How Red Hat Decides Which Open Source Companies To Buy
darthcamaro writes "You don't really buy an open source company — since the tech is all open. But then again, Red Hat 'buys' open source companies all the time, they just bought one this week. So when does it makes sense for Red Hat to buy a company versus just building it on their own? Apparently, it all comes down to community. 'When you buy an open source company, if the people aren't coming and passionate about staying then you spend a lot of money for what? Because you don't get a lot of intellectual property,' Red Hat CEO Jim Whitehurst said."
You don't really buy an open source company — since the tech is all open.
What a dumb statement. Buying an open source company is buying their copyrights, possibly any patents they hold and getting to acquire their people.
"If you honestly believe that robust community builds better softer, than this makes more sense," Whitehurst said.
Obviously some journalist isn't very familiar with their subject matter.
What I got from the article is that they are will to buy a company with:
an established community based product
centralized developers willing to work for Redhat
For companies that have distributed developers/contributors, they tend to just hire major contributors to steer it in the direction Redhat wants it to go.
I only look human.
My mother is a halfling and my dad is an ogre, so that makes me an Ogreling
As someone who worked in High Performance computing before turning to the dark side of IT/IP law, perhaps I can illustrate how intangible value is calculated
a) Patents - traditionally were the dividing line in the noo-sphere invention space marking where a claim was being worked. Apart from a signaling mechanism (mating call to VCs), most industry accept that the real-worth is the tacit knowledge in headspace of the invebtors
b) Copyrights - may be F/OSS but acquisition of legal ownership grants pararights of offering alternative licenses under different terms, management rights of deciding what to approve/release when, and ability to alter the direction/intensity of project
c) Trademark - or more accurately celebrity rights such as Torvolds by Transmeta. Bragging rights are well accepted as institutions like Harvard collect Nobel laureates. However, you might also acquire traits such as time to market, low bug defect or rapid user feedback
d) Indigenous knowledge - which is the result of domain expertise and can be unique or hard to duplicate. Mozilla in web standards formation/propagation, Apache in internet performance engineering, the complex peripheral skills that enable complex products to evolve. Building a similar socio-cultural organisation with the same incentives, rituals/story telling and passion/values is non-trivial (cf Craig List)
We're a pretty heavy postgres shop, and Red Hat employing one of the top developers of that project (Tom) was the way we chose which distro to use (and actually pay for). Not that we actually needed such support (he gives at least as support on the project mailing list) --- but for marking reasons we needed *a* tier-1 distro with "official" "support" --- so we chose to support them based on them supporting Tom.
I think you would have more return on your money by sponsoring so called Tom yourself and using CentOS or Ubuntu or whatever other free distro.
My reasoning is, why pay for support if you don't need it? If you want to support postgres developers, support them directly.
Apparently, it all comes down to community.
This is increasingly true for all start-ups. Even if a start-up has no IP, and its platform can be easily cloned, it can be valuable solely from the users it's accumulated. VCs now look for at least one of the 3Ps: people, profit, and popularity. Get the people then find a business model to exploit them.
Centos is not free, it is paid by Red hat ( cause packages and software do not write by them self ), so that also make sense to pay RH if people want to keep Centos alive.
And supporting alone a developper is slightly more expensive than paying a company that does it and that share the load among customers ( cause if the price of a RHEL subscription is counted a around 1000 to 2000$ per year, you still have to pay for his paycheck, for travel and sponsoring ( ie, getting to hackfest ), hardware, etc ), so I would say that you can take around the price of 60 subscriptions just to have someone sitting coding for postgresql for you ( and for this, you have a full time coder not working on your product, cause if you pay a postgresql developpers to make it work on somethig else than pg during work time, you are not sponsoring them that well, and he is no longer a top coder ).
In practice, I think that's indeed a good way to do this, but in practice, only few companes are willing to directly sponsor developpers, and only a few, so for the rest, mutualisation of ressources by paying companies sponsoring ( be it RH, or entreprisedb, or any others postgresql supporting companies ) is the way to go.
All this sound to me like the typical BS you can expect from a sales oriented company.
They need communities to save money on development.