A research project to develop a bibliography management tool focusing on literature search, and automating repeatable tasks, for advanced typists and small visual displays.
Literature search is both fun and time consuming. The amount of knowledge became huge. Even mathematicians research into computational proofs, and into dedicated languages to help find relevant contributions through clever algorithms. Once found, having to store them locally appears reasonable.
This little tool helps to
- keep track of all literatures relevant to my research interests across many LaTeX projects, using a tree of tags/categories,
- importing into the local database is almost automized, using a shortcut and copy+paste of a DOI/arXiv/HAL, followed by a confirmation shortcut,
- importing BibTeX manually with snippets and autocomplete (and detect correlations to existing bibliographies), and generate a unique identifier,
- quickly search for bibliographies, by a mixture of simple search, category filtering (with custom syntax), fulltext search, and explicit field search (title, author etc.),
- keep up to 9 different search result sets in memory,
- quickly preview a bibliography, with fulltext search results (similar to Google Books), a list of comments found within the PDF documents, the contents of annotated text files (sometimes used for extended summaries/analysis/etc.), and a graph of references (which is unfortunately not trivial to automize freely),
- quickly exporting bibliographies to projects using a gather/retrieve approach (a command line program reads your project BCF file along with the locally central bibliography index), populates BibTeX files as defined by your documents, and shows the protocol (exported+missing identifiers),
- quickly work on notebooks (in Linux), at home, in libraries, at airports, in the wild (no reliability on the availability of the internet),
- trust the contents, being verified/previewed upon importing, to remain as is (no internal faulty conversions),
- work in small teams or individually having a shared index (for the latter: the combination of text based backend and GIT is industry proven; there is also an sql backend),
- restore program state of principal variables between program executions,
- experiment with compact, simple and computationally efficient GUIs (not my core research focus).
This project is open for collaboration. If you feel to help with developing, with testing, with documenting, with discussing future directions, please get in touch with me.