Father of My Solution for Microsoft to Get Nerds Back
I was catching up on my Daring Fireball feed today and saw that John Gruber had put together a longer post on Microsoft’s earnings shortfall. It’s a worthwhile read, and I think that Gruber is correct in suggesting that Microsoft has begun a very difficult time and is going to be there for a while.
Today that is simply no longer the case. Microsoft has lost all but a sliver of this entire market. People who love computers overwhelmingly prefer to use a Mac today. Microsoft’s core problem is that they have lost the hearts of computer enthusiasts. (Emphasis mine) Regular people don’t think about their choice of computer platform in detail and with passion like nerds do because, duh, they are not nerds. But nerds are leading indicators.
Now, let me be particularly cautious here. This is combustible territory I realize. A couple of months ago I was asked by a couple of people for my opinion of what Microsoft should do to change direction. After considering it for a while I came up with a simple two step plan that would make me think that something big had happened at Microsoft.
Neither of these two steps I’m going to suggest are things that average consumers would notice, or even really care about. My point, agreeing with Gruber, is to do something that will get the nerds to take notice. However, before proceeding, let me come clean that I have been a huge Microsoft advocate. MarketWatch.com is one of the largest financial websites on the Internet, and is entirely built on Windows technology. I’ve advocated with people the benefits of Windows development environments. I’ve even been quoted in Microsoft case studies. With all that said, I can’t imagine using Windows in a new company, and I haven’t seen an early stage company using Windows for years now.
Unix Has Won
My first suggestion for Microsoft is to wave the white flag on the NT kernel. Dump it. Dump it and replace it with a Linux, BSD, Mach or other Unix-type kernel. I’m not going to get into arguments about the benefits of the NT kernel and its VMS lineage. This isn’t a point of threading models or memory management. My point is both a technical and cultural one. It seems unnecessary to continue to shoulder the burden of the NT kernel. Move to an open-source Linux kernel and stop carrying all the water yourself.
What happens with Windows? Not much for the user. Microsoft should do what Apple did when it made Mac OS X. Wholesale change the kernel but keep the user experience pretty much the same. The engine is new and different, but users don’t need to know that. Meanwhile, the nerds of the world can enjoy a native shell with the thousands of Unix executables available native on the platform.
If you’re saying to yourself “Jamie, the kernel is not Microsoft’s problem!” you are likely right. I’m not suggesting that blue screens of death are regular and that the kernel is broken. I’m suggesting that Microsoft needs to adopt a new culture, a new perspective on software. They need to force themselves onto a new path that acknowledges what has happened in the open-source world for the last 20 years. Adopting an open-source Unix style kernel would send a very clear, and very loud message that things have changed. That the future is going to be different. A new path is being broken.
Stop Fighting the Web Browsers
My second suggestion rings similar to the first, but is in a different area. I’m not going to suggest that Microsoft should abandon Internet Explorer. They can and should do what they can to make a great browser that works amazingly in Windows. Internet Explorer should be that. However, they should immediately ditch the rendering engine in IE and move to one of the open-source ones. WebKit, Gecko, anything other than Trident (or MSHTML).
The nerds of the world, the ones that build the websites that everyone uses, know intimately how bad Internet Explorer is. At this point, IE’s lack of compliance and its adoption of web standards is so poor that I would argue they are impeding the progress of the web. Ask any web developer and they will tell you how they could make amazingly better sites if only they could rely on the major browsers to behave well. By and large Firefox (Gecko), Safari (WebKit), Chrome (WebKit) all do. Even Opera (Presto). But Internet Explorer? No way. Microsoft has created a problem for everyone else, and they need to fix it.
Again, changing the rendering engine won’t be noticeable to most people (although when sites work better and load faster they will notice that). This is another move that would signal to the leading part of the technology market, that Microsoft “gets it”. That they have woken from a long slumber and are going to do it right now.
Two Steps, That’s it…
With those two moves Microsoft could wake me up, and I think a lot of other people too. They would embrace the open-source movement. They would stop positioning themselves as them against everyone else in the entire world. And, on top of it they could save money by owning the maintenance of less software, and they would have products that worked better.
Of course, I doubt these things would ever happen. But then again, I also didn’t think Macs would ever have 2-button mice or use Intel chips, and that all came true. Microsoft isn’t a bad company, and they aren’t going away. But they do need to recapture the hearts of people who are passionate about this stuff. This is my recommendation for how they can do that.
Apple’s Two Steps
Friends that have known me for a very long time know that I used to be a Mac guy for many years, and I gave up on Apple completely. I was a Mac user on System 5, System 6 and System 7. I really loved my Macs. But, I bid the platform farewell after experiencing one too many operating system failures. I never looked back and used Windows NT and it’s follow-on versions for a many, many years.
A couple of years ago I converted every computer in the house back to Macs and that is pretty much all I use now. What made me come back? Two things.
First, Apple bought NeXT computer and adopted it’s Mach-based kernel and underlying operating system. They didn’t change the user experience in a major way, but underneath it was all different. A real operating system, that didn’t crash.
The second change was to abandon the Motorola PowerPC processor. Intel was clearly the way to go, but it seemed unlikely that they ever would.
