Snow Leopard was a big improvement over regular Leopard. I think Snow will be the OS that "hangs on" for a long time.
"Hangs on" at what percentage? Everything hangs around at relatively small percentages.
Take a look at the chart here ( click on "Overall" -> "Major Version" while in operating system mode ) :
http://update.omnigroup.com/
The sampling is limited just to folks who would buy and report to omnigroup but is aligned with other results.
10.6 is on a steady decline (current graph covers Jan 2012 to now and it has dropped almost in half to 16% in a year.). Lion was a minor hiccup but Apple largely corrected with 10.8. If 10.9 further removes some issues ( wonky Skeuomorphic GUI ) and value added additions ( newer faster drivers ), then 10.6 will drop on an even steeper decline. 10.9 may also trigger an end to security upgrades, but as long as hovering over 10% they may continue a while longer.
Two years is enough time for even the procrastinators to chuck some of the PowerPC stuff they had been clinging to. Folks who need very long term Rosetta need to get 10.6 server and go Virtual Machine.
Sometime in late 2013 it should get to a similar relatively steady state point that Tiger and Leopard are in where it takes several years to move to get as significant declines. If the security updates stop, then it definitely will.
Snow has the more classical UI and doesn't do wonky auto-save and auto-restore.
Auto-save is largely an application, not OS feature. The library to do it is present in the OS resources provided, but apps don't have to use it.
Auto-restore. If quit apps it makes no difference. If traumatized by it just quit applications explicitly. It is a good, safe practice anyway.
P.S. pre-Tiger reading will fall out of omniware metrics because they'll stop collecting from ancient software that is restricted to OS versions older than what they support with their apps.
P.P.S. The arrival of 10.9 not only bring incremental improvement but it also brings something greater than or equal to 10.8.3. There are significant group that won't move to a 10.x.1 or 10.x.2 of OS X. Usually because it isn't stable enough versus the previous > 10.x-1.3 they are already on. In short, 10.8 should accelerate again now that 10.8.3 is out before it starts to decline.
This stability zone in the calendar is a good reason why Apple probably should move the Mac Pro release zone here. it doesn't have to be exactly the same, but away from the introduction of "new" 10.x+1.1 releases.
Last edited: