Hi EmmanuelMash, here's my 2 pennies worth - there is no scientific formula to this I'm afraid but most times, it is ends up being a case of "if it aint broke, don't fix it".
patches are usually applied to correct issues with certain functionality. Ideally, it would be best to be using the latest versions but that again comes with its risks as new software tends to throw in a few new bugs here and there and only the adventurous tend to be on the latest release straightaway. The pragmatic approach is to let the early adopters/trendsetters help Oracle iron out all the teething problems and then you can move to the latest versions when these versions have stabilised.
There is an argument that adopting the "if it aint broke, don't fix it" approach could mean people could still be stuck on 10g because it's "stable" whilst those that tend to grab hold of the new features stand the risk of exposing themselves to bugs and instability in newer versions. There's also the small matter of extended support cost for older version of software to take into account.
I guess the best approach would be to be in the middle ground and not adopt latest versions immediately and also not languish behind with older versions. Patches should be applied for security purposes as well as bug fixes which affect your implementation.
Business is more focused on stability (if its not broken don't touch it mindset)You mean the same business guys who have no issue with 15 apps updating each week on their iPhone and demand each new feature to automagically appear and compatibility with new browsers (compared to yeas old tech) to just work?Sarcasm aside - the closer to current the better.Why do you thing the common model for anything cloud is "You have no real choice but to be on the latest release"?
Joel Acha wrote:
Patches should be applied for security purposes as well as bug fixes which affect your implementation.
And "compatibility be damned"? Modern ecosystems are heterogeneous and no system must ever be regarded just on its own.
Well Christian a single Iphone's downtime wont affect decision making. Apps on the that have smaller user bases are much easier to control. We have created a monster but creating a dependable platform. Now everyone cannot do without it.
Hence my "sarcasm" bit. I fully get those points but as I said you can never discount the surroundings and security, compatibility, integration and feature creep all force you to stay on top of things as least in a close-ish manner.
Plus: The longer you wait the harsher the upgrade pain. Cumulated incremental steps efforts are often less than one huge bang effort.