> I re-read the author's essay multiple times and I finally realized his proposal rests on a (probably incorrect) assumption of corporate motives and rationale. He believes the only reason companies avoid AGPL is the extra paperwork burden it causes.
I read this very differently (or I'm misunderstanding you), the reason why companies avoid AGPL is because they do not want to open up their closed source code that they run together with (unmodified) GPL/MIT code on their servers. If they were using AGPL they would have to open source their closed code as well.
Now companies would likely be willing to pay for a commercial licence of an AGPL library they use, to avoid having to open source their closed code. However, many packages don't even have the possibility for dual-licencing (possibly partly due to the paperwork involved for the maintainers), and also having to track commercial licencing (contracts) for many small packages would put corporations off. Instead he proposes that there are "package deals" which lightens the buerocratic burden on both maintainers and "customers".
I read this very differently (or I'm misunderstanding you), the reason why companies avoid AGPL is because they do not want to open up their closed source code that they run together with (unmodified) GPL/MIT code on their servers. If they were using AGPL they would have to open source their closed code as well.
Now companies would likely be willing to pay for a commercial licence of an AGPL library they use, to avoid having to open source their closed code. However, many packages don't even have the possibility for dual-licencing (possibly partly due to the paperwork involved for the maintainers), and also having to track commercial licencing (contracts) for many small packages would put corporations off. Instead he proposes that there are "package deals" which lightens the buerocratic burden on both maintainers and "customers".