Jason Kolb is trying to reinvent the Internet.

Alex Barnett linked to Jason Kolb’s recent blog miniseries about online presence and identity.  

Here’s links to all of Jason’s posts (descriptions from Jason’s “Featured Posts” listing)

Reinventing the Internet, part one – How the evolution of social networks is going to fundamentally change the Internet and the way we use it to communicate.
Reinventing the Internet, part two – A domain name in every pot – Why and how our online identities will eventually revolve around our own personal domain names. 
Reinventing the Internet, part three – Unlocking the potential of the URI – this is really WHY everyone needs to have their own domain name.
Reinventing the Internet, part four – Connecting the dots – A look at the open peer-to-peer social network at various levels, and an overview of how it’s all hooked together.
Reinventing the Internet, part five – Decentralized network, centralized identity – Why and how our online identities should be nodes in a decentralized social network.

If you think you might someday want to have a part in the evolution of the internet, Jason’s posts are a great read.  He has some very interesting ideas that relate to how our personal data is stored and located on the internet.

After reading through these posts, I have a few thoughts to contribute.

1. DNS is not a solved problem.  Jason seems to think that since DNS has served him reliable for over a decade, it is sufficient.  There are many problems with DNS, and they mainly come down to trust.  The distributed nature of DNS makes it powerful and reliable, but it also makes it susceptible to many different attacks, including spoofing & cache poisoning.  Now this doesn’t really matter much for a lot of information, but what if we were relying on the security of DNS to verify the authenticity of stock tips coming from Warren Buffet?  As the payoff for fraud gets higher, we need to increase the security of the underlying systems.  The good news is that this problem has mostly been solved from a technology standpoint, check out http://www.dnssec.org/ for links to lots of resources, and this PDF specifically for a great overview on threats and mitigation details.  An alternative to this is requiring the personal server to have an authentication certificate from a reputable authority, and then relying on that to bootstrap any authentication.
   
2. Jason seems to focus on individuals, but this model could be applied to business entities as well.  Businesses have for the most part missed the social networking boat.  Yeah there are some entities that have set up shop on MySpace, or who publish company-focused blogs, but the value proposition hasn’t really taken off.  Jason’s model for publishing and consumption of information should apply to businesses as well, and it might be easier for early versions of it to gain traction in this space.
   
3. Something that will probably be critical in both the personal and business space is the idea of Views, or adapters that will convert the format and protocol of the data.  This way, I post some new family photos to my private data store, they get emailed to my email savvy relatives, they show up in a rss feed for those using newsreaders, they get published to a picture site for those who only want to occasionally browse my pics,  and possibly get sent off to kodak.com for printing and delivery to the grandparents. That way adoption isn’t held up because the folks on the receiving end aren’t living in the land of XMPP yet.  Of course with this last bit this beast would start overlapping with products like BizTalk.
   
4. Many of these issues have been solved in very complicated ways in the past by CORBA, and more recently HLA implementations.  These are both distributed models that allow publishing and subscribing of information, based on some predetermined schema, although HLA calls the schema the Object Model Template, and CORBA uses its Interface Definition Language.  What Jason is proposing seems much simpler at first, but the lessons learned from HLA and CORBA, especially in terms of schema development probably apply.
   
5. Many folks see this idea as being at odds with the MySpace crowd, but really it just requires that the main players allow you to use your own domain name on their servers.  In reality, not too many people want to run a server in their basement (Unless it’s dirt-simple and provides real perceived value).  It’d be great if this framework was open enough that I could own my personal domain name and get access to all of the tools of Myspace, Youtube, Flickr, etc.  Ideally, they would just be service providers (format conversion, friendly interface, etc.) and the the data would be pushed back and stored on my personal server (which is hosted by yet another company).  Microsoft and Google are already getting into this space with Live Domains, Office Live, and Google Apps.
   
6. Social networking sites provide a hub for communities to form around.  The “social momentum” that these sites have is going to make implementing the distributed model more difficult.  Right now soandso.myspace.com equates to “cool” and soandso.com means you are a geek.  Unless the “cool factor” of the distributed model can be raised above MySpace, then there’s no chance it’ll get any traction.  All of us geeks see this personal server idea and think it’s a utopia because it plays to things we think are important, data ownership & verifiable authentication without sharing personal information.  Will it really matter to the teens who sign up on MySpace because it’s “cool”?

That’s it for now.  Cool ideas Jason, it’ll be interesting to see where it goes from here.