For this reason, I find myself to have cooled a bit toward the viability of App.net. I still enjoy it far better than Twitter or Facebook, but making a walled garden that is paid rather than ad-driven seems a pretty minor improvement.

There are really only two successful, non-centralized social networks: RSS and email. Both will outlive Twitter. It is interesting that both are also basically free-form, worse-is-better type solutions. In a sense, they are two sides of the same coin. Email is the federated push protocol, and RSS is federated pull. I award the prize of "only good social network" to RSS mostly because it is simpler (as far as I can tell) and doesn't have the severe spam problems that email does. (While we have good filters now, I still see spam as an obvious mark against the protocol, even if there is no good alternative.) There is also the problem that email purports to be private communication, but doesn't offer very good security.

I am currently fascinated with the successful P2P protocols, particularly Bitcoin and BitTorrent. I recently ran across Twister, which is a fully P2P Twitter clone, built using code from those two projects. The author mentions censorship and privacy as his primary motivations, which is a whole other angle on moving the social web away from centralized services. Every message is signed and DMs are end-to-end encrypted. I suppose if we do ever end up with a successful P2P social network, it would pretty easily trump RSS as the only good social network.