We have X amount of time available during a publish cycle per discipline (Design, Engineering, Art, QA). We decide we want to do features A, B, and C. All those features get scoped out for each discipline and you come up with a timeline to hit the target date of release. For some stuff - especially holiday stuff, you have hard release dates. When you run into issues that expand beyond the original scope of a specific discipline's original estimate you have to decide what to do - either expand the scope or move on and address it later. If you decide to expand the scope enough times it leads to what we call "feature creep" which can throw off the entire schedule and what often times leads to projects/features landing in "development hell." In the case of the Trophy, our art/design/engineering resources are heavily vested in doing new/revamped creatures & functionality for the pet revamp, so we start tweaking the schedule now and it throws off that project for the future. Do it enough times and like I said, development hell.
This often begs the question, "Why don't you just release this thing or that and release whatever else later?" There's a lot of inter-dependency in the code base and doing that many forks gets messy pretty quickly, and as you can imagine after all these years its challenging enough to deal with it.
Let me give an example from 9 years ago to what Kyronix is referring to.
While we in the Americas and Europe didn't get a boxed set for Kingdom Reborn, UO Japan had a box (soundtrack CD?) with a KR install disc as a "bonus" (thinly disguised way to get a "boxed product" on the shelves there). EA Japan was FIRMLY committed to having it shipped 2Q 2007 - or, on the shelves no later than the start of July. This created a Hard release date for the rest of the world, as it would have cost them thousands, if not tens of thousands, of dollars, to cancel or delay that Japanese release (not to mention the potential embarrassment for the non-delivery for the UO staff in Japan - some Japanese are still like that).
Between "Feature Creep", program bugs, and the like, KR wasn't really ready for an INTERNAL ALPHA when it was forced into OPEN BETA by that hard deadline. And it was released at a point when it should have been, at best, put into closed Beta. The whole KR debacle was caused by deadlines. Had that hard release date not been there, the KR client would have been a 10th anniversary release that September at the EARLIEST (more likely would have been a 1Q 2008 release), and been a LOT more functional than the turd sandwich we got.
And the "just release what you got" idea resulted in KR - and the INABILITY to get later stuff to mesh, because of the inter-dependency of code, is what forced them to have to effectively reprogram the KR client from scratch to create the Stygian Abyss Enhanced Client - which DOES work in many ways that simply patching KR did not.
and what about the advance notice that it is going to the Prodo Test Shards, 24 hours would have been nice to get some Chars off to other shards and bring home some other Chars too.
