This post manages to completely miss the point of the Google letter on openness. As some have incorrectly asserted, Google's latest post is not about "everyth

This post manages to completely miss the point of the Google letter on openness. As some have incorrectly asserted, Google's latest post is not about "everything open", it is an internal memo to product managers that it occurred to them would be fit for public consumption. The central thesis of the letter is that "open" is hard to define (and defined differently by different people) and yet crucially important. Because it is hard to define, many people misuse it (OpenDNS comes to mind), and no one can call them on it because it is a vague term. Internally, Google employees argue about what open means, and the purpose of this letter was to clarify what open means to Google. They lay it out in extremely clear terms, and it is well structured. This response essay by Dixon misses all that completely - he unilaterally defines "open" to mean "open source", and then asks for Google to give away their software, claiming that if they don't, they're hypocrites. Going back to Google's definition of "open" (which is complex), one significant aspect they emphasize is the notion of "lock-in", i.e. locking in users, and locking out competitors. This is one area where Google puts their money where their mouth is; they let users leave. You can get at your email, export your Google Reader content, download all your Google Docs, archive your Picasa pictures, and export all your contacts. That means if a competitor comes along with something better, you can start using it. Try that with Facebook. Dixon acting like that is worthless because it isn't the "open-source" form of "open" is patently absurd. If this is confusing, I highly recommend reading Simon Phipps essays on open source to understand what kind of software makes a good candidate for a business to open source (operating systems, browsers, compilers...the basics that we all use to compete) and what does not (grid/cloud computing algorithms, specialized software, and other types of software that are differentiators). Essentially, the business model is to pool our collective resources in areas in which we don't want to compete (web server software, operating systems) and to spend time innovating in areas we do wish to compete (search algorithms, voice recognition, image analysis, etc.) Over time, as those specialized areas become more mainstream, they become better candidates for open-sourcing.