The beta version of Google Scholar, launched last Thursday (November 18), has been warmly welcomed in the scientific world. But questions remain about some aspects of the academic search engine.
The tool is targeted squarely at scientists and academic researchers wanting to find scholarly literature across broad areas of research. It searches a subset of the main Google index, open repositories, and the Web sites of academic publishers, professional societies, preprint repositories, and universities.
Duane Webster, head of the Association of Research Libraries, said the arrival of the new service was welcome but noted the need for an open dialog as it develops. "As their experiment moves forward, there will be a range of important questions that will need to be addressed," he told
For example, Webster pointed out that Google has declined to name the participants in the development of the service. There are concerns over this lack of information, and with "Google's unwillingness to describe how it determines what is scholarly," he said.
That said, "it is likely that their criteria over time will evolve and, like their other services, I am sure this is a capability that will improve significantly with experience," Webster added. The Association of Research Libraries is an umbrella group of 123 North American research libraries.
Although Google has been quiet on a number of matters relating to Google Scholar, the engineer who led the project, Anurag Acharya, has told the media the company received good cooperation from publishers in the sector. "Google as a company has greatly benefited from academic research, and this is one of the ways we can give back to the community," he added.
Open-access publisher BioMed Central (which is a partner with
"We, along with others in the scientific community, have been talking to Google about offering a service like this for some time. We are very pleased that they have taken this step," Velterop said.
However, Michael Eisen, co-founder of another open-access publisher, Public Library of Science (PLoS), pointed out that people using the service will face an important obstacle—many of the articles are only available to journal subscribers.
"While Google may have the noble goal of making 'the world's scientific literature universally accessible, ' the fact remains that most of the articles returned in a Google Scholar search can only be accessed by those fortunate enough to have a subscription to the relevant journal," he told
A PloS spokeswoman added: "Google will find that they can better serve their searchers' needs for access to complete scholarly articles by 'flagging' as open access or ranking more highly those that are freely available online. Such a system would minimize people's frustration at finding an article that looks perfect for their research needs but discovering that they are unable to access it. This frustration is already quite evident in the various threaded discussions occurring online."
The new search service also provides citation information about the articles retrieved by stating how often a paper has been cited in scholarly literature, Google said. But the company conceded that "your search results may include citations of older works and seminal articles that appear only in books or other offline publications."
Nor is it entirely clear yet how the citation ranks given by Google relate to those on more established services. "I think [the citations are] a potentially very valuable tool," said Tim Mark, the executive director of the Canadian Association of Research Libraries. However, "I think the research community needs to test drive [Google Scholar] for a while," he told
Mark said he was enthusiastic about any tool that brings together information that researchers need. But he cautioned, "This is not a panacea. We still need the human judgement to sift [the results]."
Mark was also concerned about the potential lack of access to subscription-only material and added that he, too, would like more information on the search algorithm.
Some features for narrowing searches will be added in the future, Acharya said, including a way to limit results by the date of publication. "We will gradually refine it once we figure out which things are important," he said.