Zotero.org and Mendeley.

The lack of an adequate method for retrieving metadata has led to proposals such as the Life Sciences Identifier (LSID) [138],[139] and BioGUID [140] (Biological Globally Unique IDentifier)

Figure 4. A typical workflow for using a digital library representing a subset of the literature.

Google Scholar [95] (e.g., [96]–[99]) is a service provided by Google (see also [100]), which indexes traditional scientific literature, as well as preprints and “grey” self-archived publications [19] from selected institutional Web site

Scopus [88] is a service provided by Reed Elsevier and seems to be the Digital Library with individually the most comprehensive coverage, claiming (June 2008) >33,000,000 records (leaving aside Web pages). As far as linking is concerned, Scopus allows links <1?show=[to]?>to its content using OpenURL [89], which provides a standard syntax for creating URIs. For example, the URI

ubMed [75] is a service provided by the National Center for Biotechnology Information (NCBI). The PubMed database includes more than 17 million citations from more than 19,600 life science journals [76],[77]. The primary mechanism for identifying publications in PubMed is the PubMed identifier (PMID);

URIs for granted, but these humble strings are fundamental to the way the Web works [58] and how libraries can exploit it, so they are a crucial part of the cyberinfrastructure [59] required for e-science on the Web. It is easy to underestimate the value of simple URIs, which can be cited in publications, bookmarked, cut-and-pasted, e-mailed, posted in blogs, added to Web pages and wikis [60]–[62], and indexed by search engines. Simple URIs are a key part of the current Web (version 1.0) and one of the reasons for the Web's phenomenal success since appearing in 1990 [63]. As we shall demonstrate with examples, each digital library has its own style of URI for being linked to (inbound links) and alternative styles of URI for linking out (outbound links) to publisher sites. Some of these links are simple, others more complex, and this has important consequences for both human and programmatic access to the resources these URIs identify.

Where available, DOIs can be used to retrieve metadata for a given publication using a DOI resolver such as CrossRef [57], a linking system developed by a consortium of publishers.

DOIs are a specific type of URI and similar to the International Standard Book Numbers (ISBN), allowing persistent and unique identification of a publication (or indeed part of a publication), independently of its location

These applications allow users to add personal metadata, notes, and keywords (simple labels or “tags” [27],[28]) to help manage, navigate, and share their personal collections

Table 1. A summary of some of the digital libraries described in this Review. https://doi.org/10.1371/journal.pcbi.1000204.t001

Advantages of using CiteULike and Connotea

74].PubMed.gov and PubMed Central.PubMed [75] is aservice provided by the National Center for BiotechnologyInformation (NCBI

Figure 2

ObjectFigure

igital library[2–4] denotes a collection of literatureand its attendant metadata (data about data) stored electronically

Warmer digital libraries cannot be achieved by software tools alone. The digital libraries themselves can take simple steps to make data and metadata more amenable to human and automated use, making their content more useful and useable

Aunque difieren en tamaño y cobertura, todas estas bibliotecas digitales brindan instalaciones básicas similares para buscar y examinar publicaciones. Estas características están bien documentadas en otros lugares, por lo que no las describiremos en detalle aquí. Con la excepción de arXiv y PubMed Central, que brindan acceso libre total a artículos completos, todas las demás bibliotecas descritas aquí brindan acceso gratuito a metadatos (autor, año, título, revista, resumen, etc.) y enlaces a datos (el texto de un artículo determinado), que el usuario puede o no tener licencia para ver

A Future with Warmer Libraries

Socializationallows users to share their personal collections and see who else isreading the same publications, including added information suchas related papers with the same keyword (or ‘‘tag’’) and what notesother people have written about a given publication.

Personalizationallows users to say this is my library,the sources I am interested in, my collection of references, as wellas literature I have authored or co-authored.

Table 1.A summary of some of the digital libraries described in this Review.

Metadata: You can’t always GET what you want

The identity crisis, inability to get metadata easily, andproliferation of metadata standards are three of the main reasonsthat libraries are particularly difficult to use and search asautomatically as one would wish.

