Bibliothèque nationale de Luxembourg
Amsterdam 16.9.2013
eluxemburgensia
Yves Maurer
www.bnl.lu 2
www.eluxemburgensia.lu
Archive of historical newspapers 11 titles 60 000 issues 400 000 pages
Data scanning Timeline 2003 – 2007 : Scan images only 2005 first trial with METS/ALTO Since 2007 : METS/ALTO
www.bnl.lu 3
Drivers for newspaper UI
Existing functionality should be kept Pick by date View images, zoom, next page, next issue etc.
Internal Driver: METS/ALTO Full text search in whole collection Search in obituaries, image captions etc. Articles should be an entity
Site survey ANDP Paperspast ...
www.bnl.lu 4
Search - challenges
What do you search in? www.eluxemburgensia.lu (Digitool)
is based on issues www.a-z.lu (Primo)
is based on articles Bad OCR Three different search engines Poor metadata, rich fulltext -> snippets
www.bnl.lu 5
Search interface 1
www.bnl.lu 6
Search interface 2
www.bnl.lu 7
Search interface 2
www.bnl.lu 8
Viewer - challenges
Need to handle: Table of contents Search Article highlighting Show OCR text Browse thumbnails Navigation in the page (zoom, drag, gestures, article reading mode, ...) Deep links to articles and pages
Technical constraints Speed No flash Browser compatibility Integration with Digitool
www.bnl.lu 9
Viewer UI
www.bnl.lu 10
User feedback
Well received Especially from expat population and researchers
Wishlist – general patrons User OCR correction Embeddable viewer for blogs User collections with annotations Android app
Wishlist – researchers Access to full text data as linked data set Printing of individual articles Simple “microfilm-like” viewer for efficient viewing of entire collection
www.bnl.lu 11
Open source viewer
URL: http://sourceforge.net/projects/bnlviewer/
Used by http://www.periodika.lv/
BSD License