Digitising Journals on Trove

A national approach to sharing content, engaging communities and collaboration in the digital world

When

Conference

VALA 2018

Author: 
Hilary Berthon Assistant Director, Trove Outreach
Julia Hickie Assistant Director, Trove Data, Discovery and Delivery

Publication date
Thursday, 22 March, 2018

Abstract: 
Trove’s success is underpinned by the availability of a significant quantity of digitised material achieved through partnerships between the National Library of Australia (National Library) and many other organisations. The National Library is using the lessons learned through earlier digitisation collaborations, as well as a history of content aggregation, and has now turned its attention to journals. Existing processes for digitisation project management and online discovery have been expanded, with new delivery and browse features introduced to facilitate new ways of navigating content. The result has been rapidly growing usage and the engagement of new communities.

Journals digitisation

We know from the enquiries we receive, from the new works that are created, and from the stories people told us on the 2017 Trove Roadshows, that digitising material and making it available in Trove gives Australians unprecedented access to their cultural heritage.

While undertaking research for this paper, the authors were able to identify the impact that full text discoverability has on use. In May 2016, after incorporating full text discovery for digitised journals, usage of journals in Trove suddenly tripled (Figure 1). That one change marked the beginning of the expansion of the journals digitisation program, and a new chapter in Trove's history of facilitating equal online access to all Australians, no matter who they are or where they live.

The journey taken to get to that point is outlined below.

Journal page views before and after full text discovery was introduced
Figure 1: Journal page views before and after full text discovery was introduced in May 2016

Using a process that works

The National Library’s approach to digitising journals builds on its previous experience with mass digitisation of newspapers, but also on the innovations and efficiencies gained as part of the Digital Library Infrastructure Replacement Program that has laid the foundations of the National Library’s virtual library. In June 2017, the National Library successfully completed its Digital Library Infrastructure Replacement Program, positioning it to meet its collecting obligations, continue a tradition of innovation and meet the challenges of being a memory institution in the digital age. This investment has transformed how users discover, consume and engage with the National Library’s vast digital collection, which has tripled in volume since the program to modernise the National Library’s digital library infrastructure commenced in 2011. The program has ensured that systems can scale to meet expected levels of collecting, support new content types and meet digital preservation requirements. It has enabled the National Library to collect and manage its digital collections securely and accountably, delivering innovative new services that improve national reach and draw in new audiences and improve productivity for collecting and collection management areas using a more sustainable IT infrastructure. The National Library now has flexible systems to support outsourced and mass scale digitisation projects.

The National Library works with a wide range of partners, from universities, historical societies, the private sector and other libraries on journal digitisation projects. The approach to determining which journals will be digitised is based on use, historical significance and geographical or regional coverage, and those journals most at risk of loss. Often these decisions are made collaboratively with partners, and examples of these collaborations are detailed later in this paper.

The procedures established for setting up projects with digitisation partners—specifying requirements for both digital capture and processing captured images into searchable text, quality assuring and transporting both source materials and files—learnt during the newspapers digitisation program have been extended to journals digitisation.

The mass digitisation of journals relies on the efficient integration of a sequence of processes, commencing with an initial pre-scanning survey. This work is undertaken within the National Library’s Digitisation and Document Delivery Branch. This includes assessing the material’s physical condition: its fragility and any preservation requirements, its size, number of pages, the tightness of binding, any changes in format or the presence of fold-outs. This work also involves identifying any supplementary material such as special issues or indexes, or extraneous material bound together with the journal of interest, and any obvious missing pages or issues. This survey allows an assessment of the best equipment for digital capture and may indicate that a search for additional or alternative source material for the digitisation project is needed.

An analysis of the content—its layout, and any changes in layout through the date range of interest—enables optical character recognition (OCR) and content categorisation requirements to be defined and will indicate the possible need for a pilot project to be undertaken for a particular title. For example, the digitisation of ‘Art in Australia’ included requirements for linking illustrations with captions as well as sequencing and linking content. Assessing the bibliographic record and noting any journal name changes are also preliminary steps required to ensure the smooth progression of the project. In addition, the National Library works with partners to locate and source copies so that, as much as possible, the full run can be digitised. This quite often means bringing copies from different holding institutions together: re-uniting issues online is one of great benefits of digitisation.

