Dr Marie-Louise Ayres
Publication date: 
Tuesday, 15 April, 2014

Just last week I saw a quote on our National Library Yammer[1] site: ‘old librarians love books, good librarians love data, great librarians love people’. The first phrase is ageist and derogatory, but the second two phrases speak to me. I would like to combine them by saying that great librarians love the way that data – and content – can serve people, can have impacts beyond our imagining and indeed can change lives. We have been thinking a lot about impact recently at the National Library, especially in relation to Trove, but the concepts are just as relevant to national bibliographic services such as Libraries Australia and Te Puna.

When Te Puna’s Jenny McDonald asked me to speak, mentioning that the other keynote would be from Pat Losinski and that he would be focusing on the ‘collective impact’ of programs of work, this idea resonated immediately with me. Just last week I saw a quote on our National Library Yammer[1] site: ‘old librarians love books, good librarians love data, great librarians love people’. The first phrase is ageist and derogatory, but the second two phrases speak to me. I would like to combine them by saying that great librarians love the way that data – and content – can serve people, can have impacts beyond our imagining and indeed can change lives. We have been thinking a lot about impact recently at the National Library, especially in relation to Trove, but the concepts are just as relevant to national bibliographic services such as Libraries Australia and Te Puna.

I do not have any doubts about the collective impact of our decades worth of national collaborations, but I do think we are now in a social, policy, technical, and fiscal environment in which we need to ask ourselves how we can maximise collective impact and change the most lives. Our predecessors in our profession took giant steps decades ago by instituting efficient and effective national services. One of the questions in my mind is – how can we know that we are still delivering those services most effectively and efficiently at the national level? And when – if ever – will come the day in which working at larger – or indeed smaller – scales will maximise the impact of our work? As a member of the National Library’s senior management group, I ponder these questions about a range of our programs, ranging from bibliographic through to digital preservation services. Today, I will focus on Libraries Australia – one of our services for libraries - before turning my attention to Trove - our service for people.

I know that some of you will be familiar with Libraries Australia and our other national services - but others will not be. I want to start where we always need to start – with the Commonwealth of Australia legislation from which we derive our authority.

[1] National Library of Australia staff share news, ideas and interesting reading material via a dedicated Yammer site.

The National Library Act (1960) includes specific provisions concerning the Library’s role in providing national services to assist Australian libraries and authorities.

Our Strategic Directions 2012-2014 includes:

  • a Direction - Deliver National Leadership
  • an aspiration - Provide national infrastructure to underpin efficient and effective library services across the country, and to support Australian libraries in the twenty-first-century digital world
  • and a specific action - Renew the Libraries Australia Search service to meet the collection management needs of over 1,000 Australian libraries.

This speaks directly to Libraries Australia’s mission, as do two other National Leadership aspirations – to engage with libraries in the Asia and Pacific regions – and to participate in international collaborations, with a focus on standards and practices for the creation and exchange of data describing collections.

But beyond this focus on data and libraries, our Strategic Directions also sets out our intention to:

  • Make our collections and services accessible to all Australians
  • Engage with diverse Australian audiences, including those in regional and remote communities, and
  • Support research and lifelong learning for all Australians by maximising online access to our collections and services.

Actions to support this aspiration in the current triennium include: developing a robust and sustainable business model for Trove, expanding digitisation of regional and community newspapers in partnership with communities and benefactors, and finding faster and cheaper ways of making our collections discoverable in the online environment. The entire focus of this work is on maximising our collective impact.

I do not have to tell this audience that maximising impact in relation to our collections depends on good data and on people who are passionate about that data. It starts with librarians working at individual libraries and expands its power through shared bibliographic services.

Libraries Australia and its predecessors Kinetica and ABN have been running successfully for more than 30 years. The services we have today are thanks to the vision of library leaders of the past, and on days when I am trying to figure out our next move, I look for inspiration at this little office 'installation

of the name badges of some of my predecessors in this role: Debbie Campbell; Warwick Cathro; Tony Boston; Roxanne Missingham; Andrew Wells... and of course those who preceded them.

 18,333,091 RIS, Cataloguing, Web input transactions in 2012-13

Over the last several decades, these leaders, other National Library staff, and the staff of libraries across Australia have built a formidable piece of national infrastructure. Libraries Australia includes 25 million bibliographic records, 50 million holdings, has more than 1200 members, and in the 2012-2013 financial year supported nearly 18 million searches and more than 18 million transactions – adds, edits and deletes of records and holdings.


