Bear Flag Republic Radio Weblog

< 2:17:05 PM        >

Robert Heverly took a lot of time out to write a blog post titled "what Scoble doesn't get." I highly recommend everyone at Microsoft read it.

The short version is that he doesn't like when I start defending a vision of the PC as an appliance.

Personally, I'm actually mostly on his side of the argument. Why? I want a system I can tinker with. Hey, look at my post about the kinds of defaults I want turned on in Longhorn. You'll notice that most people said I should not be allowed anywhere close to the designs of Longhorn. Why did they say that? Because most users are not like me.

This is where both Robert and I make a fatal mistake: we want a computer that works for us and we want to say the hell with the rest of the users.

That's one danger of talking about computers in weblogs. I'm pretty sure there are very few "normal" users reading this weblog (or Robert's). Why do I say that? Cause, there are what, several hundred million computer users in the world, and only a few thousand who hang out here. The readers I have interacted with here are way way way high on the bell curve (translation: they are advanced to super advanced users and developers).

On the other hand, in airports I talk with dozens of "regular users." None of whom yet read weblogs. Of the hundreds of "average people" I've met in the planes or at the bar in the airports, I've only met a couple who had weblogs. They both worked at Microsoft. (Yesterday I met a third, who worked as a security guard, but more on that in a later post). Which tells you how skewed this little world is.

But, is there a way we both can win? Longhorn shows we can.

By the way, Longhorn is a code name. There's no way that'll be the name of the final product. After our marketing department gets done, and our lawyers get done, and our executives get done, I'll bet we end up with a far more boring name like "Microsoft Windows 200X Client Edition." Personally, I wish we could change the name to something else, but that's another weblog post I'm working over in my head.

So, what did Robert ask for?

1) Transparency. He wants Longhorn to show him what's going on under the hood. Actually, Longhorn will be more transparent than XP is today. We haven't disclosed all the ways that'll be true, but certainly developers will be able to peek under the hood in new ways.

2) Defaults (particularly in reducing the footprint so that the OS is far more secure, he uses the example of the messenger dialog boxes that pop up if a spamware attack is going on). Absolutely agreed. Actually, in Windows XPSP2, you'll see sizeable strides made to do that. Long before Longhorn.

3) Configurability (particularly when it comes to devices and networks). We haven't yet discussed these aspects of Longhorn publicly. But, totally agreed. I think we'll have some pleasant suprises for Robert.

4) Don't include apps like Outlook Express, and don't leave them on by default. Oh, here's a point where I disagree. My vision of the computer should be that it should be easy to do a wide range of things on the computer. One of those things, I believe, is participating in an NNTP newsgroup. That's a scenario that I want normal people to be able to do. I do not want them to need to download an NNTP news reader first. So, yes, I want Outlook Express included by default. Now, the trick is to make it more secure. I do agree with Robert there.

This is a fundamental disagreement, though, between my vision (note I said my vision, cause I can't presume to speak on behalf of 55,000 employees -- you'll actually find quite a few disagreements between these two camps internally too) and Robert's vision.

I want a computer with a wide range of scenarios supported out of the box. I want "average users" to be able to do things like watch a video on CNN.com. Send email. Participate in newsgroups. Subscribe to RSS feeds. Read a web page. Watch a DVD. Listen to music. Use a calculator. Instant message with friends. Play some games. And much more.

Most importantly I want these scenarios to be enabled out of the box.

I also want users to always have the latest versions of these apps. Why? Because we'll add features. Fix bugs. Improve experiences and performance. And, add new scenarios.

But, Robert wants us to sell him a box that won't keep itself up to date.

In talking to "average users" this is NOT what normal people want. They want us to make their experiences better. They want a system that changes over time. I do know, however, that many of the people who read my weblog do not want that kind of vision for the future.

A couple of other things.

1) Robert brings out the usual "the black helicopters are coming for me" argument. He thinks Microsoft wants to watch what music he downloads and what apps he loads and more.

I've talked with quite a few execs and employees about this. They give me two arguments.

First is a simple scale argument: how many users will use Longhorn? Let's say that over the next decade a billion people will buy Longhorn in some version. That's not an unrealistic goal. Now, let's just do some thinking of what it means to have a billion machines reporting back to Microsoft. Do you have any idea of what kind of server farm would be needed to keep track of that many machines? How much bandwidth?

Second is a simple business argument. If we watched that much data coming off of your computer, it would be INSTANTLY known. Come on, our users are smarter than that. And certainly the Linux or Apple folks would LOVE it if we actually did that kind of intrusive watching. The press would love it too.

Third is a diversity argument. See, here you assume that every single one of our 55,000 employees is evil if you think we can put a massive scale watching feature into our apps like that. Microsoft simply does not work like that. We have a very diverse set of employees here. Some of whom, gasp, LOVE Linux. Yes, I've met several. Some of whom, gasp, LOVE the Macintosh. Just visit our Macintosh business unit.

Do you think that we'd all keep quiet if we added evil features like those? If you do, you just don't understand how Microsoft works. I have an open invitation to anyone who thinks this company is evil on that scale to come up and see me. I'll get you a tour of our server farms. You can see just how hard it is to data mine the simple crash data we get from apps that crash. And then you can understand the complete insanity of imagining that Bill Gates (or even Robert Scoble) is gonna look at your data of what apps you installed on your machine.

It's funny, though. Even I have these fears. Which is why I keep asking the questions.

The other thing he brings up is Trustworthy computing and the whole DRM thing. That'll be argued out for the next three years. I look at it this way: a box cutter can be used for good (opening boxes) and it can also be used for evil (9/11). Same for DRM.

Finally, where don't either Robert or I get what we really want from Microsoft? Because Microsoft has a VERY diverse set of customers. When you are building commercial software, you gotta build software for people who pay the bills and who buy it. What enterprises want in an operating system often is at odds with what Robert or I want in an operating system. Who buys more copies of Windows? Procter and Gamble or Robert Heverly? General Motors or Robert Heverly? In fact, Apple Computer buys more copies of Windows than Robert Heverly does.

It's a tough job designing software (or really any product).

For another view on this problem. I'm reading "The Second Coming of Steve Jobs" by Alan Deutschman. (Thanks to Buzz Bruggeman for sending me that as a Christmas Gift) and the author noted that Steve Jobs played a key role in the design of the iMac. The author also pointed out that the likelihood of Steve Jobs wanting to use an iMac in his own home was between none and nil.

In other words, when designing products, we need to pay attention to the customers. That's often different than what the geeks want.

Whew, that all said, I think Robert will get a lot of what he wants in Longhorn. My main frustration is that it's a few years away before I can really explain how.

comments < 8:44:57 AM        >