The National Library’s work with processing and delivering non-English language newspapers has paved the way for digitising and delivering non-English language journals. Trove’s digitised newspaper collection includes a modest range of Roman script languages, such as German, Italian, French and Estonian, as well as Chinese script. To date, Trove has one journal containing non-English language: ‘Inside  Indonesia’ (1983-2007).

In contrast to the National Library’s newspaper digitisation program, where digital capture is generally from microfilm, scanning of journals material is generally undertaken from print. The National Library has a Qidenus Robotic Book Scan 4.0 (Qidenus Technologies, 2016) which enables efficient capture and cropping. Image files are captured in colour and 400 dpi resolution. Each image will generally contain one physical page, although there may be exceptions where it is not desirable to split a two-up image, due to the layout of the page. Quality assurance, both by the operator and by peer-review, occurs following capture. Errors that occur may vary from missed or duplicate pages to image problems such as reflection due to page curvature. The scanning quality assurance stage also provides an opportunity to adjust cropping and insert pages, such as foldouts, that may have been captured on other equipment.

After the image files are loaded to the backend content management system against a catalogue record, the files are then sent to two places, one to generate derivatives for the delivery system and the other for OCR processing. The National Library uses the software program docWorks to convert its scanned journal images into searchable and categorised content for enhanced access through Trove. When OCRing is completed, the output is ingested back into the content management system ready to be indexed. Since mid-2015, the National Library has engaged OCR contractors using this software to deliver OCR output using the Remote QA (RQA) module. In the National Library’s set-up, all the master image files, along with their associated bibliographic and descriptive metadata are held within the Library’s digital content management system. The contractor generates the OCR output (METS and ALTO)[i] files, which are ingested back into this system, matched with the master images and indexed by Trove. The OCR and content analysis requirements cover aspects such as identifying and zoning articles, providing OCR text for each page and article, manually re-keying text for headlines, subheadings and authors for each article, and providing a valid METS xml file for each issue and a valid ALTO xml file for each page. Through a combination of automated processes and manual checking, categorising, ordering and grouping elements such as advertisements, illustrations, captions, tables, footnotes, headlines and authors also takes place. At the journal issue level, the structural elements of the main parts of each issue, such as cover, title and table of contents pages, are identified. A generic specification may be supplemented with additional instructions for particular journals. Quality assurance of pilot issues is a vital aspect of this process, ensuring that any systematic zoning errors are addressed. The National Library uses bespoke digital object storage systems with separate file systems for holding digital object metadata, for preservation, display and working content.

Digitising at Scale

As with newspapers, the National Library has discovered that there are considerable efficiency savings made possible by digitising journals on a large scale. These include taking advantage of volume discounts available from the panel of suppliers that the National Library works with, and the retention of staff (both permanent and contract) with the necessary skills to undertake this digitisation.[ii] This, combined with the added value realised through building a significant corpus of digitised material that is national in scope and allows in-context discovery, makes the large-scale production and delivery of digitised content in Trove attractive.

The digitisation process produces image files (masters in TIFF format, access copies in jpeg2000), OCR data (METS and ALTO files) as well as internal system data. The total storage for all digitised journals (excluding newspapers) currently comprises 35 Tb, with three copies of each digitised file kept. Additional backups are maintained by the National Library in offsite tape storage.

Making Journals Discoverable in Trove

Access to full text journal content represents an emerging and growing element of Trove. Initially digitised journals were discoverable in Trove in the same way as a physical journal, through a single MARC record describing the journal title. Full text searching of the content, processed through optical character recognition (OCR) software, was not possible.

In 2014, in addition to hosting digitised newspapers, Trove started hosting other digital collections, a component of a much larger, newly developed Digital Library Collection (DLC) system at the National Library.

