May 2001 column

May 2001 column


[ Site Index] [ Linux Index] [ Feedback ]


Shameless Microsoft Bashing

Hello, and welcome to the Computer Shopper Linux column. This is a column about Linux, not Microsoft. Microsoft has got most of the rest of the magazine to itself, along with a devoted staff of dozens (if not hundreds) of dedicated ShopperLabs corporate business analysts and other professionals who spend their waking hours ensuring that Steve Balmer's gospel is placed before the eyes of the unbelievers. These guys don't need my help. They occupy the capital city of the computer business, periodically sending out helicopter gunships and new proprietary file formats into the hills on search-and-destroy missions. Meanwhile, the open source Taliban -- a hard-core minority of steely-eyed fanatics -- hide in the mountaintops, polishing their Perl interpreters and awaiting their opportunity. One day they will come down from the mountains and bring the One True Word to the heathen masses: for there is but one god and that god is Ken Thompson, and Richard Stallman is his prophet.

To proceed with this horribly mixed metaphor, their time has come. Perestroika is gripping the guts of the giant superpower: there are signs in the heavens and wonders in the waters, for Microsoft is running scared.

(That's enough mixed metaphors -- Ed.)

Back in April 1999, Linux first surfaced on Redmond's radar screen. Steve Balmer was heard saying in public, "There is a level of flexibility, or comfort, that people have when they have the source code, just in case." Well yeah, Steve, we already knew that. In case you're new to the discussion, the whole point of using open source software is that you can't be held to ransom by a software supplier. Which matters a lot to you (the customer), and for entirely different reasons to you (the vendor).

Anatomy of a shark

Microsoft, is an empire built on rising share prices. They're notorious for paying their developers less than the market rate, making up the balance with options to buy shares at a very low price; if their share price stopped rising they'd actually have to pay people. Thus, Microsoft needs to keep their share price rising -- which means they have to keep increasing their market share (to convince the markets that their value as a company is rising). Like a shark that will asphyxiate if it stops moving forwards and sucking in freshly oxygenated waters, Microsoft will run into trouble if it stops eating up marketshare.

Having gobbled up the PC market (their first monopoly sector), Microsoft have been squeezing the pipes for some time. Between 1990 and 2000, the cost of the average new PC fell by 50%, but the cost (to the OEM) of the commercial software they had to bundle with the PC doubled. They've made no secret of their goal to ratchet up the pricing of consumer Windows licenses to the level of (corporate) Windows NT, rather than cutting the price of their commercial offerings. The reason they can do this is because buying software isn't like buying a car -- it's a highly specific service that gives you access to your data, and once you're locked into using a particular package it's hard to get out.

This, incidentally, is also why Microsoft keep changing the proprietary file formats in Office. It stops competitors coming up with packages that can work effectively with the same files, so once a business is on the treadmill of upgrading Microsoft Office products they can't switch over to a competitor's package. Moreover, it forces businesses to buy upgrades so they can remain compatible with the newest file formats people are throwing at them. People who know no better email spreadsheets and word processor documents back and forth in the latest Word file formats, and if you want to read and edit them you'll need to run the latest version of Office.

However, all has not gone swimmingly for the Redmond shark.

Faced with the threat of losing an anti-trust suit brought by the Department of Justice and a group of US States attorney-generals -- to say nothing of the looming EU anti-trust hearing (hanging fire while the US case runs its course) and the French anti-fraud case -- Microsoft began hunting for a new strategy in 1999. The one they selected was Microsoft .NET -- a bunch of services delivered over the internet from Microsoft servers. If you opt into .NET's application services you end up paying a monthly subscription charge to Microsoft, and your data is only usable as long as you keep paying the .NET subscription. They'll even provide a warehouse for your data -- so that you can only get at it over a cable modem or ADSL line as long as you pay the rent.

But .NET isn't here yet, and in the meantime Microsoft needs something else with which to shore up their revenue growth. They're already in trouble: Windows 2000 has bombed in the server marketplace, the DOJ is still threatening them with breakup, consumers are resisting the forcible upgrade cycle that keeps Office revenues high, and Linux is a growing threat.

Whistling in the dark

The upcoming Microsoft weapon is ostensibly a new consumer OS, Windows XP, codenamed "Whistler". It's designed to capture the desktop with a dumbed-down task-oriented interface designed for people who're afraid of computers. But hidden inside XP is a much more serious threat. Microsoft is attempting to grab control of the lucrative audio and video over network market that is emerging on the back of broadband communications (such as ADSL and cable modems). And thereon hangs a rope thick enough for Microsoft to hang themselves by accident.

Here are some things that intrigue me about XP:

Would you buy this pig in a poke?

