I have noted before the rise and rise of Google and its heartening ability to generate a profit from providing a service that is fairly central to the way most people work with the Web. It is still notching up big deals, most notably one with AOL recently, but just as importantly, it continues to innovate in ways that do not threaten its core strengths or values.
Some of these are readily apparent, such as the page of consolidated news stories drawn from around the world (though with a heavy US bias that seems particularly inappropriate for a global search tool). Other developments are less obvious.
For example, Google is making its search engine available as a Web service across the Internet by offering what it calls the Google Web APIs. Since this is based around standards such as Soap and WSDL, this probably makes it the best example of a real Web service to date. Help is available in various ways, including a Web-based newsgroup and a FAQ, and there is also a developer's kit available.
Another interesting area not immediately apparent to casual users is that of the Google Labs. As the FAQ explains, this is a "playground" where Google engineers and users can explore new ideas.
At the time of writing, there were four demos. Two of these showed promise: the glossary, which is just plain useful and should be elevated to official status immediately, and Google sets. The latter tries to predict the next word in a sequence of items that are related in some way.
What is interesting about the Google Labs is that they are attempts to mine new kinds of information from the basic Google search holdings. They are also notable because they invite users to help debug the ideas - employing a classic open source technique. There are even some public discussion boards attached to each demo.
Google seems to have borrowed another approach from open source for its Google Answers section, also currently in beta. One idea that was tried in the free software world was posting coding jobs on a public Web site (called SourceXchange) so that programmers could apply for and complete the work.
The idea failed to take off, in part probably because the dotcom downturn meant that there was insufficient work around to make the service self-supporting, but Google's take - letting users post questions that others can answer for a fee - is promising.
As the FAQ explains, there is a listing charge of 50 cents, plus a fee that can range from $4 to $100, of which Google takes one quarter. There are help and tips on this service, researcher guidelines, a FAQ, and a researcher training manual.
Reading through the list of questions is surprisingly entertaining and in many ways displays a number of the best aspects of Usenet without the worst - no spam or loonies. Interestingly, the main Google Answers page offers a categorisation of questions. This makes the comparison with the hierarchical Usenet even stronger, and also emphasises another suggestive parallel. For just as I wrote a few months back that eBay was a kind of Google for the physical world, so the Google Answers services can be seen as a kind of eBay for facts and knowledge.