What Happened to Microsoft?

2005-04-23T15:00:00.000-07:00

I was at Microsoft through the upsurge of the 90's and the slowdown of the past few years. A big theme in the trade press these days is the supposed decline of Microsoft, and what went wrong.

Microsoft operates within the law, and it competes fairly. It hasn't ceded any major market or screwed up in any real way. But you can't blame market trends either. By its own public statements there's something very wrong there organizationally. And they know that shipping Longhorn someday isn't going to turn things around. No one person can make sense of the entire situation, and there is no surgical solution. Any effective change is going to be broad, coarse, and approximate, going way beyond a reorg.

So what the hell; I figure I'm as qualified as anyone to suggest a fix. Microsoft has world class leadership, and a huge pool of great developers. But in between them is a byzantine mess. My experience is that the entire dysfunction at Microsoft is with the low two layers of management. Some were and are good contributors; many were washouts as devs; the problem is with their roles. To make Microsoft perform again, reassign the first two levels of managers, across the entire company.

In a conversation the other day, I said they should all be eliminated; but that's not fair -- let them code, test, PM ... anything other than posture, argue, complain, power grab, act sensitive, suppress developer creativity, and make each other feel useful. At the end of the day, these people are stakeholders, and will benefit along with everyone else from a resurgence as an innovative growth company. The competent ones would retain leadership roles through respect by their peers.

One of my last surreal experiences at Microsoft was one of these managers scheduling a morale lunch, and actually calling it "morale lunch" (as opposed to lunch, group lunch, or something else that didn't stigmatize it as a band-aid for some big problem with morale). He probably put "held morale lunch" on his status report. That's the kind of Dilbert-like, management-by-numbers mentality now widespread within the company.

Let low-level leadership roles self-select informally from the ranks of devs. Let devs peer-review. Let the company benefit from the enormous productive capacity of the individual contributor.

Then, get rid of private offices. Isolation desocializes and polarizes people, and stifles information flow. People become anxious and mean. Let everyone breathe the same air. Eliminate anonymous complaining -- all views should be open and attributable.

I was once deeply passionate about Microsoft. Even though I'm with a competitor now, I still care about it. Microsoft is genuinely dedicated to broadening choices for users. As I see it, the current uncertainty stems from changing realities in the age of ubiquitous broadband.

What's Happening to Web Development?

A large part of the anxiety relates to the state of web development. Web designers are in a state of confusion, and there is no solution in sight.

The panacea for a web app designer is a language that will run performantly client-side, execute in the browser, be secure, be portable, do everything needed by the app, and incur no download or security hurdles. No one language or model delivers that today.

Avalon addresses these things, but it won't be widely supported for years even after it ships, and portability is still a question.

Meanwhile, IE market share is dropping. So what should web developers do? For many, the answer is to fall back to basic browser features: HTML and JavaScript. There is a growing programming trend called AJaX, in which Javascript-manipulated DHTML plus small functional plugins are used to hack graphics and facilitate an interactive client experience with sustained connectivity to the server.

AJaX is truly evolution in action. It is grossly inefficient, and unintelligible. I can tell you first hand that it's excruciating for developers. I can't imagine a CS course ever teaching AJaX. But it is portable and hurdle-free for users. It depends only on basic browser features and lots of CPU. Microsoft can call these hacks all it wants; this is what is delivering the right user experience today.

An AJaX web app can't do actual graphics, but it can produce smooth graphics effects through DHTML. It can't access the file system, but it can expose an internet drive. It can't do peer networking or federation, but it can achieve peer-like effects through its server.

My favorite browser is IE. I like it because it runs each browser as a separate process and because only it supports ActiveX controls, which are pretty much the only way to get code from your server running natively on a client machine without a big installation process. My game site uses an ActiveX control because it needs to execute natively on the client for perf, and needs access to graphics, the file system, networking, and the registry.

An ActiveX control was my only choice, so my game site only works with Windows/IE/x86; that should make Microsoft happy, and you would think they'd do everything possible to maintain and enhance the attractiveness of their browser for web developers like me.

I'm not so concerned about non-portability, since IE is still so dominant. The bigger problem for my site is the security hurdle associated with the ActiveX control. ActiveX controls, even signed ones like mine, can be disabled by IE; or they may trigger a security warning on first install. I lose most of my users because of that scary security popup, despite having undergone a screening process with Verisign to obtain my signing certificate. I would gladly pay much more for a certificate or process that validated my ActiveX control as fully safe, so that users would not be scared off by the security warning.

