Several processes are occurring independently but concurrently, and the shifts in their adoption create artifacts that obscure larger issues. One example: a piece yesterday on ZDNet that offered "Without Cloud, iPad Pro is an executive's desk paperweight." The cloud is optional?
Not anymore. But the cloud is only one driver today, rendering old use cases obsolete and nit-picky hardware arguments passé.
So what are these drivers? Here's some key ones.
OS irrelevance. Some may remember heated arguments over CPU architectures, such as RISC vs CISC. Today we have x86 and ARM, and the arguments are over.
Just as CPUs are less relevant, operating systems are following a similar path. As OSs improve, their functional differences aren't worth arguing about, and attention shifts to the apps they run - which is where the action is today.
Appification of the web. 20+ years ago, at DEC, my engineering team had a hard software problem: to support non-DEC operating systems, we needed multiple clients, and we couldn't afford it. But then we - and everyone else - realized that a lightweight webserver on our storage controller could give much of the needed functionality.
So for the last 20 years we've been packing as much functionality into web protocols as possible. But the overhead has grown with functionality, so now it makes sense to build the optimized apps we couldn't afford 20 years ago. We can do that because we only have a couple of major mobile OSs, and both are built on Unix variants.
Commoditization of pro apps. This is driven by a different dynamic: the rise of the semi-pro creator. If you're editing feature length movies all day, you need - and can afford - an infrastructure optimized for that work. But more and more people are creating short videos, along with photos, text, presentations and graphics.
If you're doing all those things, then the cost of making one go faster has to be much lower than if you're doing one thing all day. And the apps need fewer functions, rather than the ability to do everything. So tablet apps are getting more capable - as are tablets like the iPad Pro and the Surface Pro - but with an emphasis on ease of use rather than lists of unused features.
Cloud integration. Adding cloud options to enterprise infrastructure - mobile, desktop, and servers - radically changes client and server requirements. Not only can client apps be simpler and cheaper, but the hardware required to run them is too.
Just because a $500 tablet won't render video as fast as an $800 notebook doesn't mean that the occasional video producer is crippled. If he can shoot, edit and render on the fly, the fact that some other platform would perform some piece faster is irrelevant.
Likewise, the server market is silently convulsing as the webscale companies drive features that no enterprise could ever cost-justify. Warehouse-sized computers force investments in ever lower-cost infrastructure - goodbye, "enterprise" routers and storage - that inevitably trickle down to enterprises whose IT groups are being pressed by cloud pricing.
The Storage Bits take
Our world now has over 1 billion people online at once, and that isn't counting the nascent Internet of Things, which promises billions more endpoints. This scale is orders of magnitude greater than 20 years ago, and the effects are ripping through computing with tsunami force.
Use cases now range from people who do everything on smartphones to people who are heads down all day on $10k workstations driving 10,000 node supercomputers in the cloud. So forget what you think you know about enterprise use cases, because the world has moved on.
Legacy environments will, like the poor, always be with us. But look instead at where the growth and opportunity is, rather than the past.
Sure, three years ago the cloud may have been optional. But to compete today it is essential.
I've been observing computing technology for almost 40 years, and the rate of change - and the vectors of change - has never been greater. Selling operating systems? That's not a business, as Microsoft's free W10 rollout underlines.
Proprietary storage arrays with limited scale? Bye-bye! Desktop PCs? A shrinking market for years to come. More powerful smartphones and tablets? Yes, and they'll keep getting better, and so will the apps and web-scale services behind them.
I have no idea whether the iPad Pro will be successful, or whether phablets will kill tablets, and don't much care. But convoluted arguments based on yesterday's infrastructure details aren't going to decide anything.
What I do know is that we're lucky to be living in an age of such rapid advances in the technology of machine intelligence. Because the human race needs all the help it can get with the many challenges we face.
Comments welcome, as always.