Microsoft Windows: The Next 30 Years

Is Satya Redmond's Atatürk? Or พระบาทสมเด็จพระปรมินทรมหาจุฬาลงกรณ์ พระจุลจอมเกล้าเจ้าอยู่หัว ?

King of Siam ... Chulalongkorn the Great พระจุลจอมเกล้าเจ้าอยู่หัว

Analysis Windows anniversaries are a bit like Halloween. You can bring out a Vista or Windows Me to scare the children into bed on time. Microsoft Windows turned 30 this month, and blogs are full of nostalgia.

(Here's our Fright Night – we dare you to click.)

But what will the next 30 years look like? Is the last 30 years of Windows a reliable guide to the future? Is Windows, as many believe, now a legacy platform, a dying franchise that's destined to suffer a slow death of maintenance releases, while more nimble mobile platforms scamper past it?

I've been as critical as anyone of Microsoft's decision to release Windows 10 on PCs and (imminent) mobile before it's ready, with a compromised design that makes life harder for users. But from Microsoft's perspective it's undertaking a gigantic exercise that might just keep the Windows code base relevant for the future.

In fact, I'd compare it to the reforms undertaken by Atatürk and successive modernising monarchs in Thailand. Particularly Chulalongkorn (aka พระบาทสมเด็จพระปรมินทรมหาจุฬาลงกรณ์ พระจุลจอมเกล้าเจ้าอยู่หัว).

Thailand is the only Asian country never to be colonised, with Chulalongkorn abolishing slavery, setting up Western-style institutions to deal with patronage and corruption, and adopting full Western dress. Atatürk took the defeated, 700-year-old Ottoman Empire and proceeded to abolish the Caliphate and create the modern, secular Turkey we know today. These were extraordinary achievements.

The analogy here is that unpopular reform came from inside, fending off future woes. It's what SatNad (Satya Nadella, CEO) has embarked upon at Microsoft.

In doing so, SatNad has got lucky. For the first decade of the smartphone wars, when every byte was sacred, we assumed that Windows would forever be too resource-intensive and crufty to make the leap from desktops and laptops to mobile computing. It wasn't a new assumption; in the early 1990s many had thought that bloated desktop platforms were completely inappropriate for mobile: hence the early 1990s hype around Newton and Go.

Photo of HP Tablet PC running Windows XP Tablet PC Edition

A svelte Tablet PC – how on earth did this fail?

Bill Gates hardly helped. His vision of mobile computing was the Tablet PC, remember. His idea of mobility was all the inconvenience of desktop Windows, with none of the convenience. And Windows was supposed to be doomed too, because it was so poorly hacked together. Modern Windows had started well, as a VMS-inspired clean room project led by VMS' designers. But expediency forced Microsoft to compromise. As the multiyear death march to release Windows 2000 drew to a close, Microsoft faced a shipping deadline only a few weeks away, with over 63,000 bugs still active.

(Only this static code analysis tool helped avoid Windows 2000 being a calamity.) By 2003, the Windows XP code base was being described as "a house of cards ... precariously shifting in the breeze".

Photo of a Microsoft Stinger phone

Microsoft's first phone, based on "Stinger", c.2002
... because you always wanted a Programs menu
on your mobe

Microsoft got lucky – and every company needs a lucky General – for two reasons. Very quietly, Microsoft began to drop support for archaic subsystems, sacrificing compatibility for performance: the MinWin refactoring project. By 2007 Microsoft could demonstrate a functional version of Windows that had the requirement of desktop operating systems from over a decade before: around a hundred files took up 25MB of disk space. The benefits weren't really felt until Windows 8 in 2012 – but Windows 8 represented such a radical disruption (and not in a good way) for users, that most people didn't notice.

WinMin meant Microsoft could take advantage of the extraordinary increase in mobile CPU power, ironically prompted by the mobile industry adopting Java as their platform of choice to take on Apple. Android needed huge improvements in graphics and CPU, but capitalism cranked into life: the demand for iPhone-alikes ensured they got these at relatively low cost.

All this means that the legacy Windows code base can now be squeezed into a modern mobile without being teased for being overweight. And it set the stage for SatNad's major modernisation effort, which extinguishes the line between mobile and PC completely.

Two new Lumia phones available in a few days will not only run "desktop" or "legacy" Windows, but they'll happily substitute for a full-blown desktop. It's a great party trick with limited use cases – and it's evident even from the demos that you're getting a subset of a PC on a Windows 10 Mobile. But the intent is clear enough, and the experience can only get better. Technically, Microsoft has already ensured Windows is capable of being relevant for the future, even if the apps, the content and the users aren't there.

As to whether Windows can survive the next thirty years, that's a harder one to answer. The world of personal computing hasn't really changed radically since 1985, it's just got more colourful, and there are more computers of different shapes and sizes. Most of them are mobile and Microsoft is almost a negligible presence here. But the modern Microsoft has added value to Windows in a way few would have thought possible a decade ago.

This leaves serious questions (typically from shareholders) whether the crowd revenue stream can replace the traditional PC OEM revenue stream. Don't forget that Windows 10 Mobile is given away for free, while Microsoft still relies on PC OEM licensing fees. But if SatNad turns out to be Microsoft's Atatürk or Chulalongkorn, he will truly have done the impossible. ®


Biting the hand that feeds IT © 1998–2017