Hi, Op woensdag 16 augustus 2006 22:45, schreef Dries Buytaert:
I sense that most people are ready for all this, but that at least a number of people are still catching up with the Drupal 4.7 madness. Just look at the translation status page (http://drupal.org/ translation-status), for example. Certainly, the translators are behind a little. Given more time, they would likely still be behind.
We have three things: themes, modules and translations. Translations are behind. Modules are far, far behind (most of the 4.6 to 4.7 modules are still suffering from instabilities and bugs) and themes are still not very stable (if 1 sept is the date, we had a cycle where most of that cycle we had less themes available, then in the old 4.6 version). My gut feeling is that 4.6 is there to stay. So many people stayed on 4.6 because critical modules for them (flexinode is one I know of from first hand experience) never became really stable on 4.7. This whole talk leads to Dries' interesting point: Even given more time, there still will be stuff never ready for 4.7, same will be for 4.8. Two things can happen: People skip complete releases (upgrade from 4.6 to 4.8) OR people stick on a release forever: * People on 4.6 could not upgrade to 4.7 because modules or themes were never ready, they lacked resources to do it themselves. * Developers never got around fixing their modules for 4.7. * There were less people around doing cool stuff for 4.8, because they spend ages getting their modules and themes and sites to 4.7. Same for 4.7->4.8 * New modules are introduced during that 4.7 cycle to *replace* broken/not upgraded 4.6 modules, but are not ready/released (at all), nor are they compatible w. eachother (cck <-> flexinode). Leaves Drupal one simple question: Should it bother? No! Upgrading is something -I estimate- that will happen a lot less, people invest too much in a project to follow up every x months and invest big amounts again. But the growth, the new sites, will build upon new versions. If Drupal were a snake, it would get a longer tail, so to say, not all people ride on the head anymore. The good news is that that means more people using, developing testing and investing in older versions: these versions simply stay alive longer. Bad news is: less hands for new versions/head. To illustrate less hands: I maintain some modules that I will probably never upgrade to 4.7, simply because they are happy on my 4.6 site(s) and because I never got proper patches from others for an upgrade to 4.7. Let alone That I or others get them ready for 4.8. Example: I have looked at the new improvements in HEAD, and found that upgrading flexinode (properly and clean) to 4.8 is going to be a hell. It will probably never happen (properly). In the same time, CCK is still not ready for prime time, and I doubt that it will get there in the 4.8 cycle. So people looking for these CCK/flexinode alike node-builder features are still best off in 4.6. And that will be the case for a long while in 4.8 too! But all the effort in getting flexinode somehow ready for 4.7 was effort not spent in getting CCK up/out there! So I do not mean "less hands to improve core for 4.9". I mean less hands to get stuff over and ready in 4.8! In short: off course we should stay improving with the current fast pace: Lets freeze asap. And lets not bother about all the backwards compatibility, nor about people not keeping up with HEAD.. Older stable releases are good as they are now. And they fullfill all that we need for that backwards compatibility. Bèr