How Open is Google?

MARCH 8, 2008

Recently I randomly came across this (fairly old) post from an environmentalist who was getting excited about Google’s sketchup. As I wrote in a comment on their blog:

I applaud your support for “an economic model that works by sharing rather than hiding” as well as your appreciation for the “power of open source” but would point out that sketchup itself is most definitely not open source (nor are pretty much any of the rest of Google’s online toolkit).

While one can be thankful that Google have provided this and other services (and in most cases have also provided support for export in an open format) we would do well to keep clearly in mind the distinction between ‘free as in beer’ and ‘free as in freedom’ i.e. free to reuse, redistribute (see for more details).

We should also remember that it has long been common in the software industry when pricing proprietary products to follow a ‘cheap (free) at the start, expensive later’ pricing scheme in order to exploit the lock-in that arises from the existence of switching costs and indirect network effects.

Thus, it might well be worth investigating (or contributing to) an 3-D design system that really is open-source on the basis that as truly ‘open’ system will prove the better investment for your time and effort over the long-run.

This is a general point. Generally Google gets very good press whether among the wider public, the ‘open’ community or techies. Yet actually almost all of the products they product that are in any way core to their business have remained completely proprietary (and in many cases secret). Consider the following list:

  • The Google File System: proprietary and partially secret (some documentation in some academic papers)
  • Google Search (the algorithm and related software): proprietary and mainly secret. Basic pagerank idea was published (not original anyway since part of bibliometrics for a long time) but changes since then are not generally published. Of course there are reasons, primarily related to spam why you would want to keep this information secret but one could always deal with this by publishing information with a lag (e.g. 2-5 years)
    • Google Stripped Down Linux: proprietary (only used in house so GPL ineffective)
  • Google Docs: proprietary (free to use) though does use open document formats.
  • Google Groups: proprietary (free to use) and with no obvious export format (though I may be missing something here).
  • Google Mail: proprietary (free to use)
  • Google Code: proprietary (free to use) (as far as I can tell)
  • Google Knol: unknown at the present time
  • Google Scholar: proprietary (free to use)

Thus one might wonder why Google is generally seen as such a supporter of ‘openness’. One simple answer is that at present our primary experience of Google is of it providing a ‘free’ (as in beer) services while making the money by selling our attention to advertisers. While all of Google’s core products remain proprietary the fact that users are not charged at the point of use makes it seem we’re getting something wonderful for free – all the costs invisible because either they occur indirectly via the impact of advertising (higher prices etc) or in the long-run (potential lock-in to a single provider whose interests in terms of innovation and competition are not likely to coincide, at least in the long run, with those of society as a whole).

It is particularly interesting here to compare Google with Microsoft (perhaps at an earlier stage of their progress to the Monopoly they now have). Microsoft also supplies a proprietary product but one which, because of the demand structure of the industry, is priced directly to the consumer. This combined with its closed nature and ubiquitous use have combined to make Microsoft, at least at present, much less popular both generally and in the free/open community than Google, with a fairly widespread belief that the effects of the Microsoft monopoly on the software industry and wider society have been, let us say, ‘mixed’.

A second point would be that because of the nature of Google’s products there are quite a few areas where it is willing to support open-source development because those open-source products are complementary (and not therefore competing with) its own activities. For example many of the open-source activities that receive support via Google’s Summer of Code are related to basic tools (e.g. python) which are an input into Google’s work but in no way compete with anything Google do (it would be interesting to see what would happen to Google SoC proposals that did propose work on projects directly competing with Google’s own activities – for example work on an open search engine).