I bought my first Mac again when Mac OS X 10.3 was out. It was still on the Motorola chip, but I took a flyer. Shortly thereafter they moved to Intel chip and I’ve not looked back.
My point in telling this is to highlight how, for me at least, these two changes made all the difference in the world. Perhaps the two changes I’ve outlined above could do the same for Microsoft.
Interesting and I agree that they need to move back to nerdom and away from “big corporate giant”. They clearly have lost that part of their culture. As you point out their technology decisions, for the most part, turns-off the geek community which has a very real impact; albeit the impact is not entirely clear to most consumers but IMO they (the consumers) are ultimately hurt by it and it is starting to really show. However, I don’t think that it is the destination that is difficult for them to define as much as the journey especially in the two solutions you have proposed.
In the case of swapping kernel’s, I question whether they will have the courage to redirect, I assume, a large portion of their current staff to move/migrate to a different kernel. This will cost $$ and time both of which are hard to swallow when earnings are shrinking, your once large lead is slowly evaporating or in some cases they are falling even further behind the category leader. Oh, and the kernel along with everything built upon it is still bringing in Billion’s (Please note uppercase B for Big Billions). Not to mention the uprising of corporate IT shops around the world that either have a perceived or realistic concern in having to support essentially a different platform they feel they are already familiar with.
That said, all of the these are short-term concerns in my mind but doubt they will be able to overcome them until they bleed a bit more.
You make a really good point about the overall cost of doing a kernel swap. I imagine it is the kind of move that you can only really justify when your back is completely against the wall — similar to Apple’s position when it had a complete failure on its hands with Mac OS 9.
So if they are essentially an unmovable tanker they probably will just have to do a better job of marketing?
You are a genius (in general, not just on this topic), and the advice is very sound (and straightforward), but it ain’t gonna happen (I know you know that).
Even though Apple eventually came around to Intel chips, the analogy to the kernel here doesn’t work. MS is who they are because of their focus on developers (one of the reasons you chose them for MarketWatch).
So, if they swap the kernel, they’ll have to port 100% of their various APIs to seamlessly work on the Linux/Unix/BSD kernel (whichever they choose), to keep the current cadre of developers, and keep their tools working the same way as well.
I know that Mono already ports the .NET framework to Linux, and with the ties to Novell, that’s possibly a start, but there are many other legacy API’s necessary.
Finally, if they did what you suggest, their biggest fear would be that projects like WINE would make people choose other “distros” (because Windows, would essentially become another distro), and then use the APIs to port/build apps elsewhere.
Don’t get me wrong, I think your ideas are good, and ultimately could be very workable for MS, but I think that they won’t bother, both out of fear, and because I believe that Windows 7 will be successful, which will buy them a big reprieve!
I’m going to choose to simply read the first four words of your comment and leave it at that. I’m done. No more that I need to see. :-)
Is it necessarily the case that they would become just another distro? I wouldn’t consider Mac OS X a distro, even though it would have those characteristics. You (and Kent above) are both correct in highlighting the enormous expense of making thousands and thousands of API methods work in a new world. For these and many other reasons I tend to agree — this isn’t feasible.
If the kernel recommendation isn’t feasible, so far the two commenters haven’t suggested that swapping out the rendering engine in Internet Explorer is a bad idea. That seems like something that could be done pretty easily, and has little downside (as compared to changing kernels).
Redmond, I suggest WebKit. Let me know when it’s ready.
I think swapping rendering engines is a good idea with little downside that I can think of other than perhaps a bruised ego. Sure it will take some time but I would think that type of commitment could be absorbed pretty easily.
The challenge is that it is not clear to me that affecting the cultural change like the one that is needed and that you have articulated will even come close to being accomplished by changing rendering engines within IE. There may be a few embers of change withinthe organization, but I don’t think would be on par with a different OS kernel.
The geek community would likely stop bashing IE as much as it does if it were to adopt say WebKit, but I don’t think that necessarily means it would adopt the platform. Although, it would certainly be a step in the right direction.
I agree with Kent’s comment entirely! The other issue is JavaScript, which is becoming increasingly critical to the operation of the web, and in which Google, Safari and Firefox lead MS by a wide margin. So, if they swap rendering engines (and I too vote for WebKit), then they should consider what they can do regarding updating their JS engine as well.
As far as “just another distro”, that would apply more if they chose Linux rather than BSD. I think of RHEL vs CentOS (and others) as an example.
If MS were to adopt a Linux kernel, they could only “protect” it, by going the support and upgrade route in stopping people from slicing and dicing it, and offering offshoots (like the dozens of Ubuntu derivatives, which itself is a Debian derivative, etc.
Still, all good food for thought! :-)
I hadn’t realized how many important legacy IE-only applications there were until I was recently re-introduced to the big company IT world. Hell, trying to make IE8 render in a more compatible mode by default caused people to lose it. I don’t think they can ditch IE’s charmingly bad ideas without pissing off all their corporate customers.
It used to be possible to really program on a windows box and feel like you were close to the box. For all the benefits of the developer tools I think they filter too much from the casual folks at the start of their career which then leads them to tools that are ‘more accessible’