Did some more testing of DTPS, my object-centric ORM, and continued to find it perform SELECT queries 3-4x faster than Propel. Additionally, I found it performed SELECT and UPDATE queries about 2x faster than Propel. INSERT queries, however, where a bit of a different story: DTPS took about 1.3x more time than Propel to complete the equivalent query.
While I'm impressed, and quite surprised, with the relative performance of DTPS, the thing that really gets me PUMPED about DTPS is how it innately supports polymorphic queries with no table duplication, and just how rapidly you can slam out a new persistent data type.
Overall I'm pretty happy with the system, and am curious how well it will work in practice.
For the past few weeks I've been experimenting with designing an object-centric MySQL database structure that allows php web applications to be written in a manner that feels like a typical single user heap program while preserving much of the efficiency and concurrency of traditional relational database design.
I've now got a nice/functional version completed and have found that its SELECT performance is 2-3x slower when compared to hand tuned prepared SQL; however, I've found that it's performance is 3-4x faster when compared to an equivalent query using the popular Propel ORM. This discovery was quite surprising since my design was focused on rapid prototyping/development and not optimal performance. Although, I'm happy with the results so far, my testing has not yet convinced me of the practical usability of my design.
FYI: I'm currently calling the system DTPS (Ductape Pirate Ship) due to the whimsical and rapid design that it encourages.
Got most all content migrated to the new web platform.
The website. It's BACK!
I was using Joomla! but found that it wasn't the most stable thing ever + I found it needlessly difficult to extend. The solution was obviously to design my own platform. It took a little more time than I wanted to take (~1 month), but I'm fairly pleased with the results and I was able to create and implement some rather novel/experimental ideas.