How Many Applications Depend On Windows?

msnomer writes: "The Cato Institute is going to publish a report this week disputing the claim that there are 70,000 Microsoft applications, according to a NY Times article. The author, economist Richard B. MacKenzie, says that the "70,000 figure might actually represent the number of applications that have been written during the entire history of the personal computer industry." The figure, however absurd, pits Microsoft marketing against the legal department, since the purported number of applications--a number not disputed by Microsoft, by the way--is a key factor in the decision to break up the company. Anyone else astonished that Microsoft marketing may have exaggerated a bit?"

Alternatively, is anyone astonished that a number so arbitrary (in either direction) is considered a serious point of evidence? Not that there's a great way to weight the importance of programs in isolation from each other, but [random shareware X, even if business related] doesn't match the importance of, say, a major word processor in the real world. Isn't the flexibility that a given OS offers to create new programs, and the rate of change in the number of available programs, more important than the existing number anyhow?

Updated 20:05 by timothy: PJ Doland, Webmaster of the Cato Institute writes: "The article on Microsoft that you mention is now online at http://www.cato.org/pubs/pas/pa-380es.html. I'm a slashdot addict. When I saw your link to the Institute's site, I convinced the higher-ups to let me post the report early for your readers. The full report is in PDF."

Comments

The report blows.

[brokeninside]

Here's my summary of some of the points from the report:

Point One: Microsoft doesn't act like a Monopoly.

Point Two: The definition of the Market that Judge Jackson used was overly narrow because Sun, Apple, and Redhat don't make operating systems for non-networked intel compatible PC's.

Point Three: There are no 'staying costs' for sticking with Microsoft Windows.

Point Four: Market-share does not define a monopoly in the "new economy"

Don't take my word for it, read it for yourselves, its very poorly reasoned and uses the same types of arguments that it claims are bad when Judge Jackson used them.

How many bugs per app....

[spankenstein]

Remeber when Windows 2000 came out? There were 65,000 documented bugs. This is eerily close to the number of "applications" that they say it has.... coincidence?

It's a hit, just not a home run..

[Thomas Charron]

The article does make several points that are indeed valid, in that the judge made a big deal about some thing he didn't have to, and didn;t about some thing he should have..

The first thing addressed, or not, in the above article is that, yes, indeed, the judge *IS* making up a new meaning for a monopoly. In the past *NOTHING* of any value could simply be copied billions and billions of times at little cost. I can mass mail an application to a million people, at the total cost of 19.95. Heck, run a little banner ad proggy, and it's *FREE*. *NEVER* in *HISTORY* has a situation like this been present. No one could ever argue that Microsoft has not stopped people from developing alternate operating systems. They haven't. What they have done is ensure a closed and private 'public' standard.

Think about someone having a patent on the internal cumbustion engine. I'm sure someone had a patent at one point in time that applied. Now, along comes the car. Suddenly, you're forced to only buy the patents owners gas. And by the way, you'll need to buy a new engine for your car every few years, so you can run the new gas. Certainly, in no way has this patent holder stopped anyone from coming up with a new means of power. But if they can manage to get enough people using it, you *DO* have a monopoly.

Personally, I do not like Microsoft. I don't like their company, or how it treats its customers. Sure, they have some good apps. And yes, they've done some damned good buisness deals, and been sucessfull as hell. But at some point, it has to stop. At some point, it has to be said that you've simply made *to* much money, and you've locked in to good of a nestbed.

Anything that ends up being used widely in the public needs to be open eventually. It's in the general good of the public. In the past, things moved slower. No one could lock the public in, becouse of the length of time it would take. Microsoft locked over 90% of the OS market into place within a few short years, becouse of the success of the PC.

Whoever came up with the tarred road had a damned good idea. What if he was still recieving money for the idea? What if it cost the government millions of dollars, paying for the use of the idea? It's certainly in the good of the public to have these roads. It's also in the good of the public to have a well defined, standardized computing environment.

I'm blathering, I'll shadap and read some other opinions now..

Re:70,000 Breakdown

[ackthpt]

I added it up with the Windows calculator ;-)

Vote Naked 2000

Richard B. MacKenzie is on crack...

[Rombuu]

70,000 figure might actually represent the number of applications that have been written during the entire history of the personal computer industry.

Oh, come on. This is what happens when you have ivory tower academics weighing in on things. I've been with companies that have serveral thousand of their own internal applications... entire history of the industry indeed.

Hell, there have probably been around 70,000 bad first person shooters in the past few years....

Re:Richard B. MacKenzie is on crack...

[GlenRaphael]

70,000 figure might actually represent the number of applications that have been written during the entire history of the personal computer industry.

We should probably wait to see the actual Cato study before lambasting Mr. MacKenzie too much based on what a New York Times reporter heard. But I agree that sentence sounds ridiculous unless he's using the word "application" in some highly restrictive sense. Look at the big shareware repositories. Jumbo.com has "over 300,000 shareware and freeware programs"; Shareware.com has "over 250,000 shareware files". Unfortunately those services don't return more than 500 results per query, but that would be a good place to start. Removing duplicates from a service of this sort would get you a list of applications that run on relatively recent hardware and were popular enough to be worth indexing, but it'd still be a small fraction of applications ever written because it wouldn't include most commercial apps, any apps for defunct platforms like the Commodore 64, apps written in-house, and apps that are or were only available in more limited distribution..

I don't know about this

[Anonymous Coward]

People didn't believe Cato when he testified at the OJ trial, and they're really not going to believe him now. This is a bit out of his league.

You'd think *programmers* would understand this

[FascDot Killed My Pr]

What question are we trying to answer?

How many applications depend on Windows?
How many applications run on Windows?
How many applications Windows users can use?

The first two are both mentioned in the Slashdot summary, but the three are largely unrelated. Consider programs that run on multiple platforms (via porting), consider Java, consider mainframe apps with customized (or telnet) front ends. I'm sure MS Marketing and MS Legal are both exploiting the vagueness of English to pull the wool over our eyes. Don't let's help them with misleading (and even contradictory!) news stories.
--

Define "Application"

[ChaoticCoyote]

Is an "application" defined as "MS Office" or "MS Word" or just any ol' executable?Does a complex DLL or COM component qualify as an "applicatiom?" What about a VBScript to manipulate an Access database?

I suspect Microsoft's "70,000" can easily be justified or villified, based on your definition of application.

It's all a matter of complexity and perspective. Like most things in life.

I tend to think of an "application" as a collection of programs and libraries that work together in to provide a solution. I wrote a specialized encryption program today, in a few hundred lines of console-based Visual C++; I don't consider this to be an application. In the last decade, I've created a couple of hundred Windows-based EXE/DLL/OCX files, but I've written perhaps a dozen "applications."

Considering the number of people who code for Windows, however, I wouldn't be at all surprised if there are 70,000 (or more) pieces of executable code for it. Hell, my Windows NT 4.0SP6 box has 3500 EXEs and DLLs alone, and I'm sure that doesn't even scratch the surface...

The URL for the Report

[pjdoland]

I'm the Cato Institute's webmaster (as well as a Slashdot reader). I convinced the people upstairs to let me post the report a early because of Slashdot. It's at http://www.cato.org/pubs/pas/pa-380es.html I hope this will clear up some questions.

Your karma recipe for today

[Enoch Root]

Write a post containing at least two of the following:

That number is way too high

That number is way too low

Are they counting all the virii/Hello Worlds/shareware programs/versions of Windows?

Kato is on crack

I coded more applications on my weekend

Linux r00lz

Any obligatory joke about Micro$oft

You're welcome.