Support for the Libraries Australia service is provided by two Divisions of the Library - my own Division of Resource Sharing which is responsible for strategic directions, the business, data management, and member support; and the Information Technology Division which is responsible for infrastructure and the many pieces of internally developed software which support the service. We have a relatively large Information Technology Division, comprising more than 40 staff or 10% of the total Library staff profile.

The National Library creates the majority of original records added to Libraries Australia; these records are created by our colleagues in the Collection Management Division. The Library is by far the biggest lender in the Australian system, with ILL and document delivery managed by our colleagues in the Australian Collections and Reader Services Division. While we manage our own billing in the Resource Sharing Division, our colleagues in the Corporate Services Division provide our financial services and assist us with cost modelling etc. This is to say that Libraries Australia, its content and services, is a major area of business for most parts of the Library, and is seen as an essential part of our national leadership role.

We offer a range of services, and perhaps the best way to talk about those is to show you this fairly scary looking diagram of the ways in which the various 'bits' of Libraries Australia work. Like most National Library of Australia services, the LA 'stack' includes some pieces of licensed software (they’re the ones outlined), and a lot of internally developed software, customised over the years.

To focus first on bibliographic services, our members can add records using our Record Import Service (RIS), the Cataloguing Client, and the Web Cataloguing form. We export records via our Record Export Service, and our Products service. The latter includes a number of free self-help data products, and a boutique service for particular customer needs. We support standard protocols such as Z39.50 to support workflows.

Addition, edits and deletes of records and holdings are all managed in OCLC's CBS software, supported primarily by OCLC in Leiden. We apply match-merge processes, and are always refining our scripting to minimise the quantity of manual review we do.

Of course we also offer our Search service, currently based on an aging TeraText database. We are redeveloping Search and I will return to this issue a little later. We offer our members a range of other targets - including Te Puna via the Search service.

And then there is Libraries Australia Document Delivery. Unlike Te Puna, we still have a local version of VDX, and indeed we upgraded to VDX 6 over the weekend of 14-15 March 2014. We are considering future options in this area.

I would first like to share some long-term trends for use of Libraries Australia services, as it is crucial to consider these as part of our service context.

As you can see, use of Search has climbed steadily over more than 10 years, and we are now supporting twice as many searches as we were when we developed Libraries Australia from its predecessor, Kinetica. Behind this search activity, we also know that more bibliographic records and holdings are being added to Libraries Australia than ever before. Under the hood, we know that use of all methods of updating the Australian National Bibliographic Database are holding their use patterns, so we know that all methods are still useful to our users.



However, on the Document Delivery side, we seem to have already passed the peak of activity peak, and we know that for a small number of Australia’s biggest libraries, Libraries Australia Document Delivery or LADD now supports just less than half of all the document delivery they undertake on behalf of their users. There are quite a few competitors out there, especially for article level service.

The peak occurred between 2008 and 2010, but as you can see, we are still running at levels above where we were 10 years ago. If I were a betting woman I would suggest that we will probably see a gradual tailing off of activity over the next several years; I would be astonished to see another peak. As you can see, LADD is still supporting at least a quarter of a million transactions every single year. That is a quarter of a million resources in the hands of people who need them but not otherwise be able to get hold of them. Despite dire predictions over several years, ILL is definitely not dead.

This is a report we have not previously run, showing Australian lending to New Zealand and New Zealand lending to Australia since Trans Tasman lending commenced in 2006. Requests to New Zealand form a relatively modest 1.1% of all LADD requests and loans to New Zealand comprise 2.3% of all transactions. Still, between them, that means we managed to get nearly 10,000 resources into the hands of Australians and New Zealanders who would otherwise have had to fly across the ditch to get access.

We have also seen a peak in Trans Tasman lending and are now in gradual decline. I am a bit of a 'why' person and in this case I have to say that I have absolutely no idea why the fall in Trans Tasman lending is running a year or two behind the bigger LADD picture!

Over the last nine months, we have been looking at our products service which supplies specialist datasets, either on a self-help or charged basis. Last July we increased our product charges for the first time in many years, and a few months later we produced a new booklet - online and hard copy - on what products could be used for. Our products manager, Mel Satrapa also spoke at last October's Libraries Australia Forum to let our members know how products could be useful to them.