If you've got any sense, the answer is "probably not": but if you buy a PC running Windows in 2002 this is what you'll get. (And an extra 100 on the price compared to a PC running Linux.)

Against this background, Steve Balmer surprised no-one by announcing publicly in January that Linux is Microsoft's #1 threat and Microsoft needs to attack Linux in the market. Linux is evolving more and more hardware support, a couple of desktop environments are competing for best-of-breed (but will work with each other's applications and plug-in objects). And there's relentless price pressure on the distributors -- while Microsoft's OS price to consumers has doubled in real terms since 1990.

Balmer was diplomatic in comparison with Jim Alchin. Alchin, head of Microsoft's Windows development group, gave a speech this February in which he denounced Linux and open source software as "un-American" and accused it of "stifling innovation". Realising he'd acquired a faceful of self-applied egg he clarified his statement a couple of days later, explaining that it was merely the horrible, evil GNU General Public License he objected to (which doesn't allow a company like Microsoft to take an open source package, splice it into the guts of their own products, and sell it without admitting what they're doing in public).

Verbal attacks aren't the only things Microsoft has been doing. 'Instead of buying out its competition, Microsoft has begun to contact Linux developers, offering job opportunities at Redmond,' Peter Revill reported in osOpinion. 'As I've always said, Linux users and developers are totally different from Microsoft's: "Microsoft is right that (most of) the Linux community dislikes the company. That's because Microsoft is an exact contradiction to everything the open source community stands for. The sheer contempt that so many Linux users have for Microsoft is genuinely understandable, as Microsoft is incredibly reluctant to do anything to help open source and is known for actively "attacking" it.'

Meanwhile. IBM are taking Linux seriously. They've got 1500 developers working on Linux already, says CEO Lou Gerstner; they're pumping a billion dollars into application development this year, and they're about to launch a huge advertising campaign promoting "peace, love, and Linux" all over Time Square and your television set. I wrote about this six months ago in Shopper 154: IBM sees Linux as the answer to some huge problems that have bedevilled it for two decades. IBM is still, despite the Microsoft hype, the real 800-pound gorilla of the computer industry.

This morning I picked up a copy of "Computing" -- a wide-circulation weekly trade rag that regurgitates only the purest management marketing speak. Lo, I read a Windows column in which a talking head not entirely dissimilar to myself (except for being less hairy) droned on for half a page evaluating the first anniversary successes of Windows 2000 in the corporate marketplace. He concluded with a phrase that induced severe cognitive dissonance in my head: "nobody ever got fired for buying UNIX".

But these aren't the only signs and portents and two-headed chickens in the manger, fortelling a sea-change in the industry.

That well-known polling organisation IDC does a regular survey of who's running what in the marketplace. This year's poll of the desktop shows no surprises -- Microsoft in #1 position. They downgraded Linux, due to difficulties in deciding how to estimate how many people run it; I think they're focussing over-hard on the corporate sector (which is always the last to change desktops) so it's no surprise that they don't see a huge Linux sector. But the story is different in servers.

The server sector breaks down as: Microsoft 41%, Linux 27%, and UNIX 14%. Leaving aside the fact that Linux and UNIX are effectively part and parcel of the same thing (and, with a combined share of 41%, are running neck and neck with Windows), they also grouped Windows NT and Windows 2000 under the same heading. Windows 2000, we are told, has only just past its first million sales. Total Windows server sales in 2000 ran to 2.5 million, so Windows 2000 sold less than 1 million units; Linux in contrast sold 1.6 million servers. So the figures are deceptive -- NT 4 was still selling until withdrawn from the channel, but Linux was out-selling Windows 2000 by more than 60% in the market sector that Microsoft was targeting their new OS on.

So.

Let me get this straight.

Dot NET is still a couple of years away from fruition. Windows 2000 has not taken the market by storm in the corporate sector; in fact, a lot of people are holding off upgrading to it, and Linux is out-selling it by a wide margin. Steve Balmer now officially rates Linux as Microsoft Enemy #1. They're trying to buy the opposition, but the opposition refuses to be bought. They can't undersell this competitor because the floor price is zero. So they're trying to lock up the consumer market -- before the sheep begin to stampede -- by rolling out a new system (Windows XP) that seeks out and destroys your MP3 collection, won't run with your existing hardware, and vapes non-Redmond-approved software. Meanwhile the corporate IT managers are clutching their Solaris licenses and beginning to chant "nobody ever got fired for buying UNIX", while the CEO of Compaq indignantly splutters "Linux will never get anywhere on the desktop" in response to IBM pumping a gigabuck into application development on that platform. Despite the fact that 70% of Compaq's alpha workstation sales are running, er, Linux.

I'm getting a hideous attack of deja vu; it's like 1993 all over again.

