Today I went to an interesting presentation on Smart Objects (in data repositories). It was interesting seeing the data/metadata from a research/library-oriented perspective. I should follow-up with writeups and citations, but I need a better way to organize all of this data.

(model w/ Drupal’s centralized taxonomy?)

This guy built a uber-cheap 1.2TB RAID system for $1,600. I spent $2,500 for my 0.75TB system back in December, but I’ve been very happy with it. Extra money was spent on getting a decent UPS, a real hardware RAID controller, and a hot-swap case w/ a sizable power supply.

Looking back, I would have spent an extra $100 for an even better case, gotten the 3Ware 8506 card instead of the LSI MegaRAID 150, and gotten a motherboard w/ 64-bit PCI (bought one on-sale that didn’t have that, oops).

Anyway, I figure I’ll be good for at least another year or two, at which point I’ll probably swap out my array w/ 1-2TB drives.

notlong is a new short url service (TinyURL, EvilURL that packs in some clever features. Short URLS are generated as subdomains and also generating a password to both edit the long URL and view some click-through stats. The author also hosts an overview of various url forwarders .

Speaking of cleverness, Dan Kaminsky‘s talk at LayerOne this weekend was very much so. It was mostly about very ways to (ab)use DNS, using distributed recursive lookups as a communications channel, for file transfers, radio broadcasts. Also, a 3D-plotting tool for visualization, and communicating via third party DNS servers via last-lookup bit leakage.