As you can see from this graph, there is no real sign that our increased charges, booklet or promotion of the service has made an appreciable difference – either up or down – to use of the service. We have seen a gradual decline in requests for products over recent years but feel it is definitely still worth supporting - and that we may yet find new uses for the service. A colleague at the State Library of Queensland told me just last week that their monthly product of new items published in Queensland significantly increases their legal deposit claiming success.

Before we move on, I would like to note that the graphs you have just seen are produced by our business manager, from an extensive storehouse of business metrics. We maintain a close eye on these metrics, and make them freely available to our members. For example, we recently provided a set of detailed datasets and graphs relating to university sector engagement with Libraries Australia services over a 10 year time period. We aim to be as transparent as possible with our members, even when we know they are asking themselves hard questions about whether they are still getting value for money from their Libraries Australia subscription. And believe me – our members do ask that, especially the university sector which accounts for half of all Libraries Australia activity, and half of all revenue.

I completely understand this question being asked, and often see a disconnect between the way library CEOs think about the Libraries Australia value proposition (that is, asking a lot of questions about value), and the way their staff using our services think about it (that is, absolutely sure that they need us!). Much is changing in the library world, and that means that every CEO worth his or her salt will indeed be thinking about where they want to invest.

The two biggest changes will be obvious to you: a major shift to acquiring or leasing ‘electronic’ rather than acquiring ‘print’; and increasing outsourcing of cataloguing especially in the university and public library sectors. These changes are happening across most sectors, with the exception of special libraries where their size and disparate nature makes it difficult to seize new opportunities or new ways of doing things.

In some Australian universities, 90% of acquisitions budgets are now being spent on electronic resources and public libraries are moving fast into the eBook world. One of the logical consequences is that libraries will need to do less copy cataloguing and therefore that one of the major efficiencies Libraries Australia offers will reduce in value over the years. On the other hand, we know that many libraries – especially universities – will be managing down their print collections over the next few years, and that Libraries Australia holdings data helps individual libraries and groups of libraries to make those decisions in a sectoral or national context.

Another consequence of the move to electronic is that, with more restrictive publisher licensing, the capacity to deliver documents may reduce at the same time that electronic access would theoretically make it easier to supply to other libraries. We are aware of the potential of these big changes to undermine Libraries Australia’s underlying value proposition – that our cataloguing and document delivery services offer Australian libraries major efficiencies in their every day business.

We know from our surveys and interviews in 2011-2012 that libraries are being asked to justify individual aspects of their expenditure to their parent bodies, rather than managing a program budget. Government libraries are feeling the pressure in Australia, with declining budgets, actual or threatened closures, and an increasingly tough environment in which to make a case for relevance. So we absolutely expect that all parts of our membership will be looking at the bottom line and asking whether they still need our services, or whether they can secure similar services for lower cost.

Another major disruption or change to the ways libraries do their business is the advent of cloud-based services. Clearly these have the capacity to transform data flows. We have been living in a world in which ILMS vendors and metadata aggregators occupied fairly separate parts of the universe. But we are now in transition to a world in which vendors of cloud-based ILMS systems are also potentially or actually metadata aggregators (we can think of Ex Libris’ Alma ‘community’ catalogue in this regard), and metadata aggregators are developing cloud-based ILMS systems (and here we can think of OCLC’s WorldShare Management Services). I will talk a little more about our responses to these changes a bit later.

We delved deeply into these and other library mission and priority changes a couple of years ago, when we decided whether to proceed with a significant investment to redevelop the Libraries Australia Search Service. As a senior manager, I needed to know that the new software asset we would create through this project would earn its keep by being useful for at least five years. We undertook a large survey of our users in late 2011 and early 2012, combined with intensive interviews with library leaders from every sector before confirming our confidence that our services are still in demand, still useful, and that providing the services ourselves still seems to offer the most effective and efficient operation for our members.

