Reply to post: Re: "software had become powerful enough to serve users well enough for years"

Re: "software had become powerful enough to serve users well enough for years"

No, there was a time when software upgrades bought improvements big enough to be worth the price and effort, even on the same hardware.

On the server side you saw and reaped real benefits from going, for example from NT to 2000 (AD, especially) to 2003 (64 bit support, especially). After that, improvements were less far reaching for the majority of customers, usually just for the high-end ones, the others saw just new costs and little more.

On the desktop side moving to 2000/XP from Win 9x meant sounder systems, while from NT it meant improved device and graphics support. 7 added the full 64 bit support XP lacked, while correcting the Vista defects. While both 8 and 10 still add improvements, again they are less compelling than before, while introducing unneeded issues.

This is true also for a lot of application software, both server and desktop side. New releases may offer little new unless you are a specific high-end power user needing exactly that feature. A new DB offering fancy NoSQL and BigData features may not interest you at all when you're just running the same accounting/invoicing workloads. Adobe moved to the subscription model too because it had issue to compel most users to upgrade Photoshop as well...

Sure, some of those upgrades often required new hardware, but not every time. I had system for which the hardware was fully exploited only after a newer OS was installed - upgrading to 64 bit for example resolved a lot of issue due to 32 bit memory constraints, and it was worth the price of the new OS (on the HW side, it may have meant just some new RAM modules).

Now hardware has outpowered software up to the point you can consolidate many servers on the same hardware and still run the same workloads - but that's an issue for hardware vendors, not software ones.