In September last year, at the ALIA 2014 conference, I spoke about Trove at 5, and asked ‘are we there yet’, focusing on Trove’s audiences and the work still ahead to extend its reach to a full cross-section of the Australian community. Today I would like to focus on where ‘there’ is, by comparing Trove
to three other cultural aggregators, Europeana
And Digital Public Library of America
These are the services with which Trove is sometimes compared and with which we have fairly close contact. I think of them as the aggregator siblings. All four aggregate metadata describing content from multiple contributors so that users can discover resources through a single portal, and developers can access resources through a single API platform. All are based on a common value of making it easier for the public to access, enjoy and use the collections of cultural institutions.
My full paper goes into some detail about the similarities and differences between these four services, but today I will focus on a broader view of the service cultures and the ways in which they might be said to reflect their local environments. As Ashley Sanders, an American historian tweeted on January 6,
‘Everything is constructed, including library and archival catalogs. Understand that construction’, and as Siva said on Tuesday ‘To organise is not a neutral term’. That is as true for cultural aggregators as it is for Google.
I will start by taking you back to the home page of Europeana as it appeared after the seventh of January 2015, and still appears today.
Perhaps because I have spent a long time thinking about Trove and its siblings, this blog post immediately signalled much more than just its content, and highlighted a key point of difference, at least between Trove and Europeana. I cannot imagine a scenario in which we the Trove team at the National Library would have written a post like this:
Or placed it so prominently on the website. This is absolutely not to say that we – at the Library, in this room, in the profession, in the wider cultural sector – do not share in that sense of shock and outrage, or that we do not – at some deep level – feel that at least part of our jobs is to strengthen the Australian citizenry and its democracy by making information freely available. Nor is it to say that we would not want to express our sense of solidarity
with the communities we serve, to feel ourselves shoulder to shoulder in such a time. But we just would not speak so directly about Trove as a political mission, and a statement like this would really jar with the culture Trove has developed over the last five years.
While I was cycling home one day, musing on this issue of how to represent the different cultural identities of these aggregator services, I started to think about the concept of the vernacular and – in particular – the way we apply that concept to architecture. Which led me to asking myself, if Trove was a house or a building, what kind of house would it be? What would Europeana, Digital NZ and DPLA be?
For me, Trove as house or building would be a country or outback homestead:
Of course homesteads come in many varieties, and I commend Trove to you as a great place to look closely at this vernacular. Bandon Grove in Queensland is clearly capacious, has a bit of a middle-aged spread about it, it’s plain, not gussied up, and it pretty clearly tells pilots exactly what it is.
Closer to my home in Canberra is the much grander Lanyon homestead, on the banks of the Murrumbidgee. Again, it is expansive, but this one is built mostly of stone and very much made to last, which is definitely how we think of Trove. Trove’s home within the National Library of Australia gives it a very solid institutional base, something it has in common with Digital NZ but not with Europeana and DPLA which both exist as distinct legal entities, funded by multiple partners but not strongly tied to any single national institution.
There are pluses and minuses to both modes of establishment, and I think it would be true to say that while the big plus of a single National Library home is the long-term commitment that national libraries have, one of the minuses is funding.
We at the Trove team do sometimes wring our hands at our modest resource base, feeling a bit like the poor cousins in the world of aggregators. Many of you know that Trove’s digitised newspaper content has been jointly funded by the National Library and other contributors, on around a 50/50 split, with the State Libraries of NSW and Victoria the largest contributors, assisted by significant state government funding for things digital. But the central service, and the National Library’s own contribution to newspaper digitisation has been funded without additional Commonwealth government appropriation. We have had to close other services, reduce our overseas collecting, and forgo many possibilities and opportunities in the digital and physical worlds in order to invest in Trove. Our friends at Digital NZ also operate on a modest internal only resource base.
By contrast, Europeana has been very generously funded by the European Union, and supported by high levels of member state investment in national digitisation programs, something we have not seen at the national level in Australia. DPLA is funded by a combination of major investments from the kinds of large Foundations that operate in the US but not in Australia or New Zealand (i.e. Mellon and Gates Foundations) AND by federal funding bodies, the National Endowment for the Humanities and the Institute for Museum and Library Sciences for which there are no Australian equivalents. DPLA has to compete for these funds, so does not have an ongoing and secure resource base. But – like Europeana – it at least has access to competitive funding programs, which allows it to source significant project funding to work on what I think of as the Big Issues. Recent examples include a $600,000 Mellon grant to research and implement a sustainable business model, and a $300,000 Knight News Foundation grant to research and simplify access and use statements across DPLA content. The National Library cannot access any competitive grant funds, meaning that these kind of Big Issues need to be worked on within existing modest resources, and at a snail’s pace.
