How Microsoft Lost the API War | The Coffeehouse

jkwuc89

Excellent read. Check it out at:
http://www.joelonsoftware.com/articles/APIWar.html

Akaina
C9 is cool, but /. has better karma.

You know, that article was so good I wanted to give him a high five. If I was more cynical I'd say he read my earlier posts on the subject.

I've been preaching this same message on C9 and I'm glad it got some large exposure:

http://channel9.msdn.com/ShowPost.aspx?PostID=5901#5901

Joel Spolsky is right on the money.

lars

Indeed. It's brilliant!

/Lars.

object88
amplify.

Wow, that's a bit of an eye-opener.

jamie

"Promising new technologies like HTAs and DHTML were stopped in their tracks. The Internet Explorer team seems to have disappeared; they have been completely missing in action for several years. There's no way Microsoft is going to allow DHTML to get any better than it already is: it's just too dangerous to their core business, the rich client. The big meme at Microsoft these days is: "Microsoft is betting the company on the rich client.""

= no more full screen except on Windows client local

Jeremy W
that blogging guy

After about a day of thinking this through, I think I've realised what the problem is.

Fundamentally what Joel is observing is more the 'death of the desktop' than the death of Windows / MS / etc.

I've blogged about this throughout the day, and it's been fun to see the reaction. Lots of people are picking up on this, but I think far too many are grabbing it from the wrong angle.

If this is an anti-desktop argument, then Linux desktops, OS2, OSX and Netware are dead as well, for most of the same core reasons.

But, really, the death of the desktop is decades away, if it ever happens. As such, I don't really see why this article is kicking up such a big stink. It's not enough to make any decisions or do any forward-thinking on...

... All of that said it is a very, very good article and makes you think a lot. It has a lot of great points. It just took me all day to realise what the problem with the conclusion was.

object88
amplify.

Jeremy W. wrote:
Fundamentally what Joel is observing is more the 'death of the desktop' than the death of Windows / MS / etc.

Interesting take. But the death of the desktop is the death of Microsoft if, in fact, the majority of Microsoft's revenue comes from the OS (more home OSs sold than server, right?) and Office. If the majority of apps go "online", then significantly fewer copies of Office are sold, and fewer OS upgrades are sold. Even the servers sold might get knocked down a few posts, if ASP.NET apps can run on other platforms.

So why isn't this the death of Linux and Apple as well? I would postulate because a) Linux is "free" and doesn't depend on revenue as much as Microsoft, and b) because a large part of Apple's marketshare comes from graphics and other apps not well suited for "online" apps.

OTOH, Windows is the primary platform for PC gamers, right? But then, you have competition from PlayStation, GameCube, and even their own X-Box.

(Now, I've written the above some time after reading the article, so I've likely got a good bit wrong. Please feel free to correct me.)

Now, as Joel points out, this is an exageration (sp?), and won't really be the death of Microsoft. They will surely find a way to survive.

Jeremy W
that blogging guy

The problem with the death of the desktop, though, is that now you're talking 10, 20, 50 years in the future. At a time when none of us can predict if Linux / Microsoft will even be around. Also, there is a huge debate about whether the web as it stands now is in any way the best way to deliver the large majority of applications.

Web-enabled thin front ends is one thing. Web apps period is another entirely, and is much farther away.

AndyC

jamie wrote:

= no more full screen except on Windows client local

Er, their is more to life than full screen IE you know.

AndyC

object88 wrote:

Interesting take. But the death of the desktop is the death of Microsoft if, in fact, the majority of Microsoft's revenue comes from the OS (more home OSs sold than server, right?) and Office. If the majority of apps go "online", then significantly fewer copies of Office are sold, and fewer OS upgrades are sold. Even the servers sold might get knocked down a few posts, if ASP.NET apps can run on other platforms.

I'm not sure I follow the logic there. Office is far and away the big revenue generator at MSFT surely? I was under the impression that the majority of client OS's were OEM and provided a small share of overall revenue.

So what would happen if Office became a web page tomorrow? Presumably there would be some sort of subscription charge for using it and that wouldn't be a million miles away from the current Licensing 6 model, would it?

Plus you'll still need some sort of OS to access the web, even if it was something closer to Windows Embedded than XP. Which leaves plenty of scope for some "improvments" to web standards to make sure Office still runs better on an MSFT platform.

Embrace and extend as the old saying goes. Wink

AndyC

object88
amplify.

AndyC wrote:

So what would happen if Office became a web page tomorrow?

Plus you'll still need some sort of OS to access the web, even if it was something closer to Windows Embedded than XP. Which leaves plenty of scope for some "improvments" to web standards to make sure Office still runs better on an MSFT platform.

Well, if Office goes online, then it no longer uses the coveted Windows API-- hence, the Death Of The API, as the article proclaimed.

And we already have OSs to access the web. Any improvements to web standards can be handled up upgrading the browser, rather than the OS, which leads back to the death of the Windows API.

Jeremy W
that blogging guy

Except that Joel was also knocking on .NET. The article's scope is greater than the title, and Joel's even clarified in his follow-on post that the title was meant to be inflammatory and that he really didn't expect Windows to die anytime soon.

Joel's a smart guy. This came out of left field to me, but now that I realise that he's not actually saying to do anything, and that this isn't about now ... Well, the article doesnt' bother me so much.

