Re: require_once vs. no require_once - please read, critical information

Hi, Gregory Beaver wrote:

I thought it might be useful to have a definitive benchmark on require_once, both of its performance AND maintainability, something often neglected when benchmarking.

My head hurts from reading messages of such length in English, so I think I'll pass further discussions on require_once issues unless you are willing to switch to Russian. After reading this diatribe I see the following reasoning: 1) We don't get noticeable performance improvements from switching require_once -> __autoload() 2) We don't get noticeable maintainability improvements from switching require_once -> __autoload() 3) We have a long track record of using require_once-based PEAR and therefore know a lot of problems with that 4) We don't know of any problems with __autoload() because PEAR2 is vapourware 5) If we think really hard we can come up with some corner cases that *may* benefit from switching to __autoload() ...therefore we should immediately switch from problem-ridden require_once to problem-free __autoload() and get huge performance and maintainability improvements. This reasoning is extremely convincing so I change my potential vote from -1 (WTF?) to 0 (don't really care). The main reason of course being 6) If we rip something out then we can easily add this something back when we have more solid evidence.