This is a guest post from LIBER sponsor RetroNews; written by Nathalie Thouny and Étienne Manchette.
To speed up the digitisation of the press and make press publications all the easier to access, BnF-Partenariats¹, the French National Library’s subsidiary in charge of advancing its digital legacy, has designed and developed, with the support of the French National Fund for a Digital Society, a new on-line service entirely dedicated to the press of the past: retronews.fr. For the past two years, the site has made it possible for interested people to view millions of archived press sources, originally published between 1631 and 1945. In so doing, they can dive into history as experienced by its contemporaries.
A Bit of Background
The French National Library, or BnF, has a special part to play in the digitisation policy of the collections it houses, in particular the press collections, especially valuable in that they are a testimonial to the events and topics that stirred each era.
Determined to make this heritage visible to the widest possible audiences, the BnF implemented a mass digitisation policy and, to extend its reach even further, decided to offer open access to its archives, both on and off premises, through its digital library Gallica — celebrating its 20th anniversary this year.
Digitising The Press: A Priority Issue
The press collections at the BnF fill 44 linear kilometres of warehouse space and are growing annually, as a result of legal deposit requirements. In 2016, 16,000 newspapers were deposited at the BnF.
They form a substantial and invaluable resource, part of which is now vulnerable to deterioration. The ink and paper are giving way to time and digitisation has, for too long, been infrequent.
The digitisation programme is thus a key component of the library’s core endeavour. Several hundreds of press publications are being digitised and published online. Within two years’ time, it is predicted that RetroNews will offer access to nearly 20 million pages of press archives.
Selecting Sources: The First Step In Building An Editorial Identity
A Bird’s Eye View of Yesteryear’s PressThe sheer breadth of the collections published since the start of the 17th century, when the first periodicals came into being, made selection a necessary first step before embarking on the process of digitisation.
Spanning the period from 1631, when La Gazette was founded by Théophraste Renaudot, to 1945, RetroNews offers viewing access to 300 press publications: a large number of general news dailies and weeklies – including Le Petit Parisien, Le Journal, Le Figaro, Le Matin and Le Gaulois – as well as political – L’Humanité or L’Action française –, regional – La Petite Gironde, Le Petit Marseillais, Le Petit Troyen – and satirical – such as Le Charivari, Le Rire or La Caricature papers.
Selecting publications for RetroNews was thus already an editorial act.
First and foremost, it implied ensuring in-depth coverage for each period. It also meant carving out the exploration paths, discovery routes and research avenues that would go on to benefit all types of users. The selection of newspapers published online each month, augmented with reference fact sheets about the history of the press, is another response to that same set of imperatives. By offering a searchable base of publications covering all periods of history, different types of press, political movements and geographical zones, it provides readers and researchers with a structured corpus offering multiple inroads.
Semantic Enrichment & Search Tools: Essential Stepping Stones In Opening Access To Experts
The second challenge which RetroNews had to address – and one just as fundamental – was the question of how to offer seamless, dedicated access to the resources.
This implied developing advanced search features to explore that content. Incorporating both indexing and viewing technologies, the site was developed to serve the needs of researchers, academics and, more broadly, all users, with its expert search engine.
Research on data mining was conducted to improve character recognition techniques and create lexicons enriched by repositories such as Rameau, Vapereau or VIAF². Applied to the archived collections, each of the aforementioned processing tools facilitated the large-scale extraction of semantic entities such as places, individuals, organisations, events. The addition of IPTC, the thematic classification tool used by all major media worldwide, brought a further angle to the experience, putting topics into perspective across history.
With its multi-faceted search filters (by theme, person, event, place, organisation, or historical era) applicable to the press collections, and its dedication to highlighting the features specific to them, in typography, writing and format, RetroNews makes it possible to explore the events, large and small, that have shaped History.