[License-discuss] Intimacy in open source (SSPL and AGPL)

[Thorsten Glaser]

Lawrence Rosen dixit:

>But I also understand and appreciate the MongoDB business case dilemma. If
>they just give their software away without some copyleft conditions for free
>network use, they will not profit much from it.

I don’t.

It’s possible to support a large enough company (I know of several ones
in the “several dozen employees” range, my employer lies in the 200–300
range) by selling service, support and customisation services on other‐
wise OSS software. This might not scale down to smaller companies (as
SLAs might be necessary to win big customers) and works better if the
actual software in question isn’t something you’d run off-the-shelf but
customer‑ or trade-specific, admittedly.

Several more companies survive on an “open core” model, which I perso‐
nally dislike (they aren’t likely to accept even third-party patches
to the “open core” if these are in concurrence with proprietary/paid-for
extensions, and it’s (as said above) not strictly necessary to survive,
it just increases the margins) but slightly proves the point.

It’s also possible to have a mix of OSS projects, customers for these
(which, besides the already mentioned service, support, customisation,
installation help, etc. also encompasses paid CRs (I pay you for adding
feature X to the OSS version everyone can then use)), and traditional
customers. Here, “traditional” does not necessarily mean closed-source…
we have customers the software we develop for is OSS, but we don’t
publish it (completely) except for giving it to the customer, although
they could do it (money-counting cheaps they are, they won’t)… but we
*can* use components from it, after cleaning off the customer name and
customer-specific data and processes, in other software, of course. But
the occasional closed source project also won’t hurt.

In the end, it hinges on reputation (both in the world of suit-wearing
people and in the community… my employer does sponsor OSS work, shows
up at conferences, etc.) and a certain minimum size, but definitely NOT
on underhanded licencing practices.

Heck, I even could see some model in which the software is non-OSS
(perhaps shared source, if not closed source) for 2–3 years and becomes
OSS afterwards (in which time a new non-OSS version with more features
will have arrived), but only if the “OSS afterwards” is done right: not
by just dumping the code, but with all the usual things (some community
around it, accepting patches, community support, ⚠ vendor-provided
security fixes). This could also offer paid support for those who need
it (either on a plan or occasionally) but can live with not-the-latest
of the software (as long as it’s still sensible enough to run, that is,
provided with security and bug fixes… if the vendor is good, these can
be easily merged forward and backward with their newer internal version…
even third-party stuff can, given a CLA or a suitable inbound licence).

bye,
//mirabilos
-- 
I believe no one can invent an algorithm. One just happens to hit upon it
when God enlightens him. Or only God invents algorithms, we merely copy them.
If you don't believe in God, just consider God as Nature if you won't deny
existence.		-- Coywolf Qi Hunt