At 03:43 29/11/2001, John Donagher wrote: >On Wed, 28 Nov 2001, Robinson, Mike wrote: > > > BC may not be of grave concern to Python developers. > > Perhaps thats why it hasn't hit the radar screens. > > You don't get to 6.5 million domains and a million ip > > addresses chomping big 'BC' pieces off of developers > > butts. > > > >Two completely different user domains; you can't measure Python usage by >mod_python installs. Just like you can't measure Perl usage by mod_perl >installs.
How do you measure Python's success? I think that many people will argue that PHP's space is much more difficult than that of Python's, and would be much less open to such changes. In my opinion, it's quite clear that breaking compatibility is a bad idea, and breaking more things is always worse than breaking less things (hence the 'hey, we're breaking compatibility in a couple of things, so it means we can break many more things' approach is bogus). Every single thing that we break means one more thing for people to watch for when they upgrade, and more broken things means more time for migration, or, in reality, it means that people won't migrate at all. Typically, you start a revolution if things go really bad, beyond your ability to handle them. We came the closest, IMHO, with the register_globals issue - and there too, we're deprecating this feature very gradually, and still give people the option of saying 'I don't care, I want it to stay on'. This is clearly not the case with PHP, which is growing and growing through its regular evolution. It may be the case with your particular project, and some other projects, but it still doesn't mean that we can afford to break PHP in order to make it more hospitable to such projects. Zeev -- PHP Development Mailing List <http://www.php.net/> To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