The storage of OCR full text in a back-end collection management system did not automatically enable search in a public-facing system such as Trove. A number of different models were considered, based on the existing digitised textual content within Trove. As the National Library’s current Director-General, Dr. Marie-Louise Ayres, noted in 2012, Trove was conceived when the National Library’s ‘habit of creating new specialist discovery services was catching up with us.’(Ayres, 2012) The idea of creating custom functionality for an additional collection, with the requisite maintenance activities and multiplicity of discovery locations, was not sustainable.

Taking advantage of Trove’s existing ingest mechanism to facilitate full-text discovery

As digitised journal content started to build up through partnerships, the pressing question for staff at the National Library was how to facilitate this full text searching of digitised journal content.

Creating a solution that would amalgamate this material with similar aggregated collection material, hosted outside the National Library, was considered very attractive. Since its launch in 2009, Trove has been an aggregator, a place to find and annotate (add tags, comments, create lists and make text corrections) Australian cultural collections in one place. As of 1 August 2017, Trove directly aggregates 328 collections, from 227 organisations, with more libraries coming into Trove through Libraries Australia. A piece of National Library-developed software called the National Library Harvester facilitates direct contribution, copying metadata records from external web services, transforming the metadata into Trove’s form of Dublin Core and loading those records into Trove. An earlier project had already pioneered the inclusion of full text alongside these metadata records, enabling this content to be searchable in Trove (Hickie & Raadgever, 2015).

With the development of a new Application Programming Interface (API) ingest module, any Representational State Transfer (RESTful) eXtensible Markup language (XML) or JavaScript Object Notation (JSON) API could be accessed by the National Library Harvester. The process starts with the National Library Harvester making an http request to the journal article repository at a regularly scheduled time each evening. The repository’s JSON API responds with metadata such as article title and full text. This update occurs incrementally, returning all digitised journal articles that have been added, updated or deleted since a given timestamp. These digitised journal articles are then ‘cross-walked’ into Trove’s form of Dublin Core, using Java regular expressions and eXtensible Stylesheet Language Transformations (XSLT) stylesheets. Finally, the records, which include metadata as well as full text, are loaded to Trove. As an example, within the almost 50,000 digitised articles of 'The Australian woman's mirror', an Australian weekly women's magazine published between 1924 and 1961, a patron can undertake keyword search for Australian literature, knitting patterns, places and people’s names.

New search and browse features

The delivery of journals on Trove incorporates new and enhanced search and browse features. All of the digitised images are in colour and can be viewed with two pages side-by-side, with page turning functionality, more akin to a physical journal reading experience than Trove’s newspapers viewer offers. A tiled thumbnail view of the individual pages makes it possible for someone to visually scan through the issue. Deep zoom capability means page images can be examined in great detail.

Searching within a single issue happens while the page image remains in view and the OCR text remains alongside. Small markers drop along the bottom of the screen to indicate where matches for the search term have been found within the issue. Clicking on the markers allows users to jump between references on different pages. A hyperlinked table of contents is presented alongside the page image (Figure 2). Users see a list of articles in the issue and can click on an article or page number to view it.

Table of Contents
Figure 2: A Table of Contents on the left hand side allows users to browse within an issue by article

In developing Trove’s journals delivery system, browsing by issue or edition as well as searching by keyword had been highlighted by staff as critical to journals. A sequential browse window was built that allows users to view 20 issue covers at a time (Figure 3). Users can view basic metadata for each issue, including the publication date and volume/issue/edition number. They can paginate through batches of 20 until the desired issue is located, then click on the cover image to view that issue.

New browse functionality
Figure 3: New browse functionality gives users a view of issue covers and basic metadata

Discovery by full text caused usage to triple

The usage statistics show that enabling full text discovery of journals dramatically increased use (Figure 1). May 2016 was the first month journals became discoverable at the individual article level, which included searching of the full text, rather than being discoverable through a bibliographic record at the journal title level. In that month journal usage tripled, jumping from 6,768 page views in April 2016 to 19,074 page views in June 2016.

