This document lists the major additions, changes and fixes since the release of refbase-0.9.0. See also the 'NEWS' file that comes with the refbase package.

refbase-0.9.5

Feature additions

Import/Add Records

  • Added support for resource types Abstract, Conference Article, Conference Volume, Magazine Article, Manual, Miscellaneous, Newspaper Article, Patent, Report, Software
  • Added import capabilities from SciFinder and Endnote XML
  • refbase now supports direct import from arXiv.org & CrossRef.org: enter one or more arXiv IDs, DOIs or OpenURLs, and refbase will fetch & import the corresponding record metadata
  • Added support for direct upload of references from Bookends

Export/Citation

  • Added support for new export formats: ADS, Atom XML, ISI Web of Science, OAI_DC XML, SRW_DC XML, and Word XML
  • Added new citation styles: AMA, APA, Chicago, Harvard (3 variants), MLA, and Vancouver
  • refbase can now output citations in LaTeX bibliography (.bbl) format
  • It is now possible to cite, group, or export ALL found records (instead of only the selected records)
  • When outputting citations as PDF, US letter is now supported as page format
  • Headers can now be added to cite formats via the headerMsg URL parameter
  • Links to corresponding files are now included in the export data if the export format and user permissions allow so

Search/Display

  • refbase sports a refined interface:
    • Revamped the refbase main page: all forms are now styled via CSS, and citations of the most recent publications are now dynamically included upon page load
    • Redesigned the page footer & header as well as the forms on the results pages
    • The "Quick Search" form is now always available in the page header
    • refbase now makes use of the Prototype & script.aculo.us JavaScript frameworks to offer auto-completions (or search suggestions) for text entered by the user in text entry fields of the refbase interface
    • Forms on search results pages can now be hidden (and if enabled by the admin, a sliding effect is now used when a user toggles the display of these forms)
    • Search results can be browsed & searched in three different views: List view, Citations, and Details
    • Additional record info (keywords, abstract, etc) and links to cite/export the record can now be displayed underneath each citation
  • refbase now supports OpenSearch which allows clients to search refbase and retrieve results in a standardized manner; as an example, this enables Firefox 3 & IE 8 users to search refbase directly from their browser's toolbar, and the browsers can now offer search suggestions for search terms entered in the browser's search box
  • Initial support for searching & managing of duplicate records
  • For the current session, refbase now offers a query history with links to any previous search results
  • The show.php API now allows to include results from a refbase server into other web pages; this makes it possible to dynamically include results from a refbase database into static HTML pages (using an iframe or AJAX-style JavaScript)
  • Added a "main fields" search option to the "Quick Search" form which allows to search all of the main fields at once; the list of "main fields" can be specified globally as well as individually for each user
  • The default number of records per page can now be specified globally as well as individually for each user
  • Added initial support for output to mobile devices (viewType=Mobile)
  • Many more minor enhancements

Localization

  • Added initial support for spanish and chinese localizations

Command Line Clients

  • The 'refbase' command line client now supports the new export formats atom, isi, oai_dc, srw_dc and word as well as --view=Mobile and --order=creation-date
  • 'refbase' command line client: added options -A|--append and -B|--update which let you append found records to a local BibTeX or MODS/SRW XML file (if they don't yet exist in that file), and update existing records in that file if their modification date on the server is more recent; by default, a backup file is created before adding or changing anything in the local file
  • 'refbase' command line client: added -E|--extract option which extracts citation IDs from a file and retrieves matching records from refbase (supported file types: .aux, .bbl, .bib, .enw, .ris, .tex, and (MODS|SRW|Endnote) .xml)

Miscellaneous

  • Added a refbase extension for MediaWiki which allows to place a serial number within <refbase>...</refbase> markup tags; metadata for the reference is retrieved from refbase and marked-up according to a MediaWiki citation template
  • Added support for the sitemaps.org protocol which allows for better indexing by search engines
  • The refbase unAPI service now supports the new export formats atom, oai_dc and srw_dc
  • Many more minor enhancements

