The innovation excitement versus maintenance work
A recent survey shows more and more large companies committing to open source in the USA.  "This is one of the huge changes in open source over the last decade -- the move from mostly volunteer developers to corporate participation"
"Open source development provides an increasingly efficient and cost effective way to deal with growing technology needs, while enabling customers to avoid being locked into one vendor or being forced to use its approved partners when they need service or support. Over the next decade, Herbsleb predicts such forces will drive the market for open-source software to 'increase dramatically.'"
Contributions in Open Source Infrastructure is not as exciting
While contributing to Innovations in Open source maybe quite exciting, contributions to infrastructure aren't as exciting. Large part of contributions to infrastructure are sponsored. E.g. Canonical sponsoring Ubuntu research. Companies sponsor contributions to maitenance because it helps maintain the infrastructure. Also its easier than funding innovation.
Consolidation of Internet services is more significant than Open Sourcing the service
While open sourcing a service such as a mail service will allow users to host their own mail service easily, replicating the infrastructure needed for a service is a still a challenge. Also consolidation of a service over the Internet ensures that there is one compliant centralized system than several deviating distributed services. Also a consolidated service is likely to have a large user base. A large number of users justifies that the service tend to be free, even if not open source. An open source service, servicing a very small user group is as good as closed for its sustenance costs, as of now, can render the service unpredictable.
Opening up - Upside and Downside ("Irrelavent discussions")
While the philosophy of open source communities drives quality, because of contributors reviewing, modifying and patching up code. Some big corporations while they support "freeware" are still largely averse to open sourcing. It is likely that they believe - opening up their code, may lead to several irrelevant discussions that may become a hindrance in a typical corporate discipline of execution.
Case of critical contributions to Wikipedia
Innovation and Contribution Excitement - wikipedia Aron's article
The tech.fortune.cnn article also says that "The key benefit for the companies that coalesce around a particular open-source platform is that open-source encourages innovation and the rapid expansion of the market can lift all of their boats." This is clearly infrastructure development. Aaron's article about the contributions to Wikipedia elucidates how critical contributions seep into open source software, by analyzing at contributions to Wikipedia. Innovations in software evolution can be compared to these critical contributions.
 Aaron explains the thesis that innovations may come from individuals but for standards and interoperability the ardent get motivated to police and maintain the presentation and usability. While the corporate world are motivated to do this for open source software; guess they can therefore also get to say what is a contribution and who are the vandals over time. Aaron analysis his finding as "everyone has a bunch of obscure things that, for one reason or another, they’ve come to know well. So they share them, clicking the edit link and adding a paragraph or two to Wikipedia. At the same time, a small number of people have become particularly involved in Wikipedia itself, learning its policies and special syntax, and spending their time tweaking the contributions of everybody else.. This second group is clearly very important — it’s thanks to them encyclopedias have a consistent look and tone — but it’s a severe exaggeration to say that they wrote the encyclopedia."
Generative nature of open source
There are other people trying to look into the nature of contributions to major open source software. [Gnome analysis one] [Ubuntu contribution analysis in-house] Kiran says it contributions to Gnome is likely to be more generative than contribution to Wikipedia, since a contributor to Gnome or Linux kernel needs to acquire specific skill and knowledge of this specific narrower domain as compared to Wikipedia contributions. It is not clear why this is not similar to Wikipedia analysis if one were to consider only the sub-group of people who are comfortable with Gnome or Linux kernel.
Contributing to infrastructure is not exciting. Another potential example: BOSS operating system maybe is an example.
Explorations: Open Content, Access, Workflows
Open Source and Collaborative Communities (like Wikipedia) hint: school software could be a collaborative community
Develop: "Open source" as concept relative to what one is utilizing as tools. (and products)