The size of the digitised corpus had grown from 69,234 articles in May 2016 to 320,090 in May 2017, a growth of 4.6 times. Over the same period usage increased 17-fold, and in April 2017 exceeded 100,000 page views in a month for the first time. This reflects that discoverability dramatically increases use of the digitised resource.

Replicating Trove's popular search and delivery features for journals

Two of the features that have made Trove’s digitised newspapers so popular, full-text search and discovery at the article level, have been replicated for search and discovery of journals in Trove. A user types in a search term, discovers an article record that contains this search term in the metadata or the full text, then clicks through to view the digitised journal issue. At this point they encounter the books and journals delivery system. The online book reader opens on the selected issue, then automatically scrolls to the first page the desired article appears on. The user’s search term is highlighted on the page image and the OCR text appears alongside. Other replicated functionality includes the ability to see metadata about the issue and copyright information, to get citation information, to download a copy or order a high-resolution copy from the National Library's Copies Direct service, and to share the item on select social media platforms.

Working with partners

Working with a range of partners has enabled the National Library to achieve the “broad access to content”, ranging from international and national to local content, over a range of subject areas, which Calhoun (2014, p.146) identifies as a key social contribution of digital libraries (Calhoun, 2014).

The initial step the Trove Outreach team makes in working with prospective partners who approach the National Library is to establish their requirements, to check that they are the rights holder and are happy to give the National Library the relevant permissions to digitise the journals and make them freely available on Trove. The National Library will then negotiate an agreement with the partner that describes the work to be undertaken and specifies the roles and responsibilities of each party.

The digitisation of ‘The Bulletin’ magazine has been achieved in partnership with the State Library of NSW, which provided access to its source material, and AustLit, the online research and publishing tool, which identified and made available its lists of literary content. Digitally captured by one of the Library’s scanning contractors, multiple copies were assessed and the best copy selected for digital capture. The National Library managed and funded the digitisation of the first six years (1880-1885) of this quintessentially Australian publication, launching this content in early 2017. Digitising the rest of 'The Bulletin' is part of the program of work being undertaken by the National Library with support from the Commonwealth Government’s Modernisation Fund and also with continuing support from the State Library of NSW and AustLit. The existence of AustLit’s lists of literary content contained within ‘The Bulletin’ has enabled identification and encoding of this content at the OCR processing and content analysis stage of digitisation. This has enabled the possibility of providing patrons with the ability to do specific searches for this content in the digitised Bulletin on Trove.

Through its Digital Excellence Program, the State Library of NSW has provided funding for the digitisation of many additional journals to Trove. Among the outstanding journals brought into Trove through this partnership are the ‘Building’ magazine (1907-1942), ‘The Home: an Australian Quarterly’ (1920-1942), ‘Decoration and Glass’ (1935-1949), the ‘Australian Women’s Mirror’ (1924-1954), ‘Pix’ (1938-1954), ‘Australasian photo- review’ (1899-1954) and ‘Photographic Review of Reviews’ (1894). In a similar partnership arrangement, the ‘Murray’s Austral-Asiatic Review’ (1828) was funded by the Tasmanian Archive and Heritage Office (TAHO).

A one-off partnership arrangement with the University of Wollongong Library and the University of NSW Library, both of which provided source material and scanned images to the Library’s specifications, resulted in the early twentieth century art magazine, ‘Art in Australia’ (1916-1942), being made permanently available online through Trove.

A partnership with the NSW Rural Fire Service brought sixty years of the 'Bush Fire Bulletin' (1952-19751978-2016) to Trove. Not only did the Fire Service fund this digitisation project, but it also provided its own copies to fill gaps in the National Library’s collection.

The Mount Isa Mining Magazine, 'MIMAG' (1947-1996), a record of the mines, miners and residents of Mount Isa from 1947 to 1996, found its way into Trove through a joint project with the Mount Isa City Library, which received funding from the State Library of Queensland, and Mount Isa Mines – Glencore, which provided permission for the magazine to be available in Trove.

