Christopher Jones has posted an announcement about the latest release of the PECL OCI8 package (version 1.3.2 Beta) hitting the streets:

I've released PECL OCI8 1.3.2 Beta - the latest release of PHP's OCI8 extension with support for Connection Pooling and Fast Application Notification. The release is based on the current PHP 5.3 development branch.

He notes another change in this release - a "session release" bit of functionality persistent connections will do when nothing is referencing them anymore, mking them work a bit more like normal connections. Issues that could be caused by this can be corrected with a new setting (oci8.old_oci_close_semantics) in your php.ini.

Nick Halstead has pointed out a website, written in PHP he's created to help make a little bit of sense out of the links that go flying past in your twitter client - tweetmeme.

What is it? It tracks the public timeline from twitter and picks up any links that get posted. It then follows each link to find final destination and then categorizes the content into blogs / video / images / audio. This project really shows what is possible using PHP if you know what you are doing.

Most of the work was done by another developer, Stuart Dallas as one of four that worked on the project together. It's written in PHP5 and uses only about 20 PHP files to get the job done. You can also check out the launch post over on tweetmeme's blog for more information on the service.

Jonathan Street has a few tips for developers out there looking to speed things up on their site - seven tips towards "lightning fast sites".

I was recently creating a small tool in PHP and found myself hitting the max execution time and getting a fatal error. As it was only for my personal use I just bumped up the max execution time but it made me stop and think about how I could improve the speed of those scripts I do put up for public use. Most people aren't going to wait for 60 seconds for a page to load. Naturally I hit the internet looking for tips.

The tips he came across (each including the simple benchmarks to show the differences) were:

• sizeof vs count
• is_int vs is_integer
• chop vs rtrim
• doubleval vs floatval
• fwrite vs fputs
• implode vs join
• ini_alter vs ini_set

Be sure to also check out his follow up post talking about building better benchmarks to test the sorts of tips he's given above.

David Coallier has posted about a database abstraction layer that he's been developing for PHP 5.2.x only systems and wants some opinions on his methods:

I made a very light DBAL that uses PHP5.2.x only (Since many people seem to want that) and it has the exact same DSN syntax as MDB2 for now and the query method are also called the same (No API Changes). [...] The main goal of the DBAL is to have a very effective and light way of switching RDBMS but also the possibility to change your DBAL to something more "0feature complete" as such as MDB2.

He includes the list of query method names and the types of databases that he wants it to support (as well as mentioning the fact that it would be unit tested for reliability).

The IBM developerWorks website has posted the first part of a series looking at boosting the performance and throughput of your PHP applications through an opcode caching software, specifically XCache.

PHP is a scripting language most often used to create Web applications. It's easy to learn and produces visible results quickly. However, because PHP is interpreted, PHP code is parsed and translated to opcodes every time it executes. An opcode cache eliminates that rework, making PHP applications faster.

They start with the installation of the software (just XCache, they assume everything else is installed) and what to edit in the php.ini file to get things up and running. They follow this up with a sample benchmark for a local phpmyadmin installation.

The IBM developerWorks site brings the community another great tutorial today, this time looking at the CakePHP framework and the creation of a simple application with it.

In "Cook up Web sites fast with CakePHP, Part 1: Adding related information and services", they build a sample application (Tor) from the ground up (installation guide and all) that will take in a username and password on a login page. Their goal is to show you how much time you could safe by using the framework over just the usual library-based (or procedural) programming methods.

You will need to be a bit familiar with PHP and the Model/View/Controller design pattern before you get started here, but if you have that down, you can just jump right in. You'll also either need to register with the site or log in with your account information to get to the good stuff.

Paul Jones has run some benchmarking on four different frameworks - CakePHP, Solar, Symfony, and the Zend Framework - to check for overall speed and has come up with some interesting results.

Anger-inducing broad-brush overview: Solar is 4x faster than Zend, and almost 2x faster than Symfony. Read on for all the nuances and caveats.

Throughout the rest of the post he talks about the methods that he took for testing the frameworks and that his goal was the fastest response time only, nothing about the actual functionality. The test was as simple of a "hello world" that he could get in each one.

He's provided the benchmarking results, as provided by Apache's ab testing script for the base (just PHP manually) and then for each of the frameworks, providing the code used as well. There's also a helpful bit on which classes get loaded for each of the frameworks to see what the overhead is. Symfony tops off this list with 43 total libraries for what ends up as a simple echo statement.

Continuing in his series of benchmarking various statistics about PHP, Sebastian Bergmann has created a few more interesting bits of information. This time, gives the results of running bench.php on the numbers he ended up with (and made some spiffy graphs out of them as well).

According to his numbers, a real speed boost was made in compile time when PHP 5.1.6 came on the scene. PHP 5.2 changed things very slightly, slowing things down just a bit (on both the GCC 3.4.6 and 4.1.1). The ICC compiles saw a similar jump in speed with 5.1.6.

So which is faster overall? Well, as the last graph shows, even starting from PHP 4.4.4, the ICC with the "fast" option beats either version of the GCC compiler by a good margin.

Oh, and if you'd like to run some stats yourself, he includes the raw data he found when making the compiles.