Slashdot Mirror
Joel On The Economics of Open Source
Stephen writes "The ever-incisive Joel Spolsky discusses the economics of open source software in his latest Joel on Software column. Why do so many large companies want to develop open source software? It's not because they have suddenly converted to Stallmanism."
Simple. With no web browsers out there, there wouldn't be much demand for web servers, would there? In this case the strategy is not to grab market share from the competition but, in the words of Dubya, to "make the pie higher!"
Let's look as his examples:
Netscape is trying to commoditize the browser market .. in order to dominate the server market. This would have been plausible in, say, 1997. I find it amazing that he tries to push this by anybody--the browser was commoditized.. and servers turned out to be irrelevant! Where is netscape now?
IBM is investing in open source software to bolster its consulting services ... -- wait a minute. IBM's fortune was made in the early 50s by being the king or proprietary--you couldn't even buy their computers--you had to lease them! The US government eventually stopped this, but IBM's greatest period of success in the computer age was when it had a complete monopoly on sales and service of its own, very closed product lines. With the IBM 360 series, IBM saw some erosion of this due to "plug compatible" peripherals produced elsewhere. With the IBM PC (btw.. the author's description of IBM's "success" in commoditizing the PC makes NO sense whatsoever), IBM did poorer still--we all know how badly they did.
But let's look at the specifics--IBM is a BIG company. Let's say (hypothetically) it could put its full weight behind OSS and therefore contribute a whopping 3% to the total corpus of reasonable OSS stuff. Suddenly, it has what--spent a lot of money for the benefit of all while increasing what it can personally consult on by a whopping 3%. Even if there are network, learning, or syndicate effects, this situation screams "free rider problem."
Ditto for Transmeta..
It's almost ironic that the author pics such dead or dying companies like Netscape, Transmeta, IBM, etc for his examples.. Look, I like these companies as much as anybody for their past, but let's face it..
I could go on, but this article is a big swing and miss.
This is why most OS projects are done as a hobby, not as a job. You give back to the community on your own time, but still put food on the table.
Good quote, too many chars. Seriously, the slashdot 120 char limit sucks!
no, it makes perfect sense, if you read it. he's describing how IBM published the specs to the interfaces so that 3rd party vendors could create plug-in cards. with cards, PCs can do more, making them more valuable in more situations, causing demand for them to increase.
-c
I have discovered a truly remarkable proof which this margin is too small to contain.
A lot of good points, but the Cathedral/Bazaar point is still a good one.
The argument here seems to be people make free-as-in-beer software because its cheap. But they may also do it because it produces better software (therefore reducing the TOC for the other products).
These two things are not necessarily in conflict.
Frankly, I also think that a number of arguments used are pure Aunt Sallys. Has anyone ever really said IBM have converted to communism? If so, which mental institution were they speaking from at the time?
I disagree. I've been using Unix since I started University, in 1991. When I have a problem with non-programming tasks it's typically a MS problem and not a Unix one (well, for HP, Sun boxen anyway).
The openness of the systems (even for non-Open systems like Solaris) makes them easy to maintain. All Unices behaves mostly alike, usually trivial to bring them to single user, fsck and reboot for example.
There are plenty of capable Unix admins, and plenty of resources for said admins - usually lab shelves are coming down with O'Reilly books, the web has plenty and if Usenet archives on Google can't solve your problem, well...
I'd argue, based PURELY on my current job experience, that the TCO of PCs is higher. We were a Unix based design lab, now we're PC based with Unix server farms. I've more calls on Support now than ever as I can't fix anything myself.
One payment you received: experiance
Another: Your contributions as well as many others have permitted FREE (beer) software to develop which costs you nothing to get.
And yet another: Friends -- the people who download your software may not be able to pay you but one day they may help you get a job by being a reference.
Linux got his job at transmeta because of what he did. Imagine if he charged for Linux from the start...
So yes there are payment methods other than $$$...
DRM? No thanks, I'll just get it somewhere else...
The basic idea of the article is that if you can make the total cost of entry for some product lower by reducing the cost of one of the product's components, you can charge more for the components that are left. If you're smart, you get the price down for these compliments that you don't control so you can up the price of the services you do.
So if PC hardware is cheap, more people can afford the price of entry and you can charge more for the OS (eg, Windows). If enterprise OSs and software are cheap, you can charge more for your consulting services (eg, IBM).