'Inside Indonesia' was digitised to Trove in a partnership between the National Library and the publisher, with funding from the Herb Feith Foundation and the Australia Indonesia Institute. As well as including some Indonesian language, this magazine has a wide variety of layouts, requiring specific instruction on content categorisation as well as an Indonesian language plug-in at the OCR processing stage of digitisation.

A highly respected and significant journal covering Pacific affairs during the period 1930-2000, 'The Pacific Islands Monthly', has become one of the few research resources from the Pacific region that has been made available online. This was supported by the publisher, the 'Fiji Times', and funded by the National Library.

‘The Newcastle and Maitland Catholic Sentinel: the official organ of the diocese of Maitland’ (1931-1968) and the Catholic Bathurst diocese’s ‘The Record’ (1877-1884; 1929-1931) were added to Trove in 2016. 'The Sentinel' contains a mixture of information and photographs about the diocese as well as news of Newcastle and its surrounds, international news and editorial content. It provides an interesting window into social attitudes and records the dates of significant events such as the changing of Mass from Latin to English (Lynch, 2016).

Communities drive usage

The desire for people to connect with their own stories is reflected in the popularity of particular journals, with communities congregating around particular titles. This is reflected in usage statistics, where the number of page views does not necessarily correspond to the number of digitised articles.

Title Number of digitised articles
Pacific islands monthly : PIM 95,343

The Australian woman's mirror

48,793
Building : the magazine for the architect, builder, property owner and merchant 26,721
Pix 23,354
Uniken 18,827
The Newcastle and Maitland Catholic Sentinel : the official organ of the diocese of Maitland 18,159
Australasian photo-review 16,093
The Bulletin 15,043
The Home : an Australian quarterly 12,329
Building and engineering 8,108

Table 1: Top 10 journals by number of digitised articles (Content) as of 31 July 2017

For example, it seems probable that former Mt Isa residents, now dispersed across Australia and the world, are using 'MIMAG' as a way to connect. Indeed, this was a motivation for the digitisation project (Berthon, 2016). This has resulted in 3,279 page views of the homepage and 6,361 page views of individual journal issues up to July 2017 (Tables 2 and 3). With this community interest, 'MIMAG' is the third most visited title. 'MIMAG' has 5,184 articles and does not rank in the top ten journals based on size.

Title

Page views of journal homepage

April 2016 – July 2017

The Home : an Australian quarterly

4,826

Building : the magazine for the architect, builder, property owner and merchant

3,631

Pacific Islands Monthly: PIM

3,586

MIMAG/Mount Isa Mines Limited

3,279

Art in Australia

2,509

The Bulletin

1,256

Inside Indonesia : bulletin of the Indonesia Resources and Information Programme

1,131

Decoration and glass

1,005

Bush fire bulletin

991

The Newcastle and Maitland Catholic Sentinel : the official organ of the diocese of Maitland

913

Table 2: Top 10 journals by views of home page (Usage)

Title

Page views of journal issues

April 2016 – July 2017

Pacific islands monthly : PIM

12,651

Building : the magazine for the architect, builder, property owner and merchant

 9,099

MIMAG/Mount Isa Mines Limited

6,361

The Home : an Australian quarterly

4,555

Bush fire bulletin

3,642

Art in Australia

2,781

Anzac bulletin : issued to members of the Australian Imperial Forces in Great Britain and France by authority of the High Commissioner for Australia

 2,425

The Newcastle and Maitland Catholic Sentinel : the official organ of the diocese of Maitland

 2,168

Decoration and glass

1,862

The Bulletin

1,498

Table 3: Top 10 journals by page views of digitised content (Usage) 

Title

Date journal appeared on Trove

Anzac bulletin : issued to members of the Australian Imperial Forces in Great Britain and France by authority of the High Commissioner for Australia

 5 September 2014

Inside Indonesia : bulletin of the Indonesia Resources and Information Programme

 12 May 2015

Uniken

13 January 2016

Building : the magazine for the architect, builder, property owner and merchant

 21 April 2016

Bush fire bulletin

16 June 2016

Building and engineering

26 July 2016

MIMAG/Mount Isa Mines Limited

1 August 2016

The Newcastle and Maitland Catholic Sentinel : the official organ of the diocese of Maitland        