We are really very conscious that Trove – like many an Australian farmhouse – has a solid core, but also has what we might think of as a set of lean-to’s, additions that we’ve squeezed in, or tacked on, and that don’t always fit neatly into a simple narrative about what Trove is. All four services provide additional pathways to freely available digital content delivered through contributors’ own websites. But only Trove exposes the entire National Bibliographic Database through the service, only Trove includes a set of licensed eResources, only Trove includes a very large body of records describing people and organisations, and of course only Trove – at this stage – provides direct access to 15 million newspaper pages. For those of you who saw Tim Sherratt’s talk on Tuesday, there are plenty of seams and edges in Trove!
Of all the many images I explored when I started pursuing this idea, perhaps this faded beauty from Museum Victoria speaks most directly to how I think of Trove as a house. No-nonsense, in a bush setting, more than one lean-to, and instantly recognisable as ‘Australian’. And – most importantly – a great meeting place. A place where people can come from far and wide to discover their stories, swap a yarn, reconnect with identities, broach the kinds of distances that are a fact of life in our country.
When I thought of Digital NZ as a house, I thought immediately of the New Zealand bach.
Modest in size, well-maintained, absolutely nothing extraneous, four-square,
Colourful, economical, models of recycling and re-use, elegant in their simplicity.
Our friends in New Zealand do such great work on a very modest resource base, and are always innovating and sharing. Some of you yesterday will have seen Julia’s and Mark’s presentation on how they harvested ABC content into Trove, using the NLA Harvester. Digital NZ’s Harvester has a beautiful name – Supplejack, a New Zealand plant name – and our colleagues are working to make it available as an open source tool, looking beyond their own immediate needs and offering their expertise to the world. Having known and loved Digital NZ for a long time, I can also say that it is
wonderfully quirky, exemplifying a very specific sense of New Zealand humour, which those of us who follow DigitalNZ’s tweets or view their curated sets of content will recognise.
What came first to mind when I thought about DPLA? Well, it was all to do with grandness of vision, alignment with great democratic aims, lofty ideals about the role of cultural collections in national life and that great American strength, entrepeneurship. Which led me to these beautiful images – all sourced from DPLA – of National Mall in Washington and its surrounds.
For those who haven’t wandered the Mall and its institutions, it is a grand boulevard lined with some of the world’s greatest museums,
Planned, geometric, a powerhouse of culture. I do not think it’s too big a stretch to say that our DPLA colleagues have a grand ambition, and that is to be a National Digital Mall – the monumental core kind, not the retail kind.
One of the key differences between Europeana and DPLA on the one hand, and Trove and Digital NZ on the other is that Europeana and DPLA work primarily by aggregating metadata from other aggregators.
DPLA harvests metadata from large single-institution ‘content hubs’ and from eleven state or regional service hubs, covering fifteen states. DPLA focuses its efforts on encouraging hub development, rather than harvesting from individual organisations. Its newly released Strategic Plan sets out its aim to ‘complete the service hub network’, and to ensure that every United States library, archive or museum with a digital collection has an intermediate hub available to join. DPLA states explicitly that institutions with fewer than 200,000 digitised items operating in a state without a service hub cannot be ‘onboarded’, as the peer-to-peer relationships involved are beyond DPLA’s capacity to manage.
This is in sharp contrast with Trove which, for the most part, must work with individual contributors. Trove’s smallest individual collections number in their tens, and many contributions number between a few hundred and a few thousand items. There are no comprehensive state-based hubs in Australia, aggregating cross-format digital content at scale from multiple sectors within the state. There are no single sector based hubs, aggregating digital content from a single sector such as Europe’s Archives Portal Europe. Collections Australia Network, which focused on museum collections, proved unsustainable and was closed in 2014, after the Cultural Ministers Council withdrew funding in 2011. This essentially means most of the burden of aggregation, working one contributor at a time, falls to the National Library. And with a limited resource base, that puts significant brakes on the speed with which we can work with contributors. If, as Siva Vaidhyanathan noted in his keynote presentation on Tuesday morning, DPLA operates on a ponds - lakes – ocean framework, then Trove is working with many, many ponds, of vastly different sizes and technical ability.
If, DPLA is more geometric in terms of its aggregation shape, a big part of their ‘monumental’ shape is directly oppositional:
This image was published with Andy Baio’s Medium blog post of just last week https://medium.com/message/never-trust-a-corporation-to-do-a-librarys-job-f58db4673351. As Siva mentioned on Tuesday, DPLA nails its colours firmly to the ‘library’s job’, with one of its three main elements stated as being ‘an advocate for a strong public option in the 21st century’.
