My take on it is that companies will pay for upgrades providing the perceived benefit is of value to those approving the upgrade.
We all know that the cost of the upgrade is trivial when compared with programmer's wages, but, unless there is a "do-or-die" reason to upgrade, there will be resistance to buying upgrades.
If the cost can be directly passed on to the client, that seems to ease the pain quite a bit (do I sound like I'm talking from life experience ....hmmmm... yes I am).

I will also say that there has been no earth-shattering data access performance breakthroughs since Rushmore.

Many moons ago, I sold foxbase (..yes foxbase) vs dbase III to a client via a simple test. dBase III code was run and timed and then the same code was run using foxbase. Foxbase compiled the code, reindexed the dbase files and ran the query in about half the time, total!!! THAT was an easy sell!!!

I really don't see any MAJOR performance breakthroughs happening. However, perhaps we could see table barriers broken: max table size, max fields, security, single file for a database (like MDX?).

When we try to sell upgrades to "those wot hold the purse strings" we need to be armed with perceived tangible benefits to avoid the "we'd really like to but..." syndrome.

(descends from soapbox)


>The real question was, in actual practice, why do many people stick to older versions, despite these improvements. I have my reasons; other people might have others.