Tuesday, July 25, 2006

Tim O'Reilly's Big Ideas

Tim O'Reilly opened OSCON 2006 with a list of big ideas. A core theme of the conference is how to think through the future meaning of open source and how the concepts of open source apply to web applications.

While many of Tim's ideas are well understood, they are interesting to think through and apply to the changing nature of opensource - which as a concept is moving away from solely source code to also include hosted applications and/or user-driven phenomena such as Flickr, Youtube, etc. that involve invoking publicly defined APIs independent of access to source.

The core ideas follow.

  • architecture of participation
    • design systems that are designed for user contribution
    • well-defined APIs
  • asymmetric competition whereby community contributions leverage the company/project
    • for example, craigslist, internet rank = 7, employees = 22
    • versus yahoo internet rank = 1, employees 9,000
    • youtube is another great example of massive page view growth uncorrelated to headcount growth
  • change in meaning of openness
    • as applications are increasingly delivered as a service versus installed applications some of the licensing issues become moot
    • with internet applications the concept of openness moves away from GPL-like licenses to discussions regarding the openness of APIs and degree to which data is portable
    • what should developers expect wrt apis, api support for certain versions, etc
    • should there be a GPL-like licenses for APIs?
  • operations as advantage
    • as applications move to the network, competitive advantage will be increasingly center on APIs, SLAs, and datacenter operations
    • Amazon's S3 is an example of the rise of services as a product offering
    • Craigslist, for example, manages to a key metric which is page views per kilowatt hour
  • open data
    • is data portable?
    • who owns data stored in flickr, Amazon, etc?
    • will the industry support microformats to standardize data representation and make data portable
    • Tim highlighted movemydata.org - which is making the case to ensure access to user data
    • Tim sees data as the Intel-inside equivalent of Internet applications. Given the tremendous value of data, issues relating to user access, ability of users to move data from one service to another, etc will become important to think through

As I mentioned in a prior post, Innovation Happens Elsewhere, the axiom of distributed innovation is premised on the fact that one can never employ, pay, or manage all the smart people in the world. Nor can one monopolize innovative ideas. Accordingly, companies need to be architected to leverage the innovation of others and find means of allowing third parties to see benefit and value in using your APIs, content, data centers, etc to add mutually beneficial value.

No comments:

Post a Comment