My experience is that ESA lets the industry (and maybe universities) do most of the SW/HW development and ESA does stuff like requirements and high level management. But I have never followed a project from start to finish, only been in the aforementioned industry :)
I develop data processing software.
You are right, ESA usually oversees the projects and hosts the main data archive. Esa also takes care of the launch vehicle and interfaces with Arianespace for example.
The hard- and software development for the payload is usually done in numerous science institutes, part of them also universities. They usually have subproject-leads, while they again subcontract to external companies, as needed.
A complex structure that has to do mainly with the way how European projects are funded.
What did you do for Esa? What is your view on these projects? We always had the impression, the companies are under a lot of budget pressure during the missions.
I've never seen code intended for space missions. I guess they don't share it as most of it can't be reused for other missions so there's really no point in doing it.
They also seemed to run their own summer of code in 2013: http://sophia.estec.esa.int/socis2013/
I work for ESA space missions (not Rosetta). Although my colleagues are certainly well qualified, I am not sure if we are really the best of the best. (we are not the typical web 2.0 programmers though, this is certainly true :o).
Usually, the software for space missions is written by electrical engineers or physicists/astronomers, who somehow specialized in software after their master or phd. Indeed, the style of that software is quite different from what is usually developed in business setups. No unnecessary complexity, just straight forward paths to the solution. We use more or less exclusively open source tools. No windows too. Works pretty well. Our results are also open source, btw. Just to answer some guys who believe, this is all military, secret and disclosed.
But we also have to fight cowboy-coding and design-pattern-gurus. We usually don't have to fight software managers though. Everybody is so well qualified, that we use to discuss once a week what we do, where we want to go and then just do it.
Worked very well, so far. My last mission was a huge success :o). The real innovation may indeed be done at universities and science institutes, not so much in startups and business. Which startup can afford to hire 200 scientists and invest a billion to do something new? I think there is indeed no competition on that level.
But we also give a lot of tasks to industry (we have to, because of the public funding). In some cases, this caused trouble, because their main focus is to make profit and still be good enough to avoid problems. Also, they usually keep their methods secret for legal reasons. If something does not work in space, it can be hard to get the necessary information to fix the problems (while the time is ticking, because many missions have finite lifetimes).
Our goal is usually 'as good as it gets, outstanding, if possible'. We can afford that and that is really cool. It doesn't work in a business setup, because it usually is expensive.