Now it is all very well to say that we did a lot of thinking before deciding to invest in redeveloping Search– because we did. But in many ways, we were being pushed. When we embarked on the project, we were supporting twice as much Search activity as was the case when Libraries Australia Search was first developed. And Search was under strain. We had a horror run of outages, were having to spoon-feed in batches of records to avoid overwhelming the system, and were in danger of not being able to provide the reliable services our members need. Our big problem was that the Teratext platform was no longer able to cope with the volume of activity we needed to support. We knew we had to replace it with a more scalable platform, and our experience using Solr Lucene to support Trove (on which, more later) meant that we were confident it would scale up if required. So our aim was to get to an infrastructure that looked like this:

I think you’ll understand from the number of connections shown here that we also needed to upgrade our hardware and many pieces of software in order to achieve this. This included replacing Oracle with MySQL which in itself was a major undertaking. Our project is called Libraries Australia Search Replacement – or LASR for short – but it has involved much more than just the search database and interface.  

In addition to beginning our replacement project, we did some interim risk mitigation.

We created a second instance of the entire Teratext database for internal use. This instance is used by National Library cataloguers and for our Records Export Service, while the external instance is used by our other members and for our Records Import Service.

While this took a bit of work - and there were associated tweaks around the edges - it has paid off. When we started what we thought would be a 12 month redevelopment project we were seriously worried about the risk that activity levels would continue to grow and that Teratext would die before we'd developed the new Search. In fact, we were able to address those risks, and have breathed easily for at least the last year, as it became apparent that splitting the databases had given us some much needed head room. I am especially glad of this because although we thought replacing Search would take us 12 months, by the time we finish in June, it will actually have taken two full financial years. With the additional costs this has brought, I am even gladder that we were sure the project was worth proceeding with. The big lesson for us here is that even where requirements are well known and we are replacing a mature system, there can be many unexpected delays and under-estimates along the way. A lesson for all libraries I think!

We wanted to improve our Search service to our members, but we were not necessarily looking to reinvent the wheel. In deciding how to proceed, we had the benefit of huge buy-in from our members. More than 500 individuals from more than 360 of our member libraries gave freely of their time, thoughts and expertise completing a detailed survey about their interactions with Libraries Australia. They included more than 3000 incredibly thoughtful and useful textual comments in that survey. We were also grateful that so many senior library leaders were willing to give up their valuable time to talk about strategic directions.

We gained a lot of information during that review period that directly fed into our decisions around the new Libraries Australia Search. We got many suggestions about what our members liked about the service, what they did not like and what they wanted to see changed. We found that members were actually happy with the existing Search (apart from the performance issues!), did not want things to change too much in the interface, put a high premium on clear and uncluttered screens, and had developed ingenious workarounds for things we are able to fix in the new Search. We heard loud and clear that our members did not want a Trove-like interface. And interestingly, we had strong messages that they did not want a FRBRised environment, but wanted to see individual bibliographic records for manifestations.

This was a major driver for deciding not to try to run Libraries Australia and Trove from a single database. Although there were theoretically significant savings to be made from having a single database and running two separate services from it - that is, Libraries Australia and Trove - in practice we decided this was too risky, would have required complex rules, and would compromise performance. The Trove database is FRBRised and trying to use it for Libraries Australia would have required disaggregation of record clusters back into individual records, and the addition of a large number of indexes which librarians need to do their work, but which Trove users do not. Every additional index increases strain on performance and we just did not think we could go there.

Nevertheless, we will gain significant efficiencies by using the basic toolset - Solr Lucene - re-using many indexes and many pieces of code, and these efficiencies will stand us in good stead in out years.

So where are we now, and what will we be offering our members when we release the new Libraries Australia Search into production? I am happy to say that although the Libraries Australia Search Redevelopment project – LASR for short – has taken twice as long as anticipated, we are nearly there. We anticipate commencing user acceptance testing in the second half of May, and putting the new service into production before the end of June.

I mentioned earlier that we did an extensive survey and review of Libraries Australia before commencing LASR. But our communication with our members did not stop there. We sent a business analyst and designer to observe our members using the service – and learned a lot from doing so. We maintained a project blog.

We developed two Alpha versions, let our members know about what functionality was available and what was not, and encouraged them to test out the Alphas and provide us with feedback – which they did.

I am going to move quite fast though a few screens from Alpha version 2. II should just note that the colour scheme of the Alphas was deliberately muted so testers could focus on functionality. We have actually opted to stick with our existing Libraries Australia livery, so the production service will be in the familiar shades of blue and green.

We encouraged members to test Simple and Advanced search, and Browse.