 8 August 2016

Pacific islands monthly : PIM

15 September 2016

Art in Australia

7 November 2016

Decoration and Glass (1935-1949)

5 January 2017

The Home: an Australian Quarterly (1920-1942)

25 January 2017

The Australian Woman’s Mirror (1924-1954)

11 April 2017

Australasian photo-review (1899-1954)

21 April 2017

The Bulletin

3 May 2017

Pix (1938-1954)

8 June 2017

Table 4: Date that journal's full-text became keyword searchable on Trove

Community impacts: drawing together communities and exposing collections to the world

According to Calhoun (2014, Ch. 6), “conventional thinking has tended to emphasize the collections of libraries over their societal or community based roles.” In this context, it is interesting to trace the impact on communities of making digitised journals available on Trove. The 'Bush Fire Bulletin' is an example of how providing access to digitised content on Trove can draw a community together: in this case the dispersed workforce of the world’s largest volunteer fire service, the NSW Rural Fire Service (RFS).

According to Senior Librarian with the Fire Service, Katie Vandine, interest in digitising this magazine was heightened as many of the brigades celebrating anniversary years were looking for archival content. She said the service hoped that this digitisation project would be received as a gift to its 74,000 volunteer members. An unintentional consequence of having this magazine online has been its use by the Fire Service in preparing to publish new issues of the 'Bush Fire Bulletin,' where the archive is consulted frequently for modern day articles detailing the particular fires that led to the creation and growth of individual brigades. The potential to reach even wider audiences was highlighted by NSW RFS Commissioner Shane Fitzsimmons who, as the retrospectively-digitised issues came onto Trove, noted that “this history of the NSW RFS will be available for all of Australia and the world to see.’’

The digitisation of 'Art in Australia' has enabled free online access to the full publication in high resolution to researchers and art-lovers alike, and has resulted in some interesting connections being made, demonstrating Trove’s role as a ‘facilitator of conversations’ (Lankes, Silverstei, & Nicholson, 2007). For example, the University of Wollongong was contacted by the son of one of the artists represented in the magazine, who had been trying to locate one of his father’s paintings without success. The University received another communication from a researcher who had been studying the work of Sydney Ure Smith, who, with Hartland and Hyde (photo-engravers), devised a method of including quality coloured reproductions in their periodicals. “Having them digitised on Trove is actually better than the paper originals as I can often magnify the images to such an extent I can make more precise judgements about the quality of the 'Art in Australia' reproduction as compared with the surviving originals. I've learnt a lot about which colour separations worked best back then and which did not.”

Digitised journals have the capacity to draw together communities of readers and researchers. According to Monash University Research Fellow and Chair of the publishing board of 'Inside Indonesia', Jemma Purdey, the impact of the 'Inside Indonesia' digitisation project for the Indonesian Resources and Information Program Inc., the publishers of the magazine, has been significant. “As a publication of more than three decades standing, this archive represents two generations of critical writing and reporting on Indonesia and Australia's relations with Indonesia. The archive serves to ensure that this resource is both preserved and made accessible to readers and researchers across the world.” As well as providing an invaluable source for research and teaching, having the journal online on Trove has facilitated access by the journal’s growing overseas readership, especially in Indonesia. Purdey says,

“We were aware that for many newer readers of the magazine who had never known it as a print magazine [back issues of 'Inside Indonesia' was not a resource they had access to]. At the same time we knew that only a few complete collections exist, and that they are all in Australia. Increasingly the readership of 'Inside Indonesia' is in Indonesia and elsewhere overseas, and so we were looking for a way to enable them to access this important and rich back catalogue.” She says, “the feedback so far, especially from young Indonesians has been great, as they encounter for the first time some of the voices that were taboo during the New Order period.” Preservation has been an important driver says Purdey, “We began this process of making digital copies of the archive but soon realized that we did not have the capacity to create the archive in the way we wanted, nor to 'future proof' it and store it for the long term. When we approached the National Library with our project it was a perfect fit for our needs.”

