Open source: The numbers don't tell the whole story

Savio has had some interesting posts lately crunching numbers related to open source investments and, most currently, the FSF's financial ability to litigate GPLv3. In both cases, though, I think he's getting a bit too attached to the calculator and detached from reality. (Not that I ever stray from reality.... :-)

(That said, in both cases I appreciate the exercise that he walks us through. It's useful.)

With regard to the FSF's ability to litigate over GPLv3, this misses the point. The FSF has never needed to rely on litigation to enforce the GPL. That's what public pressure is all about, and the open source community can deliver that in abundance. I've yet to see it fail.

To the extent that the FSF needed money and it was over a principle that affected all, I would imagine that Savio's employer, IBM, as well as others might find cash in their wallets that they could put to such a good cause as public flagellation of would-be GPL scoffers.

On the open source investment front, Savio suggested that open source has yet to produce stellar returns for investors. This is true, so far as it goes. But his analysis requires carefully overlooking the obvious: we're still early into the real money being pumped into open source. (He also finds a number - 57% - that may be accurate but it represents nine complete failures to find one Google, so maybe the criticism should be that we haven't produced a Google yet...?) Most of the early money invested in open source wasn't invested in open source companies at all or, rather, into companies that styled themselves as such.

So, if you focus on pureplay open source companies, you'd have great returns on companies like Red Hat, VA (momentarily :-), and JBoss, but negative returns on Linuxcare, Lineo (my alma mater - just doing my part to keep the market rational), and Caldera.

The real analysis needs to begin in 2002 or so with the rise of the real open source investment market: MySQL, Hyperic, MuleSource, JBoss, SugarCRM, Alfresco, Pentaho, JasperSoft, etc. (and Gluecode, with which Savio is no doubt familiar). I'm confident that VCs will see strong returns from this crop. Most of these companies have had strong "up" rounds in their funding, indicating that the market is already valuing them well.

So, to Savio's point, I'd suggest that it's possible that VCs will slow their investments for a year or two in early stage open source ventures, until the money starts to return to them via acquisitions and IPOs. I'm guessing 2009 will be the banner year, but we'll see some strong action in 2008, too.

Regardless, given that VCs are in the business of diversifying their risk, as Larry Augustin notes, it's probably unreasonable to expect them to load up on open source startups given that having one to three open source startups is probably all they should take on right now, rationally speaking.

Once the IPOs hit, however, rationality will probably flee. So that's the time when the investments will start in earnest, and the returns will be just as goofy as in other inflated markets (like Web 2.0).