Changes

  • Introduced measures to better prevent cross-site scripting (XSS) attacks
  • For security reasons, HTML tags are now stripped from the headerMsg URL parameter; string formatting can now be done using the refbase markup syntax
  • Improved support for PHP installations running with safe_mode=On
  • By default, export of data via the refbase GUI is now enabled for anonymous users
  • The citation output options (previously located in the results footer) have been moved to the Display Options form of the results header in Citation view
  • refbase now remembers many more search & display options while navigating or refining the results list
  • Changed the HTTP transfer method in most forms from POST to GET; this helps to avoid warnings about re-sending POST data in some browsers (such as Firefox)
  • We now use a show.php URL for the "show details" link in the search results list since it is much shorter and easier to bookmark as a permanent link
  • When linking to particular records via show.php, one can now use consecutive serial number ranges as values of the records parameter; examples: show.php?records=123-131 or show.php?records=123-141,145,147,150-152
  • When viewing/outputting citations, the list of citations can now be sorted by record creation date (use citeOrder=creation-date in show.php URLs); this allows to sort records such that newly added/edited records get listed top of the list
  • Redesigned the page footer to make it fit on one line (and thus less heavy)
  • Added many CSS attributes (id/class names) to HTML output where appropriate; this allows to easier refine the CSS styling of the refbase interface
  • The admin can now define any of the supported views (List view, Citations, Details) as the default view
  • For each view, the admin can now define whether the results header & footer should be displayed open or closed by default, or whether they should be entirely hidden from the interface
  • The admin can now customize the list of fields that are displayed by default in List & Details view; similarly, the list of fields that are available in dropdown menus of the results header can also be customized
  • The admin can now define accesskey values which allow for keyboard navigation of the main parts of the refbase interface
  • The $databaseBaseURL in ini.inc.php is now auto-generated by default
  • The refbase logo image files have been updated to the new refbase logo, and the logo URL path & dimensions can now be customized easily via ini.inc.php
  • The session/temp dir path can now be specified explicitly in ini.inc.php; this allows you to define a custom directory path that's used on your server to save session data and to write any temporary files
  • The edition field now accepts string values
  • After any add, edit, or delete action, refbase now includes a link to the last multi-record search result list in the feedback header message if the previous query resulted in a single record; together with other changes, this obviates receipt.php
  • By default, the SRU formats SRW_DC XML and SRW_MODS XML are now transformed to HTML via XSL stylesheets
  • Authenticated users can now use opensearch.php or sru.php to query any of their user-specific fields
  • For OpenSearch & SRU searches, the keywords field will now be queried by default if no context set & index name is given ("cql.serverChoice")
  • Renamed the Bibutils xml2word command line tool to xml2wordbib according to the name changes in Bibutils v3.40; note that if you're using a Bibutils version prior to Bibutils v3.40, you must revert this change in file export/bibutils/export_xml2word.php
  • Underline fontshape markup is now supported in various import and citation formats as well as in refbase markup (use: __underlined text__)
  • When importing multiple records, ALL records are now imported by default
  • Improved UTF-8 handling on import so that e.g. a best-effort conversion is done for UTF-8 data which are imported into a latin1-based database
  • When importing (or exporting) RIS records, the type of thesis (such as "Masters thesis" or "Ph.D. thesis") is now taken from (or exported to) the U1 field (as is supported by Bibutils v3.35 or greater);
  • Upon import, if the given URL actually is a DOI prefixed with http://dx.doi.org/, we'll extract the DOI and move it to the doi field
  • Import of user-specific fields into the "Add record" form is now possible
  • We now generate (or honour existing) cite keys/IDs for the import data, and copy them to the user's cite_key field; the uniqueness of cite keys is ensured among all imported records as well as the user's existing records
  • Improved recognition of PubMed resource types
  • For the "Extract citations" functionality, we now check whether the extracted serial numbers and cite keys exist in the database and report any missing record identifiers
  • Changed or merged some localization strings to allow for better translations to languages such as Spanish or Chinese
  • Refined layout of user_receipt.php
  • The 'refbase_import' command line client now supports common output options from the 'refbase' CLI client; this allows to instruct the refbase server to return all imported records in any of the supported export or citation formats & styles
  • Many more minor changes

Bug fixes

  • Fixed bug that prevented correct directory creation/renaming and/or file upload on some platforms
  • Fixed preg_match() compilation errors when importing BibTeX records into a refbase UTF-8 database
  • When importing PubMed MEDLINE source data, we now extract author information from the AU field if the FAU field is not available
  • We now convert Endnote XML text style markup into appropriate refbase markup
  • Upon RIS import, we now make sure that HTML encoded source data (such as &auml;, &#xF6; or &#233;) get decoded before import
  • Records of unrecognized resource type were omitted upon citation output; fixed
  • When outputting to LaTeX or RIS, curly brackets are now escaped to avoid incorrect output
  • Fixed an issue were it wasn't possible to correctly import (or cite) records which contained non-ASCII characters in author's given names
  • Worked around an error in Internet Explorer when importing single records via the web interface
  • Fixed an error when exporting data to ODF XML and when the user-specific fields were missing from the SQL query
  • Fixed referrer-related problems that could cause unexpected behaviour and incorrect warning/error messages; however note that working simultaneously in different browser windows or tabs can still cause some unexpected results
  • Worked around PREG compilation errors on some UTF8-based systems
  • For theses, the corporate author (i.e. the corporate body granting the degree) is now labeled correctly in MODS XML output; as a result, the degree grantor is now recognized correctly by Bibutils and copied to the BibTeX school field
  • Many more minor bug fixes