Upgrades were a decision-less choice in the early days of computing. When a new version came out, you installed it. No question.
Today, it's the opposite. Microsoft comes out with a new operating system after five years, and most corporations say they will wait a year to two before thinking about upgrading. In the 25 year history of personal computers, two things have changed:
1. Software has become good enough. This is the easy one to figure out. Early software was so lacking, that upgrades were eagerly awaited. For instance, the first release of Borland Turbo C lacked graphics libraries. V1.5 was a no brainer upgrade, because it made it easy to draw graphics.
For the last decade, word processing software has become so mundane that I use Atlantis from Rising Sun Solutions. It features incremental upgrades: a few new features every couple of months that are downloaded over the Internet in a minute or two. When the software starts up, all my open documents are open again, and a dialog box points out the new features. There is no major release that changes the user interface and my way of working. I don't use Word, so the discussion of whether to upgrade to Word 2007 is immaterial to me. (For university students on tight budget, I recommend the free Abiword.)
2. Disruption has become a negative. The early days of personal computers were disruptive. That was the whole point to them. Getting your own was disruptive to your pocketbook (my first pc system cost me $6,000); figuring out how to use the undocumented MS-DOS v1.1 and badly documented CP/M 86 was disruptive; figuring out how to communicate with other computers was disruptive (a null modem cable was the cheapest); upgrading to the next release of software was disruptive (MS-DOS v.2 was soooo different!).
But that was alright, because we wanted to be disruptive, and be disrupted. Disruption was good in the 1980s, because it meant progress, like converting the dual floppy drives to double-sided so that each diskette could now hold 1.2MB.
The first sign that disruption had become bad was WordPerfect 6.0. This major release changed the user interface from efficient text to slow-as-molasses graphics. WP v6.1 didn't do enough to fix the problems. WP for Windows was as bad. In the year 1993, "upgrade" became a dirty word.
And the feeling around disruption changed from positive to negative. Today we scowl at the disruptions caused by upgrades, and how they hurt our profits:
- loss of profits from the cost of software upgrades.
- loss of income while employees learn the new interfaces and ways of doing things.
- loss of profits from upgrading hardware.
- loss of prestige when the upgrades provide negative benefit.
The same for hardware. Before CPUs speeds ground to halt in the mid-3.0GHz range, I had already stopped upgrading my computers for speed reasons -- because there was none. I had a rule-of-thumb of upgrading when new CPUs were 3x faster than my existing desktop computer's speed. My current, several-years-old desktop computer runs at 2.4GHz. I don't anticipate that a 7.2GHz model will be available this decade or next.
About the only thing I upgrade these days are hard drives, and that's for added capacity. Even when I make the purchase, I wonder how short a time it'll be before the new unit disappoints me by breaking down.