Europeana echoes this opposition to the commercial, but has a much stronger sense of mission, or, as they state themselves, Europeana is a ‘big political idea to unite Europe through culture’ … and that culture is ‘too important to leave to market forces alone to digitise and make available’.
So, when I thought about Europeana the vernacular that came to mind was the European Parliament. Europeana has an overt mandate from the European Union and its member states and there is a very real sense in which Europeana is part of the same project as the EU itself. It has a pan-European vision. And it is this vision that means that Europeana’s blog post makes perfect sense when it would be a bad fit for Trove, which simply has no government mandate of this kind.
Now, the problem with thinking about Europeana as the European Parliament building is that there is more than one of those. Unable to decide where the Parliament should reside, the entire Parliament moves between Strasbourg and Brussels on a regular basis, and there are therefore two European Parliament Houses.
Here is the Parliament in Strasbourg
and the debating chamber in Brussels. If you look at these two images, I think you will be struck by the elegance of these chambers, the circular shapes, ample seating for members of the public, also by the sheer number of seats in the Parliament. Inclusive – yes, undoubtedly. Effective? I think there’d be a few different answers to that! A reasonable portion of one of these chambers could be filled by the many people across Europe participating in Europeana’s complex multi-level governance model. It is in the area of governance that we see the most striking differences between these four services, all with pretty much the same mission.
Europeana’s governance is firmly representational, reflecting the significant investment by the EU, member states, individual cultural organisations and professional associations. While there is a very high emphasis on bringing together a community of expertise, there is also a sense in which that old term ‘follow the money’ applies. Europeana’s latest strategic plan states that the organisation wants to achieve a ‘more transparently democratic organisation’ and will likely require increased direct investment by contributors and member states to deal with significantly reduced EU funding over the next few years. In other words, Europeana operates on the ‘no taxation without representation’ and I suspect ‘no representation without taxation’ model.
DPLA, meanwhile, is a tax exempt non-profit entity funded primarily by philanthropists. It is governed by a Board which includes public and research librarians, technologists, intellectual property scholars and business experts from around the country and its Board is chaired by the CEO of Creative Commons. It thrives on the US culture of providing private time and money to support a public good. DPLA had 12 staff at the end of 2014, just a little under Trove’s 14.5 full time equivalents at the same time. But its very open governance models and ability to attract people to a great public cause, means that it has at least 50 individuals involved – in some way – in DPLA’s direction setting.
Trove’s central infrastructure is entirely funded by the National Library and its governance is – for the present – entirely internal to the Library, with updates provided to a small number of key stakeholder groups. We do think this will need to change, and have experience with many different models, but do not plan to change Trove’s governance for the next couple of years at least, as we are very focused on supporting aspects of the Library’s own Digital Library Infrastructure Replacement Program and, frankly, we simply do not have the resource to divert to a broader representational or expertise model at this stage. DigitalNZ is also funded by its National Library, and although it has a small external Board, this is not representational, but aims to bring specific sets of expertise to the table.
Having delved a little into detail, here is my final image – perhaps exemplifying all four services. In many ways, our services are all Crystal Palaces, triumphs of engineering designed to allow the general public to see the marvels of the world. The ways we engage with our publics share some characteristics – all four services are heavy users of social media – but are also very divergent.
Where Europeana could invest in sending teams to physically visit towns and villages across Europe so that members of the public could digitise and contribute their own personal heritage to the broader service, DPLA’s governance model includes regular ‘community’ Board meetings, where any member of the public can dial in on an 800 number to participate, DigitalNZ has strong links to New Zealand’s education community and Trove – well, Trove has its suite of user engagement services. And our public have – to date – given us the equivalent of 382 work years or $25million worth of their time to correct OCR’d text, and use of all Trove’s user engagement features grows by 10-15% every year. And while I do not know how many Europeans visited Europeana last year, how many Americans visited DPLA, how many New Zealanders visited Digital NZ – because those usage figures are not readily available, I do know that Trove received more than 24 million visits last year, that we regularly have more than 50,000 unique daily visitors, and that every time I walk past the Trove Raspberry Pi, somewhere around 700-800 people are engaging with Trove.
In my heart, I know that even with a very modest resource base, without the beautiful curated exhibitions our friends at Europeana and DPLA can create, without the terrific alternative ways of exploring content like DPLA’s timeline and maps based interface, and without the access to project funding that would allow us to really push ahead with some big innovations, … with all its seams and edges and lumps and bumps, well, I know that Trove has been taken to Australians’ hearts, that it is making a real difference to our national community, and that knowledge alone is enough to drive the Library to keep Trove strong, to restump the farmhouse every so often when it needs doing, and to work a bit on those lean-tos, so that we can all share its wide verandahs for a long time to come.