On Wed, 27 Feb 2008, Lew wrote:
One has only to look at how many organizations still use Oracle 8, or Java 1.3, for example, to see how conservative many shops are with respect to upgrades. I'm not saying they should be that conservative, but many organizations are and we must be ready to deal with that.
Companies that act so conversatively are already getting nailed by lack of support in the public versions of software. For example, in 2007 DST was moved around in the US for no good reason, requiring an update to the Olson Timezone Database. If you're a Java user, and you're on 1.3, you couldn't get that update unless you have a support contract--the free version won't do it. (ref: http://java.sun.com/javase/timezones/DST_faq.html ) Even there only a small subset of platforms are supported. Getting older Oracle versions to work there obviously requires the appropriate support contract to see the Metalink update, and as I can tell only 8.1 was updated, people running 8.0 were left out.
If some big-iron shop who is so blind to security issues that they want to keep 7.4 on life support, they certainly can find someone to deliver such a support agreement on a contract basis. But they shouldn't expect the public project to keep them afloat for free, and saying this project "must be ready" to handle them is quite debatable. Given the limited resources of the public volunteers here, supporting ancient versions is a drain it's hard to justify outside the context of such a support agreement. Using your own examples, Oracle and Sun sure don't, why should PostgreSQL?
-- * Greg Smith gsmith@xxxxxxxxxxxxx http://www.gregsmith.com Baltimore, MD ---------------------------(end of broadcast)--------------------------- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
