"Suddenly stopped working"

[MikeSchinkel]

Continuing the discussion from The PHP 7 Revolution: Return Types and Removed Artifacts:

@tony_marston404

Confidence in the language is adversely affected when applications which have run happily for years suddenly stop working.

Tony, I am somewhat sympathetic to your argument. OTOH, I would challenge that statement. The app does not “suddenly stop working”, it stops working at the point the underlying platform is upgraded. So an enterprise would not one day walk in to work and find a broken app unless their IT management for platform upgrade process is haphazard.

Yes, they can’t upgrade to PHP7 without code changes, but that is not the same as “suddenly stop working” and I believe weakens your argument. Note I am on the fence on this albeit leaning slightly toward removal; IOW I don’t really care but if I had my druthers I’d see it removed.

BTW, would it not be possible to recognize a PHP4_CONSTRUCTORS constant and if set to true then behave the old way? Would that not provide the best of both worlds? (albeit I’m someone else has proposed this before and it was shot down by numerous people as language crud, or some such.)

[oddz]

Realistically upgrading from 5 to 7 is going to be a major undertaking for any organization. For that reason I would rather PHP 7 enforce using proper methodologies rather than allowing that goal to be obscured by allowing usage of old, antiquated code. As MikeSchinkel pointed out the process of moving from 5 to 7 isn’t going to be just flicking a switch. It is going to be a process that will in many cases force organizations to rethink the architecture of their old software. Which in many ways is a good thing all around.