Our simple search screen now shows highly used databases and offers the ability to add and remove databases on the same page – dramatically reducing the number of clicks our members have to use. Authority databases can now also be searched at the same time as bibliographic databases

We have also made significant changes to the Advanced Search Screen and indexes.

This includes adding ‘author’ to Advanced Search defaults, splitting number indexes into two – one for control numbers and one for standard numbers – and giving members a new option to search for the date a record was created, not the date it was added to Libraries Australia.

Search limits are displayed at the top of the results screen, but we have made many other changes to results screens.

This includes previewing a record by hovering

and previewing holdings by hovering.

In both cases, the behaviour can be switched off by clicking the ‘Hide preview’ button.

These previews were added to assist librarians to "find the right record" with a minimum of clicks. Additionally, some inter-library loan officers asked if it could be easier and faster to see holdings statements.

We have also changed screen layout and sorting, with ‘latest first’ as the default given that this was the most requested default sort.

Our members will have new sort options, including relevance ranking (which we had to switch off in the existing service several years ago because of its impact on performance), and the number of holdings. We now present information in tabular formats as we found that our members spend a lot of time scanning information and we wanted to make it as easy as possible.

I want to highlight two specific interface improvements, the first of which is favourite libraries.

This improvement was a direct result of us sending our business analyst and designer to observe real people undertaking every day workflows. They found one lady who had memorised a list of 40 libraries which she knew offered a fast and efficient ILL service. We thought we should be able to do better than that! So the new Favourite Libraries enables users to select any number of favourite libraries which might be reciprocal, preferred or affiliated libraries.

A favourite library is represented by a heart icon. Records will display a heart and a number which indicates the number of favourite libraries that the item is held in. It is very easy to edit favourites from multiple screens.

The other important interface change – one that we think will make a big difference to our members – is to support comparing records and reporting duplicates.

Our members will now be able to view between two and four records in one screen. Records are selected from the search results or from saved searches.

Users can view records in full and MARC views. The records are presented in a table, and differences between values for a particular field or tag are highlighted to enable easier scanning of information.

Users can also expand holdings to view the libraries that hold the record, with libraries displayed by State for easy scanning.

Once again, the records are presented in a table and differences between values for a particular field or tag are highlighted to enable easier scanning of information.

In this view the 850 holdings are lined up and displayed by National Union Catalogue number. These comparison screens will allow users to select the best record

but also easily report a duplicate record instead of our present rather clunky method.

Of course, Libraries Australia Search is just one part of the Libraries Australia suite, and we continue to implement new services behind the scenes – many in conjunction with our partner and supplier OCLC. We upgrade our CBS versions and configurations on a regular basis.

In 2011-2012, we implemented Duplicate Detection Software. In the course of a year, almost half a million duplicate records were removed from the database, something we could not have achieved using human review.

Last financial year we implemented the Relate software. The software matches headings in bibliographic records with the corresponding authority record, improving the quality of bibliographic records by replacing non-preferred headings with the preferred form. Prior to this implementation, the only links between bibliographic and authority records were those added manually by cataloguers using the Cataloguing Client. In its first 15 months of operation, Relate matched nearly 7.5 million authorities and records, something that could clearly never have been achieved at human scale. Both pieces of software now run in the background, continuously improving database quality.

We implemented version 6 of the VDX software for our LADD customers over the weekend of 15-16 March 2014. This is still a version behind some of our members and indeed the Te Puna hosted service. Nevertheless, upgrading from version 4.3 to version 6 offers substantial benefits to our users. The following weekend – 22-23 March 2014, we implemented the CBS Job Manager – known as CJM. This piece of OCLC software has been a long time in the coming, but now offers us new opportunities to undertake large scale database maintenance changes that previously required assistance from programmers. The jobs can be automated, and to give you an idea about what we will now be able to do, it will allow changes to bibliographic data such as:

  • conditionally modifying or deleting specified parts of records, e.g. for fixing incorrect codes, or changing obsolete subject headings to new ones
  • conditionally transferring or copying the contents of a subfield to elsewhere in the field or to another field.
  • adding new fields to specified records

We can also combine these changes, allowing us to do more sophisticated changes such as for the conversion of AACR2 records to RDA.

CJM will also allow us to make changes to member holdings, such as

  • deletions
  • transfer to another customer – important in these days of merging libraries
  • deletion of local system numbers
  • changes to holdings text

These changes can be applied to all of the member’s holdings, or to ones that meet a condition such as a certain phrase being in the holding statement, or for a list of record numbers. We think that the speed of our new Libraries Australia search database and the new CJM tools may assist us to speed up the process of full holdings refreshes.

