BibGlimpse: the light-weight literature manager based on Webglimpse

Detailed information on how to install BibGlimpse is provided on the BibGlimpse setup page.

The BibGlimpse application provides a light-weight PDF reprint manager with automated bibliography retrieval for most PubMed listed papers, allowing scientists to create, manage, and share their personal collections of annotated reprints. Building on WebGlimpse, it inherits sophisticated structured full text search.

Unique features

• Personal collections and annotation support. Groups of researchers can compile subject or project specific collections of manuscripts, and manage and share these annotated content enriched collections for collaborative literature research. Public repositories lack both these features.
• Advanced full-text search.
  • Full text search improves document retrieval compared to pure abstracts based search engines like PubMed.
  • Building on WebGlimpse, simple and complex queries are permitted: misspellings can be tolerated if requested (approximate matching). WebGlimpse queries further support phrase search, soundex search, and regular expressions. Logical operators (AND, OR, NOT) and constraints (like NEAR) are available for construction of complex queries.
• Coverage is only limited by what files a researcher has access to. Manuscripts not listed in PubMed or elsewhere can also be annotated and searched. This is particularly interesting for technical reports, which have not been abstracted, as well as articles which are only available through subscriptions (for example all Elsevier titles and all recent papers by Oxford University Press are not accessible by Google Scholar).
• Light-weight and ease of day-to-day use. Indexing and retrieval of bibliographic records (matches 95% of PubMed listed papers) is fully automatic and transparent. Reprints can be added with a simple upload button (no forms to be filled), or by just saving the PDF file to the appropriate network folder. References can then be searched or exported in MedLine or BibTeX formats for citation or reference managers.
• Built to support value-added content. Making use of WebGlimpse structured queries, we support multiple fields, which can be searched individually or transparently together. Fields supported out of the box are bibliographic record, user annotation, and the full text of the paper. Making these available in plain text files allows, however, easy integration with external tools, particularly text-mining. It is known that mining of full text articles can extract more information than available only in abstracts. Making user annotation generally available to mining tools, which is content rich and can spell out knowledge implicit in the paper, is expected to significantly further improve performance. Results of such tools can be made available as additional fields for queries. While BibGlimpse does not itself provide such tools it has been designed to allow straightforward integration.

We have compiled a feature comparison with other popular tools, indicating the special niche filled by BibGlimpse.

Installation

Detailed information on how to install BibGlimpse is provided on the BibGlimpse setup page.

Additional Notes

User friendlyness in every day use was a key criterion in the development of BibGlimpse: So to add a new PDF file to an existing BibGlimpse archive, it is sufficient to save it to an indexed directory. BibGlimpse will all by itself take care about obtaining the correct bibliography from PubMed. There is no typing of author names or journal titles required. Just saving the PDF file to disk and adding some personal comment or annotation, if one likes, is enough to add the file to the archive. BibGlimpse will automatically provide appropriate Medline and BibTeX records for each PDF and it will consequently render the whole collection full-text searchable. Note that BibGlimpse does support structured queries. Queries for an author name can thus be limited to the bibliography field and will not produce hits from the references in the full-text for example. This is a considerable advantage over simple full-text searches on a number of PDF reprints. By virtue of the underlying Webglimpse search software, every paper can therfore conveniently be recalled and annoted, domain specific PDF collections of whole research groups can be turned into powerful knowledge databases.

One will not have to change his way of literature research for BibGlimpse, but BibGlimpse will definitely change how one benefits from the papers he and his collaborators have found. For a demo PDF archive and more details about the sophisticated Medline retrieval please see the BibGlimpse supplement page. In case you encounter any difficulties using BibGlimpse please contact Thomas Tuechler, bibglimpse08[at]boku.ac.at.