I like the concept, and it's relevant to my line of work so I'll probably be keeping a regular eye on it. But the layout is pretty awful, the two-uneven-columns approach really doesn't do it for me. It's hard to read, and particular it's hard to read one article summary then just move on to the next. Could it not just use the normal slashdot style single list ?

I would suggest that the "bouncing from router to router" is most likely the cause of your performance issues. TBH where feasible an application server should always be on the same local network as it's database server, particularly when the application server is as database intensive as sharepoint. It's also not uncommon for there to be a dedicated gigabit (or higher) link. As for the storage model, I'd agree SQL Server (or any other RDBMS I know of) isn't really optimum, it's great for list based data, site config etc, but I suspect it's not the ideal solution for document storage (I don't have hard facts to back that up but I don't think too many will disagree). But at the same time I think it is a step on a standard file system when it comes to the kind of dynamic structures, metadata storage, search/indexing etc. I suspect though that when (if?) WinFS is released Sharepoint will move to some combination of a SQL Server DB with a WinFS file store.

Wow, even for /. there's a lot of people who didn't even read the summary, let alone TFA. And there's a lot of FUD being spread. What this means is that IP address information might be considered personal data under EU data protection laws. This means that companies/corporations/organisations which log your IP address will have to have a privacy policy in place governing how that information is used. There are also certain requirements, such as they have to make people's own information available to them if requested, they have to disclose breaches of information to those affected and so on. It doesn't stop logging IP addresses, it won't stop webservers using client IPs to maintain statefull connections, it won't stop google associating IP addresses with search data, it won't stop wikipedia or forums storing the IP of posters. It just means that organisations doing this need need a privacy policy in place to protect this data (which most of them already have to protect other private data they store). It's just acknowledging that IP addresses can/may be used, in some cases (the summary points out that they already acknowledge IP addresses are often dynamic), to identify a person and deserves the same level of protection that things like phone numbers and home addresses already have.