We are also working in another new area with our OCLC colleagues in Melbourne. We have had a very effective working relationship over the last couple of years as Australian libraries have started implementing the cloud-based WorldShare Management Services (WMS) ILMS. We have added the holdings of 14 WMS holdings to Libraries Australia, and are nearly finished work on another 3 more complex WMS library implementations, where the implementing libraries already have holdings on Libraries Australia. This is a new world for us. Hitherto, our data flows have primarily been from an individual library to Libraries Australia – using our various add, edit and delete services. We then push the records and holdings to OCLC’s WorldCat. But with WMS, library records and holdings are being added to WorldCat and then pushed back to Libraries Australia. Obviously we have had to do a bit of work to ensure that records do not just loop endlessly between the two aggregators.

WMS is used by small to medium sized libraries. But we also know that at least 17 large member libraries have or plan to implement Ex Libris’ cloud-based Alma system. At this stage, we expect data flows to remain as they have been, that is, that individual Alma implementers will send records and holdings to Libraries Australia, using what we hope will be a standardised profile. However, it is possible that may change in the future, and the work we have done with WMS will stand us in good stead if large scale data flow changes occur in a world that will see more and more libraries opt for cloud-based systems.

I have just been talking about data flows, and our Libraries Australia work is all about making data creation and flows as effortless as possible for our member libraries. But while that is important, we also put a lot of effort into maximising the value of the data our member libraries create.

We know that others are investing and experimenting with linked open data – and we find this fascinating. At this stage, we are really happy to watch the OCLC research and development team research and test theories in this area, and see no need to replicate their work. With all Libraries Australia records and holdings being pushed through to WorldCat, we watch with interest as records are turned into linked open data and then made useful in Wikipedia, Google and other really big information aggregators.

Our work to increase the value of bibliographic data and other content is principally focused on Trove, our service for the Australian people, and anybody else in the world who finds the service useful.

So I cannot leave you without talking about the relationship between Libraries Australia and Trove.

Almost every bibliographic record - with the exception of purchased Serials Solutions records - and every authority record in Libraries Australia is exposed through Trove. Libraries Australia is for librarians and is available to members only. Trove is for Australians and is available to everybody, anywhere in the world. I recently had a conversation with a senior representative of another GLAM sector, who noted that his constituents see Trove as 'built by libraries, for libraries'. I was surprised at this because the reality is that Trove was built by the National Library of Australia, is populated with content from every sector, and of course we think it is for Australians, not for libraries at all. At the risk of overwhelming with complex systems diagrams, this one provides a simplified look at how Trove works

Libraries Australia – on the left – is just one of the feeds into Trove. The LA records that flow to Trove are aimed at supporting discovery and are converted from MARCXML to a fairly simple DC format. Trove is indexed by Google, Trove’s size and consistency means that it is ranked highly there, and in fact more than 75% of all visitors to Trove arrive via Google. We believe that the existence of Trove significantly increases return on libraries’ cataloguing investment, by making their collections more discoverable and more useful to a wider audience. Trove also has an API allowing individuals and organisations to easily access large datasets for interrogation or for re-use in other apps or web services. So Trove is itself an important syndicator of bibliographic and authority records. Between Trove and our contribution to WorldCat, I think we can be confident that resources held by Australian libraries will be discoverable well beyond the limitations of the home catalogue.

Like our colleagues at Digital NZ we work directly with custodians of digital collections - most outside the Library sector - to add their metadata to Trove. The scale of this work is fairly similar to the scale of Digital NZ work. Digital NZ have just celebrated their 150th content partner; we are pretty close to that in terms of metadata flowing directly into Trove from museums, galleries, the Australian Broadcasting Commission, historical societies etc.

We also acquire article level metadata from two vendors – Australia’s Informit, and Cengage Gale - for inclusion in Trove. The metadata is from licensed resources, and Trove users who are members of libraries who subscribe to these resources are able to access the content. In earlier, headier days we had hoped to ramp up this part of the service - to effectively offer a full discovery layer - but experience taught us that this is a expensive undertaking and not something we can offer in a service which is fully funded by the National Library, or in an environment where – unlike New Zealand – it seems unlikely that we will be able to achieve national licences for a core set of databases.

