Day: October 22, 2015

Where Will the PC Go? -- Part 1: Identifying the Problem

The other day, Ars Technica posted an article called Cringe-worthy “PC Does What?” campaign wants you to upgrade, about a new ad campaign the PC industry is pushing to try and convince users to buy new computers.

The PC industry is in trouble. It's built around a pattern of regular upgrades that customers just aren't buying anymore. And it's trying whatever it can to stop the bleeding.

On the other hand, rumors of its demise have been greatly exaggerated. In the comments thread on the Ars article, someone named erikbc said:

Well, if anyone believes PC is dead they need to get their head checked.
And understand some numbers:

https://en.wikipedia.org/wiki/Usage_share_of_operating_systems#Desktop_and_laptop_computers

A user named has responded:

…said every horse-and-buggy salesman in 1900 ever.

Which, okay, doesn't actually make a whole lot of sense. (In fact I am fairly confident that very few horse-and-buggy salesmen in 1900 ever said "If anyone believes PC is dead they need to get their head checked" and then linked to Wikipedia.) But, like many shitty analogies do, it got me thinking about why it was a shitty analogy.

Mainly, I don't think the PC will go away to the extent that horse-drawn carriages have. I think it's possible that tablets could completely replace desktop and laptop computers, but I don't think that can happen until they effectively duplicate the functionality of PC's -- in effect not actually replacing PC's but becoming them.

General Case: Typical End Users

While it's easy to point to the rise of the smartphone as the reason for declining PC sales, it's only one of the reasons. There's another one: the last processor most end users will ever need was released in 2006.

A typical end user only needs a few things in a PC: a web browser, an office suite, music, and videos. (And those last three are, increasingly, integrated into the first one; I'll circle back to that in a later post.)

In 2006, Intel released the Core 2 Duo, which, paired with even a low-end onboard graphics chip, could handle HD video and drive two 1920x1080 monitors. And it's 64-bit, so it can handle more than the 3GB of RAM that 32-bit processors max out at.

There have been plenty more, and plenty better, processors in the 9 years since. But they're not better for people who only use their computer for browsing, Office, listening to music, and watching videos. The Core 2 Duo was good enough for them.

There are people who greatly benefit from newer and better processors -- gamers and people who produce media rather than just consuming it. But they're special cases; I'll get to them later. For the average user, the difference between a Core 2 Duo and a Core i7 isn't even noticeable.

The computer industry grew up in the 1990's around the expectation that people would upgrade their computer every few years to handle new software. And people just don't do that anymore. They buy a new PC when the old one quits working; not before.

But, at least at this point, they still need a PC. People may be buying more phones than PC's, but, at least in America, a phone is not a replacement for a PC.

Problem 1: Screen Keyboards

Screen keyboards are a pain in the ass.

They're fine for short communication -- text messages and tweets -- but they're just too slow and imprecise for long-form writing. (I thought of writing this post entirely on a screen keyboard -- like last week's handwritten post -- but I think that would make me want to gouge my eyes out.)

There are still plenty of requirements for longform writing in day-to-day life -- reports for school and reports for work, for starters. And that's even in professions where you don't have to write for a living, never mind ones where you do. People who write articles, and especially people who write books, are best served with a keyboard to type on.

And maybe that won't always be the case. Maybe kids growing up with screen keyboards aren't learning to type on traditional keyboards; maybe they're faster with screen keyboards than they are with hardware ones. Maybe, within a generation, we will see essays, reports, articles, even books, all written with screen keyboards. I suspect that if we do, they'll look a whole lot different than they do today.

Or maybe screen keyboards will get better. Maybe predictions and autocorrect will improve. Maybe a faster paradigm than qwerty + swipe will catch on. There's a lot that can happen in this space.

Or maybe we won't be using keyboards at all.

Problem 2: Text-to-Speech

Speech recognition software has grown by leaps and bounds. Terry Pratchett used Dragon Dictate and TalkingPoint to write his last few novels.

But being good enough for a first draft, for a user who is no longer physically capable of using a keyboard, isn't the same thing as being able to recognize a full range of standard and nonstandard grammars and sentence structures, pick correct homonyms, and understand slang and regional dialects. (Pratchett liked to tell the story of how he had to train his text-to-speech software to recognize the word "arsehole".)

Text-to-speech software might be good enough for simple, clear documents, such as manuals, lists, daily work logs, AP-style newsbriefs, and technical writing (provided you're writing on a subject that doesn't have a lot of jargon words that don't appear in a simple dictionary). But for writing that's meant to convey personality -- editorials, reviews, fiction, even this blog post -- text-to-speech algorithms have a long way to go.

So, for now at least, a good old hardware keyboard remains the best way to input large blocks of text into a computer. In my next post, I'll examine why a dedicated PC is still the best thing to connect to that keyboard, and how phone and tablet OS's are (or aren't) working to bridge that gap.