Behind the story: "why tablets are killing PCs" - no, really, they are. Here is how.

[liam_on_linux]

The title is from this Guardian story by Charles Arthur: 'My iPad has Netflix, Spotify, Twitter – everything': why tablets are killing PCs.

There are a whole bunch of competing factors here which seems to baffle many observers. This isn't an encyclopaedic list, but...

• The rise in x86 performance fell off a cliff in about 2007 or so.

PCs aren't getting much quicker any more.

This means that the less-informed (or more specialist) media report rises in multicore performance & GPU performance, which are irrelevant to most people. That blurs the picture -- most people, including in the PC industry, don't understand the difference.

It's also meant that PC sales have stalled. 2013 PCs aren't vastly quicker than 2008 PCs.

The few upgrades that make a real difference now are things like SSDs. Otherwise, capabilities now exceed the needs of typical users & only improving in irrelevant ways (e.g. surplus RAM; surplus disk storage; surplus unusable CPU cores.)

Now, as Arthur says, it's turning into a replacement-only market.

Big deal, you may say -- but this fact alone will kill most of the players in the PC industry. It is an industry built and predicated on rapid performance increases leading to one of most rapid replacement cycles in hi-tech manufacturing. Like BA competing with Ryanair & Easyjet et al, it /cannot/ transform itself into the rivals that are killing it.

Also note that in that same timeframe, *the* big PC S/W vendor starting doing backflips in an effort to force upgrades, and many of its efforts are duds -- e.g. Vista & Win8. But more importantly, consider Office 2007 et seq: a needless & ineffective UI revamp, a file format change, both desperate bids to make people buy new stuff. The same is happening on the server side -- Windows Server 2003 & Exchange 2003 work as well as they ever did. But upgrading is a massive (and very expensive) PITA, so one might as well shrug & outsource it to a hosted solution. Cheaper /and/ much easier.

Also see: storage capacity increasing to the point of irrelevance for most users.

• PC OSes are complex & require considerable maintenance.

However, now, the main use for PCs for most users is Internet access, which is technically quite easy by modern standards, now offers a rich environment, app-like functionality, personalisation, data storage, etc.

You don't need a PC for this. A phone or a tablet does it, is easier to use, and requires way less maintenance and tech savvy. PC technical types are mostly utterly blind to just how complex & arcane PCs are. Yes, this includes Macs.

In fact, mainly thanks to Ubuntu, Linux is less work than either now. Both PC and Mac types are also blind to this as they cannot see past to-them essential tools and apps which Linux doesn't offer. But that doesn't matter, as we're entering a post-PC era.

Servers are as hard as PCs to run, but using virtualisation and datacentre hosting, they become disposable -- you keep a small herd of fungible tech staff, in 2 grades: [1] Windows grunts to do the basic 1st/2nd line stuff by remote control, tossing & replacing broken VMs rather than fixing them. This is coming to desktops too, fast. [2] A small number of expensive *skilled* techies, increasingly running *huge* numbers of FOSS servers where licencing doesn't matter. Doing this makes servers relatively cheap & easy.

Result: commercial OS vendors entrapped in a pincer attack: replaceable client devices, FOSS servers in the cloud, declining need for expert tech staff.

• Older users as ever wedded to legacy kit & methods.

This is just human nature. Few in IT are old enough to remember the last big transition -- to GUIs, 20-25y ago. Essentially *all* IT experts dismissed the new tech as a useless toy, a pointless distraction. Now, a precisely analagous transition is occurring again, but everyone's forgotten.

There's a lot of legacy tech that now seems essential to old-timers which are actually clunky, cumbersome distractions that can be discarded with little or no loss: windowing "desktops"; menus; hardware keyboards; pointing devices; expansion slots; removable media.

It is the late '80s/early '90s all over again. Then, ordinary users didn't need GUIs, GUIs were toys, that wasted power and time. Ordinary users didn't need multitasking. Ordinary users didn't need multimedia. But to get work done, people needed efficient text screens, they needed the performance of text-based software, they needed the power & reliability of Big Systems with Big OSes.

Now, ordinary users don't need touchscreens. They don't need tablets. But they do need mice. (_Mice_, FFS.) They absolutely must have windows so that they can see more than one thing at once, side by side. The new devices are /unusable/ because they will get fingerprints on their screen. (Oh the horror!)

The majority of the players in the PC hardware industry cannot and will not transition to a small, slow, gradual-replacement-based sales model. Ergo, they're dead, like everyone from Apricot through Micropolis to Wabash and Xebec, who failed to make the last transition.

The PC software industry is similarly predicated on rich clients running rich local apps, with regular replacement based on much faster clients & added features. That model is dead, too. And with it goes the model of expensive proprietary back-end systems running hardware-restricted licenced software. Both of those are as dead as minis and mainframes, they just haven't seen it yet.

When DVDs replaced videocassettes, everyone didn't throw out their VCRs overnight. It took years. Ditto CDs and compact cassettes & LPs. Many cling on to their old media & old players. I do myself. I'm typing on a 22yo keyboard, almost totally surrounded by print books.

So sure, yes, today, tablet owners have PCs too.

But the PCs won't be replaced & will eventually be thrown out. The toy tablets will get locked-down deskbound versions that will fill millions of office desks & those offices won't /have/ server rooms or local servers or local IT tech staff.

But it won't be overnight.

And I confidently predict that every single case of "essential", "indispensable", tech will in a few years seem laughably retro to 99.9% of users. WIMPs will be as retro as a commercial Unix box running Motif apps on CDE is today: in other words, tiny numbers of users will pay lots of money to do it, but it will be utterly irrelevant to all but a tiny number of specialist users.

"... a new scientific truth does not triumph by  convincing its opponents and making them see the light, but rather because its  opponents eventually die, and a new  generation grows up that is familiar with it." (Max Planck, 1949.)

Well, IT is the same... only the old techies don't need to die, they just need to be promoted into management.