In addition, we acquire all metadata from Australian university research repositories. There is no other aggregation of this metadata, so Trove is effectively the national discovery service for recent Australian research. We would like to refine discovery of these research outputs, especially now that National Health and Medical Research Council and the Australian Research Council grants are contingent on this publicly funded research being made freely available to the public. But as good librarians you will all know that what we can do with the data is only as good as the data we get – and Australia’s universities will have to improve their metadata to achieve their aim of showcasing Australian research.

And of course then there is Trove newspapers. We provide access to more than 12 million digitised and searchable pages of Australian newspapers, with this number increasing all the time.

Put together, all these resources are very attractive, and useful for Australians. We are delighted with the results of a large scale 2013 survey of Trove users which proved that Trove usage is proportional to Australia’s geographic population distribution, and that we have finally been able to reach out to Australians in big and regional cities, regional areas, and even the very remote outback. This is a big deal in a country as big as ours.

Trove is big in every way. Nearly 400 million resources, attracting more than 20 million visitors in a year, from every corner of Australia, and an average of around 75,000 unique visitors each day. While four-fifths of that use is clustered around newspapers, this still means that Trove is bringing a large number of eyeballs to other content, including all those bibliographic and authority records that our members create and add to Libraries Australia.

Trove’s content is growing fast, as is use

with use nearly doubling over the last 2 years.

Engagement with Trove is also increasing. We engage with our users through a blog, Twitter and face to face presentations. They engage with each other through their own blogs and our user forum. And of course they engage with our content. In Trove’s less than five year history, the equivalent of more than 310 full time equivalent volunteer years have been devoted to correcting newspaper text in Trove, not to mention addition of tags, comments and the creation of user lists. Trove is becoming a part of the lexicon. Over the last 18 months ago we have realised that our users have use the verb ‘to trove’, refer to themselves as ‘Troveites’, collaborate on Trove Tuesdays, where they share discoveries on a single blog, and just last week our Trove Manager, Tim Sherratt heard a bunch of digital humanists asking whether a particular data source was ‘troveable’. So in the spirit of this linguistic creativity, we’ll be celebrating our five year anniversary this year with ‘Trovember’...

If you want to know about Trove, do explore the service, and our many presentations available via the National Library’s staff papers site.

None of this work – in the Libraries Australia or Trove spaces – is easy, all of it requires considerable intellectual, creative, technical and financial resources and we receive no special funding from Government to create or sustain our services. We operate both services on lean budgets, and our dreams and desires for ours services always exceeds what we can achieve.

We have a clear sense of direction for both services,

but behind them are always bigger questions. How can we maximise the impact of work done by Australian libraries and collecting institutions? Where can we – as a National Library – have the most impact for the people of Australia? Libraries Australia has – for decades – been the best, the most efficient and the most efficient way of maximising the effort put into describing library collections. But will this continue to be so? Will it be the case that this work will be more effectively done at the international level in the near or far future?

I have spent much of my career working with archives, so for me, context is everything. So when we ask ourselves these questions at the National Library of Australia, we need to consider the context of what we know about use of our services, the context of what our users tell us about their use of these services and the value they place on them, the context of what new technologies make possible, and – this is very important – a context in which we have a sophisticated IT department which is able to work directly with our Australian users to deliver a service which will be really meet their needs.

If any of these contexts changes – if service use plummets, or our big libraries decide they can get better value elsewhere, or there is another, as yet unknown disruptive technology, or if our Government decides that shared services should dominate and that Government agencies should not deliver services which can be purchased in the open marketplace, even if less effectively and efficiently – then obviously our thinking would need to change. I have no crystal ball, but I am confident that our current suite of Libraries Australia services – a national collaboration – will offer the most effective and efficient service for the next five years or so. Beyond that, I think it likely that international aggregation may be more effective, in which case the National Library’s main aims will be to ensure the best possible Australian National Bibliographic Database to support our Australian focused services, and to find new ways to harness and work with the wonderful community now clustered around Libraries Australia.

National collaboration on bibliographic services has been our focus for 30 years, but might it be that national collaboration around digitisation, or digital preservation, or access to commercial digital content, or user engagement with digital content will be where we can be most efficient and most effective, where we will maximise the collective impact of our work? Ask me to come back in 10 years, and I might be able to answer you!