[Fsfe-uk] Re: Talk on Thu: The Economics of Open Source Software - Prospects, Pitfalls and Politics

[Wookey]

+++ MJ Ray [04-01-14 08:29 +0000]:
> On 2004-01-13 19:17:13 +0000 Martin Michlmayr <address@hidden> 
> forwarded:
> 
> >>Open Source Software does not represent a suitable alternative to the
> >>commercial software market from an economic point of view [...]
> 
> If this really is the speaker's view, I hope that someone attending 
> will point out that "Open Source" (or, more properly, Free, Libre and 
> Open Source) is not exclusive with "commercial". 

To be fair, he did understand that, and used GPL vs. Proprietary as his
basis for examination, although he did then spoil it a bit by explaining
that he meant 'Proprietary in the commercial sense'.

That chap got what can reasonably be described as a 'highly skeptical and
critical' audience, picking holes in his argument throughout the talk. With
some particularly fine points made by Ross Anderson and <erm not sure of his
name, mate of James McKenzie, been on the affs stand, he might be alex
hudson?>.

The full analysis (100 pages) is probably worth a read, and Dr Kooths
recommended we do that rather than rely on the presentation or the executive
summary. I can't find that yet though, but the presentation is here:
http://mice.uni-muenster.de/presentations/index.htm 

The study admits that customised FS development works fine as a development
model, but claims that packaged software does not because of a free-rider
problem making the software effectively zero-price (on questioning he said
that prop.software has a free-rider problem too, but that in this case it
puts the price up: this inconsistency is worth investigation).

There was an intersting discussion with the floor about how FS development
can be paid for in various ways (donations, getting enough people to commit,
people paying whatever they think is fair), but his point that conventional
sales are economically efficient was probably fair (most of the other
schemes involve more work). However he doesn't address the argument that FS
development is more efficient than PS - this could counter the
inefficiencies in the more complex payment models. The study ignores the
actual software, i.e wether an FS or PS solution is better or worse and
anything to do with software development costs. As someone said it appears
to assume that software has a unitary development cost, whatever
development/licensing model you choose. this may or may not be true.

The study goes on to show why pricing is useful in economics in order to get
consumers what they are prepared to pay for and producers to decide what to
produce. This is all fair enough, but does ignore non-monetary
considerations entirely, which is clearly not entirely realistic in a FS
world (although developers do still have to eat).

His thesis relies on software developers being a scarce resource, and the
bulk of software development being packaged, rather than customised,
software. That might currently be true but it might not remain so in an FS
world.

Some useful comparisons/analogies were made with mathematics/academia. Are
they 'economically inefficient' because there is no priced market for their
outputs (research papers)? 

He makes a reasonable case why using cross-subsidy from service agreements
to pay for development will not work well (because people who just provide
service agreements will always be better value), although again  I think
this ignores the intrinsic efficiencies in supporting FS.

He also says some very non-PC things about IPR, at least from a FS
perspective. Alex asked (paraphrasing) 'You say the GPL provides no protection 
of
intellectual property rights, but that's nonsensical to me, so which word
are you using which is different to my understanding - protection,
intellectual, property, rights or no'?

It concludes by saying that Govts shouldn't specify FLOSS in general as
the economics doesn't justify it - they should go on TCO alone in specific
cases. (The study was commissioned because Munich chose Linux even though
the TCO was higher than Windows).

In summary I don't think we are in too much danger from this analysis. It
makes some economic sense, but is so clearly divorced from reality that the
effects he is ignoring must be more important than the effects he is
considering.

Hope that was useful for those of you not in a position to dash to Cambridge
for lunchtime next day.

Wookey
-- 
Aleph One Ltd, Bottisham, CAMBRIDGE, CB5 9BA, UK  Tel +44 (0) 1223 811679
work: http://www.aleph1.co.uk/     play: http://www.chaos.org.uk/~wookey/