On the third Thursday of November, wine enthusiasts the world over traditionally race to be first to try out l'arrivée of Beaujolais Nouveau, a barely fermented wine that has rewarded harvest workers for seven centuries. This year, the same Thursday saw a second race as scientists and other information addicts rushed to try out a new vintage: ‘Google Scholar’, a test version of Google search that restricts web searches to academic texts (see page 423).

That the arrival of Google Scholar is an ‘event’ can be sensed from the waves of giddy excitement across Internet discussion groups as people posted their own musings on the free service. By its sheer scale and functionality, Google Scholar is a disruptive technology heralding profound changes in the way that many will access scientific information.

Google deserves congratulations. With a speed that only the private sector can deliver, it has addressed the most basic need of users accessing the literature: the ability to search the full text of a broad swathe of scientific information, cutting across publishers, journals and other sources. Two reservations: some publishers have not made their text available, and Google has not yet indexed all of the text made available to it by others, including Nature Publishing Group.

But Google's database is orders of magnitude bigger than anything that exists today, and its search algorithms comparatively of lightning speed. The top search hits are the most cited — your search results won't be clogged by pages of poorly cited papers — and, in a further tweak, the hits at the top are those that are most cited by journals that themselves are the most cited.

Some worriers, including more than a few librarians, fear that Google Scholar will push scientists and students further towards what one describes as a “trend towards sloppy Internet-based research”, instead of professional searches using abstracting and indexing databases, often bought at great cost by libraries. There is a grain of legitimate, if exaggerated, concern here. But one lesson of the Internet is that expensive, centralized ways of doing things tend to get undercut.

Google Scholar is far from perfect — it lacks basic functions such as search by date, characteristic of more costly databases. But the latest offering from Château Google is young, and will no doubt spur further innovation and competition in search. We at Nature drink to that, and look forward to the surprises that future years will bring.