The Simple Economics of Open Source

Gordon McMillan gmcm at
Thu Apr 27 23:57:08 EDT 2000

Raffael Cavallaro wrote:

[I wrote]
> > The closed source world is based on a diseased model
> > (pretending software is a product).
> This "diseased model" has made billions for a number of firms,
> so, from an economic standpoint, your argument rings rather
> hollow, especially when compared to numerous open source ventures
> which have yet to post a cent of profit.

So has bashing people over the head and taking all their 
money. And many proprietary ventures have yet to make a 
cent of profit (like all those dot coms with ridiculous stock 

> > Scarcity of an intellectual secret may make a software
> > package containing that secret more valuable, but scarcity of a
> > software package is likely to make it less valuable (since
> > support will be hard to find, and upgrades unlikely). OTOH it's
> > scarcity of a piece of artwork that makes it valuable.
> You're conflating separate markets. Unique artworks (paintings,
> for example) have no real equivalent in the software world,
> because software sales are based on essentially unlimited
> identical copies.

'Twas your analogy.
> My example referred to IP in the art world - secret working
> methods that allowed artists to produce works their
> contemporaries could not - methods of bronze casting, for
> example, or formulas for oil painting media. These IPs are like
> closed source in that, without them, competitors are shut out of
> a lucrative market.

There's very little of that in code. The question "how do they 
do that?" is rarely unanswered for longer than a few months.
> My point was, and is, that it makes no economic sense to open
> source an IP that could be used to create a monopoly. MS
> understands this, and works ceaslessly to create such monopolies.
> Note that this has *nothing whatsoever* to do with what is best
> for consumers. 

You really think MS's "monopoly" is based on IP?

> I think part of the open source community's problem is this naive
> belief that just because a certain model (open source) leads to
> products that benefit consumers more, markets will allow
> consumers to choose this production model. This belief is naive
> because it fails to consider that the producers' interests are in
> no way congruent with the consumer's.

An indicator that the model is diseased.

> The producer has every
> reason to go closed source and attempt to secure a quick
> monopoly. We live in this world now, and there's no reason to
> believe that it's going to go away any time soon. I think open
> source advocates grossly underestimate the power of monopolies,
> and the unwillingness of consumers to break ranks even though it
> *may* provide long term benefits, because breaking ranks entails
> large short term costs, especially in network utility - i.e., the
> loss of interoperability with the rest of the monopolist's
> customers.

They are not "monopolies". They are large companies with a 
lot of money, salesman, distribution channels, lawyers and 
programmers. The competition has quite a few programmer, 
and just started to get a little money.

If instead of analyzing profits you looked at market share, in 
the area of web servers at least, you would have to draw the 
opposite conclusion.
> BTW, all software is *not* support. This is another open source
> community bias which comes from living in the server room. Lots
> of software is *client* software, which requires little or no
> support from the vendor. 

And client software is often free. Where software has client 
and server, the money's all made from the server. If the 
vendor's smart, they'll open up the client side so they won't 
have to waste time developing it.

> Instead, these support services are
> provided by full time IT people, who may be certified by the
> vendor, but who are paid by the enterprise. In other words, MS
> makes their money on enterprise licences for Office, not for
> troubleshooting the enterprise's problems with Office.

Bzzt. A 10,000 seat license complains that X is broken, X will 
get fixed. (And when enterprise learns that Y requires only 
50% the support costs, they will switch.)

>  For the
> most lucrative software businesses, software *is* a product, not
> a service.

No. The only thing they share is that they both require 
investment, as does almost anything else.

Congress does not understand software. Lawyers and judges 
do not. Most businesses do not. Most people do not. The 
world has never before encountered anything with a 
vanishingly small marginal cost of production, but something 
that has so much "value".

Software companies adopted the "product" model. The fact 
that some of them make money does not make the model 
sane. It just means software can be lucrative.

In fact, let's compare Apple and MS. Apple could be where 
MS is now, but Steve Jobs repeatedly insisted on shutting out 
the competition, keeping developers ignorant and making 
them pay. MS, believe it or not, gives away a great deal. 
Anybody can browse all the MSDN docs. If you buy MSVC, 
you have the source to the c runtime library. MS publishes 
large amounts of specs. Yes, they usually withhold something 
- about a six month to one year lead on the market. IP in this 
market has a very short shelf life.

They're commanding position has nothing to do with IP. Most 
of the maybe-almost-original things they've done (ODBC and 
COM) are pretty thoroughly documented. Their success is due 
to business practices that are common in, say, the food 
industry or the mega-store segment, but are (rightfully) 
considered evil in the software market.

The open source model(s) will almost certainly require 
adjustment - they are mostly a reflection of the fact that the 
current model is warped. But the big boys are dinosaurs. They 
will change their models, or die (Bill won't make it, but neither 
will Scott or Larry).

But before things become sane, people will have to figure out 
a "sale" is; and we're nowhere close to that :-(.

- Gordon

More information about the Python-list mailing list