A discussion of hybrid open source and proprietary licensing models.

[Marshall Van Alstyne <marshall AT MIT.EDU>]

At 05:43 PM 8/22/2004, Rob Myers wrote:
> I'm writing this on MacOSX, not Linux. Linux's GUI options are about as 
> good as Windows 3.1 at the moment. But there's Linux software I can't get, 
> or get reliably, on MacOSX, or Windows for that matter. And I really 
> should have done a recent project in Struts or PHP rather than WebObjects. 
> The proprietary system that was once world-beating has not had the same 
> evolutionary capability as Free Software.

Yes, a good friend of mine uses MacOSX for the same reasons.  ... and so 
does my wife (although she doesn't do programming :)

> > If we accept that such users exist, can we understand what reasoning 
> > would make their choices rational?  I'm not a big fan of the "this or 
> > that party needs education" school of thought because intelligent people, 
> > when presented with the same facts, frequently disagree.  If we can 
> > understand why their choices make sense, we're more likely to find 
> > grounds for mutual gain.
>
> I accept that we need to understand the context in which conclusions are 
> reached. I am interested in this.

For me, at least, candid discussion almost never fails to teach me something.


> > So, is there any reason why paying for software makes sense?
>
> Patronage (getting something made that wouldn't be otherwise).
> Direction (making sure what's made includes one's needs).
> Continuity (making sure something one relies on continues to remain current).

Nice.  These are also good because the represent choices of an informed 
consumer, not simply someone responding to marketing and branding as a way 
to cope with uncertainty.

How about also:

Time (needing it soon and someone else is a better programmer)
Opportunity (as a lead user experiencing a problem, you sense a market 
niche and want to capture it)


> You *always* pay for software. Even if you only give your ISP a few 
> pennies for download time or buy a friend a drink for a CD, you pay. Then 
> we get to whether that payment reaches the authors.

Excellent, couldn't agree more ... but now you're thinking like an economist!


> > The SUV example, however, doesn't seem to hold up.  As a rival good, it 
> > can't be perfectly copied so the person with physical possession is the 
> > only one who gets to sell it.
>
> How perfect a copy is isn't always important. In conventional wisdom, 
> pirate copies are often very low quality (I mean analogue media and 
> physical products), yet they impact on sales of the originals.
>
> Whoever came up with the idea of an SUV has seen it copied and has to 
> compete against the copies.

Interesting.  But, the folks who are capable of copying an SUV are 
relatively few in number.

So let's take an idea put forward in the recording industry debates.  Is 
the pirate copy a complement or a substitute?  If the copy is sufficiently 
imperfect, it's more of an appetizer, a sample of the main course that 
leaves you hungry for more.  This would be a complement.

But if the copy is perfect, as it could be for most information content, 
then it could sate an appetite and limit sales.  This would be a substitute.

So the real implication would be a measure of complementarity or 
substitutability perhaps in turn measured by precision and completeness of 
the copy...


> > My point is different.  Suppose we were to make all intellectual property 
> > suddenly free in the liberty sense -- whatever it is, you have the right 
> > to take it, modify it, use it, and redistribute it -- and so does 
> > everyone else.  Then it necessarily also becomes free in the pricing 
> > sense.  You could never charge expressly for the value of a new idea per se.
>
> IIRC the patent system doesn't support the value of ideas per se: it 
> supports the value of realising that idea. This is a useful distinction 
> for FOSS and Open Content, as having the code, or even the binary, isn't 
> the same as having the software running to your satisfaction.

Agreed, this is a really thorny problem and a great observation.  In fact, 
the value of information generally is extremely hard to assess before 
someone has actually tried to use it in the marketplace.  So, you're right, 
patent systems reward the value of realizing the idea not  some arbitrary 
notion of value before it's been tried.

There's another critical observation here from von Hayek.  The "market" 
broadly construed as a means of allocating resources does so in a hugely 
decentralized fashion based on information contained in prices.  Each 
person, knowing his or her own needs and wants, can decide what to 
consume.  In so doing, scarce resources are automatically conserved 
(because they're expensive) while abundant resources are automatically used 
(because they're cheap).  Systems that destroy price signals can lead to 
the misallocation of resources, as seen in many centrally planned economies.

What this means for open content is that we may not want to destroy price 
signals entirely...


> > If this were to happen, could Pfizer afford to invest in developing 
> > Lipitor or Glaxo invest in antiretrovirals? Could Intel invest billions 
> > in next generation chips?  Society loses if these investments in forward 
> > innovation don't happen.
>
> However society may lose with the (coarse) granularity and (profit-led) 
> direction of research undertaken by such large organisations and with 
> their need to recoup such large costs. *Failed* corporate innovation can 
> also have a negative value for societies.




> > The problem seems one of balance.  What's the best way to reward 
> > innovation?  What's the best way diffusion innovation once it's 
> > occurred?   The answers to these questions seem frequently in tension.
>
> There's also the question of how relevant innovation is as a primary 
> criterion:
>
> http://www.lessig.org/blog/archives/002106.shtml

Really interesting post.  I'll have to think more about the connectedness 
possible via sharing (perhaps RMS' "freedom to share") as a distinct source 
of value apart from innovation per se.


> IMHO there are very few companies who genuinely innovate and who aren't 
> just commodifying other people's ideas.

It's why information reuse is so important!

> > Actually, there are other ways to make forks seem economically 
> > unattractive and these forces need to be understood in the marketplace.
> >
> > Here are three articles using information economics that show how 
> > packaging can make life tough for competing products.  One is mine with a 
> > colleague, the others are by Bakos & BrynjolfssonNalebuff, and by Nalebuff:
> >
> > 1) on how free information can be profit maximizing and create barriers 
> > to entry for other products:
> >     http://papers.ssrn.com/sol3/papers.cfm?abstract_id=249585
> > 2) on how bundling can be profit maximizing
> >     http://papers.ssrn.com/sol3/papers.cfm?abstract_id=231598
> > 3) on how bundling creates barriers to entry for other products
> >     http://papers.ssrn.com/sol3/papers.cfm?abstract_id=185193
>
> Thank you for these references, I'll take a look.

All are technical academic papers -- sorry for that.  They all also need to 
be expressed in straightforward language.


> > All hail the GPL!  Is there anything it can't do? :)
>
> It doesn't handle patents, and it refuses to make my tea. :-)

I can see that it would refuse the former on moral grounds but the 
latter?  I think you just have to ask it nicely, perhaps using the 
appropriate tea .c makefile :)


> > > Under BSD, the *future* of the commons is not ensured. Code goes
> > > stale, technology moves on. If code gets taken from a BSD project and
> > > improved but not released by a proprietary project, the tragedy of the
> > > commons has occurred in all but name. That does not happen under the
> > > GPL. This is an important point quite separate from zero marginals.
> >
> > Yes, this is like the forking problem earlier.  Here's why expiration of 
> > the proprietary period makes sense.
>
> Eventual release may be better than non-release.
> Unless it is gamed to affect competition.
> Or it is considered in terms of efficiency.

How do you mean gamed to affect competition and considered in terms of 
efficiency?  I'm curious.

MVA