MetadataMetadata are data about data, e.g., publication metadata include author, date, publisher, etc

exposed and readily available, programmatically and manually,from URIs, HTML [199], and PDF files of publications.Identifying publications.URNs (such as Digital ObjectIdentifiers) should be used toidentify publications whereverpossible. Most large publishers already do this, although there arestill many confounding exceptions.Identifying people.This problem is twofold: people need tobe identified as users of a system and as authors of publications. Totackle the first issue, tools and libraries should use Single Sign On(SSO) schemes, such as OpenID [187] to provide access topersonalized features where possible, as this prevents the endlessand frustrating proliferation of username/passwords to identifyusers in Web applications. The second requires unique authoridentification, an ongoing and as yet unsolved issue for digitallibraries.By following these recommendations, publishers, scientists, andlibraries of all kinds can add significant value to the informationthey manage for the digital library

Simple URIs.URIs for human use should be as simple aspossible, to allow easy linking to individual publications and theirauthors. Short URIs are much more likely to be used and cited[192] than longer, more complicated URIs.Persistent URIs.It has been noted many times before[193,194], but it is worth repeatedly restating: persistent URIsmake digital libraries a much more useful and usable place.Although URIs will inevitably decay [195,196], many (but not all)will be preserved by the Internet Archive [197,198], and everyeffort should be made to keep them persistent where possible.Exposing metadata.Publication metadata, in whatever style(EndNote, BibTeX, XML, RDF, etc.), should be transparently

A typical workflow for using a digital library representing a subset of the literature

the size, coverage, and style of metadata used (summarized in Table 1and Figure 2). Where available, DOIs can be used to retrievemetadata for a given publication using a DOI resolver such asCrossRef [57], a linking system developed by a consortium ofpublishers. We illustrate with specific examples how URIs and DOIsare used by each library to identify, name, and locate resources,particularly individual publications and their author(s). We often takeURIs for granted, but these humble strings are fundamental to theway the Web works [58] and how libraries can exploit it, so they area crucial part of the cyberinfrastructure [59] required for e-scienceon the Web. It is easy to underestimate the value of simple URIs,which can be cited in publications, bookmarked, cut-and-pasted, e-mailed, posted in blogs, added to Web pages and wikis [60–62], andindexed by search engines. Simple URIs are a key part of the currentWeb (version 1.0) and one of the reasons for the Web’s phenomenalsuccess since appearing in 1990 [63]. As we shall demonstrate withexamples, each digital library has its own style of URI for beinglinked to (inbound links) and alternative styles of URI for linking out(outbound links) to publisher sites. Some of these links are simple,others more complex, and this has important consequences for bothhuman and programmatic access to the resources these URIsidentify.The ACM Digital Library.The Association for ComputingMachinery (ACM), probably best known for the Turing award,makes their digital library available on the Web [64]. The librarycurrently contains more than 54,000 articles from 30 journals and900 conference proceedings dating back to 1947, focusingprimarily on computer science. Like many other largepublishers, the ACM uses Digital Object Identifiers (DOI) toidentify publications. So, for example, a publication on scientificworkflows [65] from the 16th International World Wide WebConference (WWW2007) is identified by the Digital ObjectFigure 1. A mind map[207]summarizing the contents of this article in a convenient manner.doi:10.1371/journal.pcbi.1000204.g001

They rely on Uniform Resource Identifiers (URIs [25] or ‘‘links’’) toidentify, name, and locate resources such as publications and theirauthors.

we shall define a digital library morebroadly as a database of scientific and technical articles,conference publications, and books that can be searched andbrowsed using a Web browser.

Tagging.Tags are just keywords, but these allow bothpersonalisation and socialisation of bibliographic data

Managing.When authoring manuscripts, managingreferences in a Web-based repository can save some of the painof re-typing metadata (e.g., author names) for a given publication

CiteU-Like, Connotea, and HubMed

otero, MyNCBI, andPapers

Socialization

Personalization

Zotero.org and Mendeley

metadata, we again mean the basic metadata about a publication,such as the author, date, journal, volume, page number, publisher,etc. In practice, this means any information typically found in anEndNote or BibTeX entry;

Where these libraries differ is in the subscription, personaliza-tion, and citation-tracking features

arXiv.org.arXiv [103] provides open access to more than44,000 e-prints in physics, mathematics, computer science,quantitative biology, and statistic

Google Scholar.Google Scholar [95] (e.g., [96–99]) is aservice provided by Google (see also [100]), which indexestraditional scientific literature, as well as preprints and ‘‘grey’’self-archived publication

Scopus.com.Scopus [88] is a service provided by ReedElsevier and seems to be the Digital Library with individually themost comprehensive coverage

The primary mechanism for identifying publications inPubMed is the PubMed identifier (PMID)