A limited test of the feature has been running since February. According to Amy Brand, CrossRef's director of business and product development, the company initiated the feature after noticing that researchers were adding DOIs directly into documents. Initially, the feature was restricted to CrossRef members only. Brand explained, "In the February trial launch, you needed a password, but we quickly removed that when we realized that researchers could handle looking up DOIs at the authoring stage and then cut and paste the references."

The Simple-Text Query can retrieve DOIs for journal articles, books, and chapters in any reference citation style, although it works best with standard styles. In the CrossRef system, each DOI comprises an alphanumeric string that connects to the basic metadata and a URL pointer to the full text, which together form a unique identifier for the content item and a persistent link to its location. (Those interested in further information on the entire DOI array should check out the Web site of the International DOI Foundation at http://www.doi.org.)

Bruce Rosenblum, CEO of Inera, assured me that the feature is style-format independent. "It is not tied to format styles; it uses a heuristic breakdown of references, parsing elements directly. It can even handle non-standard formats that the author develops or take a reference list with different format styles for different entries."

Nonetheless, instructions on the CrossRef site for Simple-Text Query urge users to separate individual references with a blank line and to submit lists in alphabetical or numbered, sequential order. Though the system can accommodate some variations, such as special characters and midreference line breaks, CrossRef recommends using hard returns only at the ends of references and converting entries to plain ASCII text. The feature can handle multiple bibliographic citations in a single reference, but users should remove annotations before submitting their bibliographies.

After retrieving the DOIs, CrossRef recommends that users check accuracy by clicking on the DOI links. In some cases, single references may retrieve more than one DOI. Also, some scholarly material may not find matches—e.g., conference proceedings and presentations, where CrossRef's coverage is still weak. However, CrossRef is continuing to expand its outreach to more types of scholarship. Last April, CrossRef announced that it had registered more than 20 million content entities since the inception of its system in early 2000. The vast majority of these DOIs went to online journal articles, but more than 1.6 million were assigned to conference proceedings and books at the chapter as well as title level. In 2005, it began supporting technical reports, working papers, dissertations, standards, and data elements or data sets. Overall, its growth rate is accelerating rapidly. It took CrossRef 4 years to reach 10 million DOIs, but only 2 more years to reach 20 million.

When it comes to DOI assignments, CrossRef is not the only player. R.R. Bowker, LLC (http://www.bowker.com), publisher of Books in Print and the U.S. assignment agency for ISBNs, is also a DOI registration agency for books and other publications qualifying for an ISBN, as well as individual chapters, excerpts, and illustrations available in either physical or digital formats.

I asked Brand if CrossRef could reach open access material. She assured me it could, but it clearly did not give the free and sometimes underdefined material any preference. For open access sources such as institutional repositories, she said: "We have very explicit policies. We put a lot of work into defining what is an original work. For example, a working paper or technical report represents work with a different status coming from an earlier stage of the research process. So we consider them different works. If something is published as a preprint or postprint, we don't give it a DOI. When we work with institutional repositories, we only accept the material that is not published." CrossRef actively encourages DSpace repositories to assign DOIs to original, nonduplicative works and register their DOIs with CrossRef, rather than just relying on registration with CNRI.

CrossRef has grown rapidly in the last few years, both in numbers and in types of members. As well as its 2,204 publishers (both commercial and scholarly societies), it now has 1,115 library members, plus intermediates including affiliates (33), associated organizations (5), agents (7), and linking solutions partners (8). The link resolver platforms with which it works closely are Endeavor's LinkFinder Plus, EBSCO's LinkSource, Ex Libris' SFX, Innovate Interfaces' Web Bridge, OCLC's Pica OL2, Ovid's LinkSolver, Serials Solutions' Article Linker, and SIRSI. Intermediates also include secondary publishers, aggregators, and abstracting and indexing services.

CrossRef also supports OpenURL links, using the OpenURL syntax in its own system and making all its publishers "OpenURL compliant" for its library participants. It also works with services such as Google Scholar, Microsoft's Windows Live Academic Search, and Elsevier's Scirus to connect content to the leading Web search engines. With Simple-Text Query, it hopes to reach out to researchers directly for spreading the use of DOIs in scholarly publishing.

The new feature is also expected to have special appeal to smaller publishers in CrossRef. According to Brand: "We have seen with the Simple-Text Query trial that more publishers are now able to implement reference linking. CrossRef serves a very broad range of publishers, including many members with limited resources and technical expertise, and we are always striving to level the technological playing field for our members." CrossRef expects many smaller publishers to use the easier interface as an alternative to CrossRef's XML-based batch query interface. It also expects soon to add a member-only version of the Simple-Text Query that will be able to upload whole files.

The new CrossRef feature comes from a collaboration with Inera, a firm that provides software solutions to editorial and production areas of publishing. Inera's eXtyles product family was originally designed around Microsoft Word and content management systems that automate editorial, reference linking, and XML production processes. According to Rosenblum, Inera products now work with some 600 journals from around 70 different publishers. The Simple-Text Query is a customization of eXtyles refXpress, which, in its full form, parses references in an unlimited number of styles without preconfiguration and can then automatically reformat references to a specific editorial style. Rosenblum warns that if a user retrieves an indicator that a reference has no match, particularly if the reference comes from a well-established publication, the reference might just be fraudulent. Just another little service for editors, publishers, librarians, and peer reviewers from your friend CrossRef and its friend Inera.