Moving the goalposts

I based my computation on real life and verifiable numbers on the web ( and I would say, even stuff that are kinda legally required to reflect the truth ). You just take yours out of nowhere, forgetting lots of stuff that need to be paid. Just look at any foundation accounting to see how much this cost.

And both Mozilla and RH have highly decentralized workforce, both are in the same market segment as Canonical, so I am pretty sure that if Canonical started to publish now enough information to compare, the numbers would be the same. But feel free to find me a real life example of a organisation that match your numbers.

Also, 33 developpers are just 10% of the current canonical workforce. That's enough to make a small distribution, but in no way enough to sustain several complex project like Unity, Upstart, Bzr, Launchpad, Mir, Gwibber, Lightdm, Juju, Maas, AppArmor, software center, UbuntuOne, and all the distribution ( and port on others platforms like TV, phone, etc ).

You need people for QA, documentation, translation, system administration, etc.

Before people were fired from Mandriva to create Mageia, the company was around 50 person, split between Brasil and France. And this was barely enough to produce a OS and maintain in house code ( installer, drakxtools, etc ). There was less projects than Canonical, and they had income.

Not to mention that if 90% of your workforce leave, there is a high chance that all of them will not work for free for the project anymore ( see Stormy Peters talk about it, something around "would you do it again for free" ). There is also a high probability that you have to remove people who act as bridges with community, thus making the community angry ( again, look at Mandriva and the way people reacted when Adam Williamson or Gael Duval were asked to leave ).

And while you think people will pay to sustain Ubuntu, you are also not taking in account the fact that Canonical will not disappear in 1 day if the worst happen. If something happen that result into using the fund to create a fundation as a measure against Canonical financial problem, the whole project slowly start to go down, and then community will slowly leave ( again, Mandriva ), and so there is likely no others interested organisations to give any money.

A saner model would be to do it right now, when everything go well and to drop the whole CLA idea with copyright assignment to Canonical, cause this prevent others interested companies from investing time ( at least, not without lots of paperwork ). AFAIK, Suse, Google, RH and likely others have such policy, and while they can do it, they would not do for smaller projects.

This would also restore the confidence of community, after the various issues like "amazon integration dropped after the freeze", or "UDS announced 2 weeks before, thus causing issues to people who booked their holidays for it".