User engagement with journals on Trove

Communities don’t just use—they improve. Although text correction is not yet available for journals on Trove, Trove does offer the capacity for users to curate lists and add tags and comments to journal material. A key challenge for digital libraries to remain successful is to engage the communities they serve (Calhoun 2014, p. 76). Some valuable pieces of research are being conducted in this way. For example, a user has created a list entitled 'Writers interviewed by Zora Cross (pseudonym Bernice May) in the Australian Woman's Mirror' (Cperkins, 2017). The description this user provides is:

“In the late 1920s and early 30s, Zora Cross interviewed her fellow women writers for the popular magazine the Australian Woman’s Mirror. Without these entertaining profiles, there would be little trace of many of these writers in literary history or biography.”

The list has collected together 26 interviews with female writers of the period, a new resource in its own right.

Pathways for data out of Trove

Although the digitised journal corpus lives within Trove, there are many possible places to discover and interact with this content. Organisations can get information out of Trove and expose it on their own websites. The simplest way is to embed links to Trove in a library catalogue, by including the journal URL in a bibliographic record. As part of the digitisation process, the National Library upgrades all relevant catalogue records with links to the online digitised version; these are shared and available for download via Libraries Australia.

For a richer experience, digitisation partners can create their own portals to expose material they have digitised, while still taking advantage of the National Library’s large-scale infrastructure, digitisation expertise and commitment to preservation and access. The NSW Rural Fire Service has created an online portal to the 'Bush Fire Bulletin' by embedding a Trove search widget in its website (Figure 4). When a user visits the Rural Fire Service web page, types in a search term and clicks the ‘Go’ button, they are then taken to a list of search results on the Trove website, in Trove’s journals zone (Figure 5). In the background, the widget has directed Trove to limit results specifically to issues of the 'Bush Fire Bulletin', by adding the search qualifiers AND ("nla.obj-284283986" OR "nla.obj-284265362") to the query. These are the specific identifiers for the two runs (1952-1975 and 1978-2017) of the 'Bush Fire Bulletin' and limit Trove’s search results to articles that contain those identifiers. Not only is the user saved from re-keying their search term, they are also given extra limiting functionality that is not currently present in the Trove interface. That the widget is facilitating access to this magazine is demonstrated in the relatively large number of page views of journal issues compared with home page views (see Tables 2 and 3).

This functionality has the potential to be replicated for other journals, or other types of content, on any website outside of Trove.

NSW Rural Fire Service
Figure 4: The NSW Rural Fire Service website has a portal where users enter search terms to search across only the digitised issues of the Bush Fire Bulletin in Trove

 

NSW Rural Fire Service portal on Trove
Figure 5: Search results on the Trove website, where a patron immediately sees results that match their search terms after coming from the NSW Rural Fire Service portal

This customisation can be taken a step further with discovery happening entirely within the digitisation partner’s own website, using Trove’s Application Programming Interface (API). The Trove API allows anyone to nominate search criteria such as a keyword term, subject, contributing partner or the specific identifier of a particular journal title. This can then be used to extract Trove records that meet those criteria, thus enabling access to curated data sets. DigitalNZ has used the Trove API to include the 'Pacific Islands Monthly' journal articles in its discovery service (Figure 6). DigitalNZ users discover articles, including title and thumbnail, on the DigitalNZ site then click through to Trove to view the digitised issue. Making data available where communities are searching, rather than expecting them to come to Trove, is a great way to maximise the usage of digitised content and, as has been noted, the opportunity to embed digital libraries much more effectively in the web’s discovery environments is a key success factor in ensuring their relevance and longevity (Calhoun 2014, p. 77).

