[development] Go PHP 5, Go!

[Rob Thorne]

Karen Stevenson wrote:
> I think it's great that there's an effort to get many open source projects behind this. We really do need to get a migration to php5 going, however I do worry about an absolute php 5 requirement since there are so many users that have no real control over what is available to them. 
>
> Rather than say, 'Drupal 7 is php5 only', why not say 'Works best with php5', then have some really nice features and modules that only work with php5. I think it would be perfectly fine if some core modules and features are php5-only, so long as it is possible to create a basic working site without them.
>   
We're close to "Works Best With PHP5" even in Drupal 5.x, if we're being 
honest.

I suspect that most of us are doing most of our development on PHP5 
already.  I know that for me, even in PHP4 deployments, I'm effectively 
doing most of the work in PHP5, and coding around the PHP4 related issues.

I suspect that if you want to treat this as a marketing problem, it's 
worth presenting Drupal 6 as "Works best in... We encourage..." mode.  
We may even want to say that some PHP4 related issues, with the 
exception of security, will be at "lowered priority".  This is to signal 
to our user base that they should be thinking about migration, and to 
the hosting community so that they can start to prepare themselves for 
upgrading their offerings.

There really *are* serious benefits to building a large application like 
a CMS in PHP5 over PHP4.  There's a reason why people use languages like 
Java, Python and Ruby -- some things are just easier to code and to 
understand, and especially to maintain.  If you've done much work in any 
of these, you know what I'm talking about.  PHP5 has these features, and 
PHP4 does not.  But to take advantage of these benefits, we'll want to 
take a close look at the low-level architecture and make some serious 
changes.  So if we do it, we need to do it "whole hog".  "Half-hog" 
won't do it.

But if we won't do it, some other project will, and some of the loss of 
share will be our own.  We're much better off helping to lead the change 
than be caught trying to catch up with another project.  This is true 
generally, but if you're doing this for a living, there's a real risk of 
loss of business if the transition to PHP5 doesn't start before too long.


Rob