It's still well worth talking about though Wink

bitmask

object88 wrote:

Now, as Joel points out, this is an exageration (sp?), and won't really be the death of Microsoft. They will surely find a way to survive.

Exactly. In the last decade or so, the following entities and events have been predicted as potential murderers:

The Network Computer
The Internet
Java
Netscape
Linux
Star Office
The Justice Department
The European Union
Google
The State Of Massachusetts
The City Of Munich
Rampant viruses
OS/2
Microsoft Bob
Apple
Novell
Palm
IBM
Time Warner
The Register

Scott McNealy (with his bare hands)

Larry Ellison (with a yacht)

Col. Mustard (in the Library, with a candlestick)

I'm sure I've missed a few. Of course, some of these are still competition, and competition is good. So is deprecating a 10 year old API. I mean, geez, maybe Microsoft should never have moved us beyond programming with INT 21h. Everything would still be backward compatible, and still suck.

I don't know who these people are who Joel talks to, the people who "don't seem to care about the little UI glitches and slowness of web interfaces" because the glitches bother me quite a bit. In fact, I don't know what I'm doing in here right this very moment. I'm going back to Outlook.

jsrfc58

This is another one of those articles that I wanted to sit and think about for a day before reacting to...Joel brings up some excellent points, but in retrospect, I think it is often a waste of time to try and "predict" things like the "death of the desktop" or that web apps are going to replace products like Office. A few years back there were predictions that "Internet appliance-PCs" were going to replace desktop PCs as we knew it. That sounded like a throwback to the "dumb terminal" days (without the green screen) and I don't know of a lot of people that bought into it. To say the war is "lost" is overstated, but I think a secondary purpose of this article was to drum up business on his web site. It obviously worked.

JParrish

I read this article and discussed it with a good friend of mine to compare points. One thing that bothered me was his back and forth point/counterpoints. It seemed less like he had a point than a philosophical expose.

As for the web being something that is driving desktop developers to abandon their desktop roots (win api), to me this is a dead horse, and we should stop beating it. It is true that a lot of developers jumped on the web bandwagon, but MS is addressing the benefits that web development brought, namely deployment and re-deployment in their single click install for windows forms.

I think multi platform will still mean ASP.NET for MS developers, but for intranet applications, and additionally commercial windows desktop applications, the deployment difficulties have largely been solved, and whidbey is bringing with it even more work addressing this subject. Just my opinion Smiley

lars

I'd say the API lock-in is dead. You can do alot of things as web apps. Or try to run .NET applications on non-microsoft server. Or (sometimes) even run Win32 apps on Linux with Wine. Simply put there are more options today. In that respect Joel is right.

But that does not mean that one option will rule them all. People have been trying very hard to replace the desktop for 10 years. I've spent 5 years building these kinds of application for customers. Some things work just as well as web apps. Some don't. In a sense it's been 10 lost years in terms of usability. There was a time when people accepted poor usability because webapps were a cool thing. Not anymore. The most spectacular failure was Sun's thin-client JavaStation. Both ActiveX and Java Applets failed to bring the real power to the web platform. It looks like Flash is the best contender so far. ASP.Net have also made it alot easier to build good interfaces with less effort.

Microsoft is the one killing Win32. With .NET. And instead of locking people in with the old API they present such a great new API (.NET) that people want to use it. Win32 is getting old and Microsoft have done the smart thing: replace it before someone else does.

I have not read that much of Joels' previous articles. But I already respect him very much. He is a very smart guy. And he sure know how to put it in writing.

/Lars.

eagle

WINFX

Akaina
C9 is cool, but /. has better karma.

I don't think Joel is pointing at web apps so much as he's pointing out the loss of long-time loyalty to Windows app development.

His point about Windows programmers requiring $130k in NY vs $80k for Java speaks volumes.

Let's assume that for the forseable future OpenOffice, MSOffice, and whoever else demand a local installation. Office loses its appeal just by not being free. So they have to justify this somehow - by taking full advantage of what the competition lacks - namely conferencing capabilities (because of poor multimedia driver support in *nix).

If they don't have die-hard developer support no one will really care what they start touting for their next version because the side with all the die hard coders will have tons of new features too. The bottom line is that Office is running scared and OSS is just strolling right along behind them.

NOW enter the death of their API and programmer loyalty, and Microsoft is in a world of hurt. That's without taking into account that .NET might one day run seemlessly on Linux via Mono or DotGNU.

On a side note, about the only thing you can't (and I say that in the Mission To Mars sense of the word) migrate to the web is a programming interface. The file sensitivity is a question, and there's no way in hell people are going to agree to a micropayment per compile. 9 times out of 10 people will just chose a free tool - with another API, which people seem to be doing ANYWAY.

With respect to his title "How Microsoft Lost the API War" I think his point is that no one fully realiezes the impact this will have.

Joel: ~5~

object88
amplify.

Akaina wrote:
On a side note, about the only thing you can't (and I say that in the Mission To Mars sense of the word) migrate to the web is a programming interface.

Not to question your points, but what of SOAP and WebServices? Aren't those (slow) web-based programming interfaces?

interscape

I talked about my comments here:

http://www.longhornblogs.com/robert/archive/2004/06/18/3731.aspx

Akaina
C9 is cool, but /. has better karma.

If Longhorn can still run a 20 year old program, it's pretty safe to say that Microsoft is going the extra mile to keep people as unwilling to change as humanly possible.

So long as Windows can run Java, and Linux can run .NET