Why is Mozilla "cheap"?
[Given that IE is free, what is the incentive for Netscape to make the browser "even cheaper"? It's a preemptive move. They need to prevent Microsoft getting a complete monopoly in web browsers, even free web browsers, because that would theoretically give Microsoft an opportunity to increase the cost of web browsing in other ways -- say, by increasing the price of Windows.]
Java does exactly what Sun *didn't* want:
[If you can run your software anywhere, that makes hardware more of a commodity. As hardware prices go down, the market expands, driving more demand for software (and leaving customers with extra money to spend on software which can now be more expensive.)
Sun's enthusiasm for WORA is, um, strange, because Sun is a hardware company. Making hardware a commodity is the last thing they want to do.
Oooooooooooooooooooooops!]
It's all 0s and 1s. Or it's not.
There's this zen koan:
A lord asked Takuan, a Zen teacher, to suggest how he might pass the time. He felt his days very long attending his office and sitting stiffly to receive the homage of others. Takuan wrote eight Chinese characters and gave them to the man:
Not twice this day
Inch time foot gem.
The day in which you coded that software you gave away for free will not come again. A small bit of your time is more valuable than the largest diamond. It's limited and you can never buy more. Never put a price on your time. It cheapens it.
(BTW, if anyone knows exactly which characters Takuan wrote down, I'd be eternally grateful if you told/showed me, email is jcsehak.at.yahoo.com)
c-hack.com |
What I most like of this theory is that hardware is a commodity today. If open source can turn software into a commodity, the real value will be in the people putting systems together (as the IBM example shows).
Most of the slashdot crowd are technical heads so I would say that it is in the best interest of most of us to get GPL'd stuff working, with the possible exception of packaged closed software developers, about 5% of all developers.
This way the money will go to us, instead of CEOs or marketing departments.
When his defense asked, "Which computer has Jon Johansen trespassed upon?" the answer was: "His own."
Joel holds forth the position that since Sun sells processors, it's against their interest for them to invest in Java, which allows people to write software for any processor, thus "commoditizing" Sun's product line.
A few paragraphs above that, he holds that is in Transmeta's interest to invest in software that runs on anything (Linux), because it commoditizes the "complement" of Transmeta's product (also processors).
Joel, why doesn't Linux and other Free Software commoditize Transmeta's processors as much as they do Sun's ? Why does linux make only make a "complement" a commodity in Transmeta's case, but Java makes the complement a commodity and the code product a commodity in Sun's case ?
The fact is, Joel got a little carried away. If the alternative is not having your product purchased no matter how good because of artificial constraints, then operating in a commodity market looks appealing. Both Sun and Transmeta believe they have superior products to Intel/AMD, even if those products are superior in a niche. They need the death of (or the splitting of) the Microsoft/Intel axis, so that they can compete on any grounds at all.
If everyone uses Linux, they may mostly use x86 chips most of the time but sometimes they'll need somthing really low power and low temp and buy Transmeta, and all their software will just re-compile no problem. If everyone write Java programs, they may use x86 chips most of the time but sometimes they'll need something really powerful and stable and reliable, and they'll buy a Solaris machine and just run their java code with no problem. (Of course they can also write Java and run it on Transmeta chips, and write for linux and run it on Sun hardware.)
So what is lacking in Joel's analysis is that therw are some things worse than competeing in a commodity market, and Free Software is sometimes funded to purposely give companies a chance at a commodity market.
Sun would alread be quite dead when they _didn't_ release Java to the developer community.
;-)) would use Unix/Solaris as their development environment.
:-)
Without Java and the ever increasing PC Power (CPU for a long time, I/O catching up slowly) only a small minority (same size as the zOS/CICS/Cobol developers
Given the cheap developer costs (PC with WinNT, Intellij and Oracle 8i), a large amount (if not most) people in the corparate (in house) software development are using Java.
Solaris is a good server OS, and everybody that has the money will prefer the high quality Sun hardware over 99,9% of the PC crap hardware, when it comes to 7/24.
I think Joel missed that (minor) point because he focus on his "shrinkwraped" business, where the rulese are sometimes quite different.
To sum it up: good article, Joel rules, but he won't win the nobel price for economics
Bye,
Jürgen