So here we are with a near-perfect (albeit non-portable) web development paradigm from Microsoft; one that would be good enough for most developers, and that would help keep IE dominant. The only thing preventing that hurdle-free experience developers want is Microsoft's security paranoia. For signed ActiveX controls, Microsoft should relax these security hurdles.

For example, the warnings might be limited to a trial period, after which the control would be deemed fully safe if no one complained; the warning could then be removed or scaled down. Or, Microsoft could charge for a review program that certified the control as fully safe. Or, some funds could be held in escrow in case the control was found to have malicious content. There are many simple solutions.

Unlike most sites, I can't just switch mine to Javascript because it is so perf-critical (I found JavaScript to be about 100 times slower than hand-optimized x86 code for number-crunching) and it needs various OS features. But the more CPU power is out there, and the more of these little extensions there are, the more attractive AJaX becomes. Furthermore, the more IE's browser share keeps dropping, the less attractive ActiveX controls are. At some point, I'll have to abandon the IE-centric solution, whether there is a good alternative or not. Microsoft should pay attention to that.

What's Happening to OS's ?

Linux works. You can order a machine from Dell today and Linux is no longer a specialty option -- it's offered right there on the same level as Windows, especially for servers. Dell is no longer afraid of the support costs related to a less intuitive OS.

When you get Linux, you're getting a full blown web server, a bunch of office productivity tools, etc. Almost every big money-making Microsoft product has a counterpart in Red Hat Linux Pro; many Linux features are perceived as higher quality than the Microsoft counterpart; others have no counterpart.

Many people have pointed out the threat to Windows as Linux improves usability. That's only part of the equation. The other half is that as the public becomes more tech savvy, they can deal better with any remaining obscurities in Linux. Linux usability and the average user's tech savviness are racing towards each other; when they meet, Windows may really get hit. I'm not saying anything brilliant here, but I've never heard that latter point made: the usability threat is not just a slow trend; it's a collision waiting to happen.

A lot of that usability in Windows comes from GUI components, e.g. wizards, that step users through management tasks. Well, on the server, that can be a minus. Unix text-based commands are easier to communicate and script; Microsoft is now backpedalling somewhat to make sure that all server management tasks are scriptable.

By the way, there are rumors of Apple's OS X being ported to x86 this year. That would be an immediate highly usable alternative OS on the PC.

The other big Linux driver of course is dropping PC prices; the price of the OS is felt more as it becomes a larger relative component of the PC purchase. PC manufacturers see the OS as a component, whose cost should be minimized.

Finally, three well-understood technological trends diminish the relevance of the OS, which favors the least-cost alternative:

First, class drivers: peripheral busses like USB allow devices to be self-describing, so that general class drivers can be written to support peripherals such as cameras, regardless of the manufacturer; it is now easier to produce an OS with good built-in device support.
Second, full digitization eliminates some peripherals altogether; e.g. you may feel less and less like you ever need a printer. If you get all your docs in digital form, then you never need a scanner.
Third, web apps diminish the OS's role as an application platform. And ubiquitous broadband makes us care less if the application gets sucked across the internet; or if the app's storage is to an internet drive.

Windows is by far the best application platform, but that's becoming less relevant. Modern users increasingly don't want state on the PC -- be it applications, data, or attached peripherals. They want their state to be on the web, with the same functionality accessible everywhere. To have your state everywhere, you can carry it around in your pocket, or you can put it on the web; just don't leave it on a PC.

Back in the 90's, I would have had a hard time believing that Microsoft or Windows dominance could be threatened. Even now, most of the bad news is perceptual, and amplified through the media and stock market. But no one speaks of the benevolent cycle of Windows and application developers anymore. It's now about the browser and web applications. What is Microsoft doing to make its browser a preferred platform for web developers, right now, today?

---------------
Ervin Peretz is a software engineer at a major web search company.
He is author of HoldemKiller, a free online poker bot.
He is author of BibleCodex, a free Bible Codes research tool.
He is founder of Terra Bite Lounge, a voluntary-payment cafe/restaurant chain.

Ervin's blog

What Happened to Microsoft?