In the old days, Microsoft dealt with rivals using a three step program -- embrace, extend, and assimilate. In the first phase, they'd embrace a new technology area that some other company had opened up, bringing out their own competing clone of the market-leading software (and making it fully compatible with the market leader's file structures). In the second phase, they'd extend their own software, introducing incompatible features designed to break the market leader. Finally, they'd buy or browbeat their rival into oblivion, if need be giving away product (like Internet Explorer) until the enemy is driven into bankruptcy.

But Microsoft is suffering from hardening of the arteries. They could have embraced and extended Linux by pre-announcing their own distribution ("Microsoft Linux Server for Windows! Read all about it!"). They can't totally assimilate Linux because it's free, but they could have stamped out the commercial competitors and marginalised the open source community. They're used to building applications on top of operating systems they don't own; in fact, Macintosh owners on average spend more on Microsoft applications than Windows users do. All they had to do was add Linux to their stable and treat it as a cash cow.

But they didn't do that.

Instead, they succumbed to "not invented here" syndrome: Linux was alien and foreign and would not be dealt with as a platform. They took to throwing abuse. Alchin even forgot Microsoft's first law for dealing with a rival -- show no sign of fear. His attack on Linux and open source in general ("it's un-American! It stifles innovation!") betrays a pathological phobia of losing control, and shows weakness where Microsoft can least afford it -- in their ability to stay ahead of the ball.

Microsoft's market dominance was built on the foundations of IT managers' collective belief in their infallibility. As long as Microsoft looks like a monolith, nobody will ever get sacked for buying Microsoft products. But if Microsoft looks uncertain, those IT managers will lose their confidence that Microsoft will still be showing them what to do in ten years' time. Then they'll look elsewhere for reassurance. As more and more of them stop automatically believing the spin coming out of Redmond, other IT managers will feel the chill wind of insecurity breathing down their neck. The house of cards that is Microsoft's reputation for infallability will crumble, leaving nothing but another medium-sized software company in its wake.

It happened to IBM in 1993-4, and it's happening to Microsoft in 2001-2002.

Linux is on the edge of capturing and dominating the server market -- it already outsells Microsoft's main server OS by a margin of 3:2, and it's steadily gaining market share. Linux has the backing of IBM from the top down, and a sizeable internal group within Compaq is pushing that other behemoth towards it. In the hand-held field, various big names (notably Sharp) are rumbling about moving away from Windows CE and bringing out palmtops based on Linux; the first are already here, and some sources are predicting a 40% market share for Linux by the end of 2002. The embedded systems market, where Microsoft effectively has no presence, is already succumbing to Linux. By the end of 2001, the only market sector where Linux has not crushed the Microsoft hegemony will be the desktop PC's -- and with IBM advertising and promoting it, and the likes of GNOME 2 and KDE 2.1 as the interface and OpenOffice and Mozilla as the application base, Linux will be a credible office desktop system.

Just as IBM didn't realise that the new-fangled PC's were undermining their corporate strategies before they were in place, Microsoft doesn't realise yet -- at an institutional level -- that this new-fangled Linux and open source stuff is undermining their long term strategy. But the writing is already on the wall.

The outlook for the future

First, let's get one thing straight: Microsoft is not going away. Their market share on the desktop is so huge that even if they decline by 10% per year for the forseeable future, it would take more than a decade for them to shrink to the size of Apple. And, realistically, they aren't going to decline that way, unless an outside force makes a determined effort to stamp them out (using the same techniques they pioneered for grabbing an information industry by the throat) and they forget to resist.

However, the desktop PC isn't the entire computer market -- or even the largest part of it. Microsoft's ten-year attack on the server market has clearly failed: within this year it will become apparent that a new contender has come out of nowhere to eat their lunch. Likewise, a number of PDA manufacturers (notably Sharp, Samsung, and Compaq) are moving towards Linux as a cheap alternative to Windows CE. Microsoft has made little more than half-hearted efforts in the huge embedded systems market -- the 80% of computers that are invisible, locked in car engine systems and video recorders -- and Symbian brutally kicked them out of the 3G phone market before they even noticed it's existence.

Microsoft are therefore bottled up on the desktop PC, and desktop PC sales are flattening, due in part to a recession in the US, and also because the market is now saturated (everyone who wants or can use a PC already owns one). All they can do is keep selling new applications and operating systems to a market that is no longer growing: this won't bring in the structural revenue growth that they've become accustomed to for so long.

The prognosis is this: by 2005, Microsoft will still own 50-90% of the desktop. But 90% of the computers out there will not be running Microsoft products, and what Microsoft chooses to do will be largely irrelevent.

They'll be back to being just another software company: the end of an era will have arrived.


[ Site Index] [ Linux Index] [ Feedback ]