I noted some sites started pushing to PHP5 with the announcement last year that PHP4 would be EOL (end-of-life) in 2007. In truth I understand that there is no longer a compelling reason to remain with PHP4. The biggest obstacle was older software that did not support PHP5 (since version 5 is incompatible with version 4 in some respects). However there is no reason why most of that software cannot be updated, and if so I am pretty sure that some alternate version 5 compatible software exists. I also read some claims that in simpler configurations PHP4 could be faster/less memory than PHP5. I don’t know if that’s true or not. Finally the biggest problem: most web hosting providers are content with PHP4 meeting all their needs and have no reason to upgrade. I do agree to this in some ways.
My biggest problems with PHP in general are its very poor track record for security, its flaws in design/configuration and finally it’s tendency to break software on updates. In some ways: its just a nuisance to maintain!!! PHP4 will be supported for critical security issues till 2008-08-08 by the PHP developers. However being open source there is no reason why someone else could not support it after that. Redhat Enterprise Linux still supports a PHP 4.3.9 package. I was supporting a 4.3.11 package for Fedora which I updated last in 2007 for Fedora 7. However I only installed it on my personal web server on a development box. I guess it would be really easy to finally abandon supporting old packages and just move to version 5, however I don’t know the effect it may have on my public server with a dozen or so websites.
For now I will set a deadline for myself to migrate to PHP5 by the August deadline. However for the time being, if people would find it useful I am considering repackaging the RHEL PHP 4.3.9 for Fedora 8. My 4.3.11 package is greatly out of date. On my development server, I’ll just go ahead and install the PHP5.2 included in Fedora 8. That will be my testing ground for my server updates coming soon.
Realistically in the long run I should just slowly stop using PHP altogether, given that PHP6 will be another mess very soon. Perhaps I will look into Python or J2EE options, not sure yet.