A blog for the enterprise open source community
Open core licensing is free software’s evil twinMatthew Aslett, November 10, 2010 @ 10:50 am ET
Or, why free software advocates love to hate open core
I’ve been trying to figure out why it is that free software advocates are so fixated on the open core licensing strategy and recently came to the conclusion that there is only one explanation: open core is free software’s evil twin.
To clarify I do not believe that open core is evil, but that the relationship between free software and open core is the equivalent of the literary device where two protagonists share certain characteristics (such as general appearance) but have inverted moralities and visual differentiators (usually a goatee beard).
If we look at the relationship between open core licensing and free software we also see common characteristics along with diverging moralities.
With regards to the common characteristics witness the fact that the two strategies are united by a dependence on strong copyleft licensing.
According to our recent research on open-source-related business strategies 67% of vendors utilizing the open core licensing strategy are associated with a project that uses a strong copyleft license. We have also found that 52% of vendors taking a single open source licensing approach to open source use a strong copyleft license.
Looking at it another way we see that 30% of vendors associated with strong copyleft licenses are using open core licensing, while a very similar number – 29% – of vendors associated with strong copyleft licenses are using single open source licensing.
That is where the similar character traits end and the differences begin.
While free software projects utilize strong copyleft to ensure that the software in question remains open, vendors using the open core licensing strategy use strong copyleft licenses, along with copyright ownership, to ensure that only they have the opportunity to take it closed.
While 83% of vendors utilizing the open core licensing strategy are associated with a project for which the vendor owns the copyright, 88% of vendors associated with foundation-owned copyright were using open source licensing.
Meanwhile 56% of vendors taking a single open source licensing approach were using the bazaar development model, compared to 61% of those taking an open core approach using the cathedral development model.
Similarly 43% of vendors taking a single open source licensing approach were using the community-led development model, compared to 80% of those taking an open core approach using the vendor-led development model.
Finally, while 96% of vendors utilizing the open core licensing strategy generate the largest proportion of their revenue from closed source software, 32% of those associated with single open source licensing generate revenue from support subscriptions, and the same proportion from ad hoc support/services.
The evil twin theory doesn’t explain why the debate is so enduring however, or why free software advocates seem to be so fixated on open core. That is unless you add in the theory that the twins are also symbiotically dependent on each other.
You don’t have to look hard for evidence that open core is dependent on free software – the statistics above demonstrate how open core related to the strong copyleft licensing strategy – but what of free software’s dependence on open core?
It seems to me that in a world where the line between proprietary and free and open source is increasingly blurred advocates of free software are becoming increasingly dependent on open core as the bogeyman to define the line and differentiate the free software approach. Where once ‘proprietary’ was considered the opposite of free, now it is open core that is considered the opposite of open source.
The dependence has gone so far, in fact, that we have seen examples of free software advocates labeling projects and vendors as open core, even when they are not, in order to highlight the benefits of a pure open source approach. Witness Bradley M Kuhn and Alexandre Oliva attempting to pin open core’s goatee beard on Canonical and the Linux kernel respectively.
If we look back at the creation of the term ‘open core’ it was coined in order to provide an alternative to terms such as ‘bait and switch’. In hindsight it was inevitable that the negative connotations would simply be applied to the new terminology.
What wasn’t obvious was how important open core would become to the software freedom movement in articulating the benefits of software freedom. That is why free software advocates love to hate open core, and that is why the open core debate will endure.
Comments (3) Categories: Licensing,Software