Pacific Islands Monthly on DigitalNZ
Figure 6: 'Pacific Islands Monthly' is discoverable on DigitalNZ, beyond the Trove web site, thanks to the Trove API

Coordination of digitisation

Coordination of the digitisation of Australian journals to avoid duplication of effort, to inform Australian institutions’ project planning and bring coherence to the national effort to make Australian journal material accessible is clearly desirable and has been a strong theme in the National Library’s conversations with its partners, particularly universities. Significantly, the coordination of a national approach and the creation of a register of digitisation projects Australia-wide were key recommendations of Katherine Howard’s 2017 report for GLAM peak—'International Comparators: How does Australia compare internationally' (Howard, 2017), a finding that mirrors overseas experience (Rikowski, 2011). The National Library has extended its listing of the newspaper titles that will be digitised to include journal content also.

Conclusion

The National Library’s newspaper digitisation program has provided a useful foundation upon which to embark on journals digitisation and the enabling of a search and delivery functionality for journals on Trove. Implementing full-text discovery and new delivery features has contributed significantly to the popularity of digitised journals on Trove, increasing return on investment for partners. The National Library has used a variety of models for working with partners to increase both the volume and variety of digitised journals content on Trove. The impacts of having this digitised content available have included bringing communities and researchers together and enabling partner organisations to expose their collections to the world. Congruent with Van House’s conception of a digital library as “a locus of shared work” (Calhoun 2014, p. 150), Trove’s capacity for engaging its users in enriching the journal content on Trove and using it to create further content and knowledge as well as customising it will be crucial to its relevance. Trove endeavours to work with its partners to build and continually improve how Trove meets the needs and behaviours of its community.

References

Ayres, M 2012, 'Digging deep in Trove: Success, challenge and uncertainty', paper presented at the OCLC Asia Pacific Conference, 3-4 September 2012, Kuala Lumpur Malaysia, viewed 17 August 2017, http://www.nla.gov.au/our-publications/staff- papers/digging-deep-in-trove-success-challenge-and-uncertainty

Berthon, H 2016, 'Discover the Mt Isa Mining Magazine. Mount Isa Local Residents and MIMAG in Trove', Trove Blog, web log post, 23 November, viewed 17 August 2017 http://www.nla.gov.au/blogs/trove/2016/11/23/discover-the-mt-isa-mining-magazine

Calhoun, K 2014 ‘Exploring Digital Libraries. Foundations, Practice, Prospects’, Neal- Schuman, Chicago.

Cperkins 2017, ' Writers interviewed by Zora Cross (pseudonym Bernice May) in the Australian Woman's Mirror', Trove, viewed 17 August 2017, http://trove.nla.gov.au/list?id=104858

Hickie, J & Raadgever, M, 2015, 'Easy as ABC A Triumph of Re-usable Metadata', paper presented at ALIA Online, 2-5 February 2015, Sydney Australia, viewed 17 August 2017 http://www.nla.gov.au/our-publications/staff-papers/easy-as-abc

Howard, K 2017 ‘International Comparators: How does Australia Compare internationally?’, GLAM Peak Report, viewed 17 August 2017, http://www.digitalcollections.org.au/framework/international-comparators-research-report

Lankes, R, Silverstei, J & Nicholson, S 2007, 'Participatory Networks: ‘The Library As Conversation’', Information Technology and Libraries, vol. 26, no. 4

Lynch, G 2016, 'Searching for The Sentinel', mmnews.today, 29 January, viewed 17 August 2017 https://mnnews.today/aurora-magazine/february-2016/7641-searching-for- the-sentinel/

Rikowski, R 2011, ‘Digitisation Perspectives’, Sense Publishers, Rotterdam.

Qidenus Technologies 2016, ‘Book Scan 4.0, Robotic Series’ viewed 14 November 2017, http://qidenus.com/technologies/products/robotic-book-scanner/

 

End Notes

 

[i] These xml files hold details of the physical and logical structure of the content including issue dates and page numbers, the re-keyed and automatically OCRed text, the location of each content type on the page and links between content.

[ii] The National Library’s digitisation team comprises two full-time permanent staff members, two to four casual staff and a proportion of the management time of three further permanent staff.