Whenever I finish an interview, I always am embarrassed by the one obvious follow-up question I didn't ask.
Shortly after I hosted a conference call last Thursday on the GPL v3 draft process, I smacked myself in the forehead for not following up on this comment by Matt Asay:
"I would argue that it's true that giving the Web properties a free ride encouraged them to get started on a dime rather than a dollar; I'm not sure I'd go so far to say that Google would be out of business or never would have gotten off the ground if it had paid more than $69.99 to Red Hat ..."
Asay was discussing the prospects of closing down the "ASP Loophole" in the GPL, which now is alive and well in the v3 draft but, call participants generally agreed, is still in the cross hairs of the open source community as a general shortcoming in OSS licensing. Google is often cited as a likely source of pressure on the Free Software Foundation -- which administers the GPL -- that caused it to drop provisions in v3 to force services companies that modify open source code to open up that modified code, even if they don't re-distribute the software directly.
My question to Asay should have been:
"But Google isn't in the business of not going out of business, right? It's in the business of putting other folks out of business."
That's overly Darwinian, or maybe Patton-esque, come to think of it. But the reality is that businesses are not communal; they can cooperate with each other, but that's typically only in times when there seems to be plenty of revenue to go around, and even then allegiances posture themselves against other businesses.
This has always seemed to me to be the ultimate conundrum of how open source -- I mean the fervent, no-software-patents kind of open source the FSF started out the v3 draft process with -- will mesh with business, which is far better described as an ecosystem (think Whiskers v. Lazuli) than a community.
I'm even more upset that I didn't ask that question, because Asay had an interesting answer that he volunteered as a concept later in the call (fat lot of value I added). He suggested that in a future licensing scheme, companies that employ open source code internally could elect to contribute cash instead of open code.
Obviously, details on how this cash contribution might be metered or accomplished are unclear -- some might argue that Google is already making a pretty big cash contribution to open source by encouraging its developers to work on side projects while they are on the G-clock -- but Asay's general suggestion caught my interest, as well as that of Rob Enderle, who commented on the call today in his blog. (The hour-long podcast of the call is definitely worth a listen.)
Seems to me that businesses, particularly SaaS vendors who have benefited enormously from the ASP loophole, would be far more inclined to spend money than to open up code that they considered to be strategically important. If every smart piece of development in, say, the CRM SaaS space, were to be opened up, you'd end up with a sector driven by marketing spend, which ultimately is good for nobody -- certainly not customers.
Of course, there are a ton of questions that would have to be answered before a cash contribution model would be viable. How many mod generations back would the cash contribution need to go? Would companies be able to agree to a percentage of future revenues as opposed to up-front outlays, to encourage vigorous open source adoption in niche SaaS sectors? All kind of annoying business stuff.
But clearly these are the kinds of answers open source vendors need to find if they intend to participate in the SaaS boom many of them are driving.