Trove

So this blog is primarily about the website Trove. But first! I would like to comment on my Google Analytics. After not really checking my Google analytics for a while, I thought I would take a quick look today. Google Analytics allows me to track who is looking at my blog, from what country etc. Today I saw that I have a number of views from people in America, more specifically that I have 3 views from Charleston, Illinois! Not only this, but they spent an average time of 7 minutes and 25 seconds on my website! Now obviously this could be a complete mistake on the part of the viewer. But in my own little naive world, I'd like to pretend that someone in America is actually interested in what I have to say. I'm really enjoying using Google Analytics and it's a great way of viewing statistics of who is viewing your blog!

Anyway! back to Trove. As part of our assessment, we have been asked to write a 500 word critique on a Digital History website. At first I was a bit dubious about posting it on here and to be honest I have no idea why. So! here is my (not quite) 500 word critique. It is only a first draft and so it is not in its final form as of yet.

My Critique! 

Trove is a federated search engine that was launched by the National Library of Australia to the public in 2009 to gather material related to Australia. This site provides metadata for over ninety million items. It is a ‘federated’ site, which means that it centralises material and collates it from one thousand libraries via the Australian National Bibliographic Database. This is opposed to a unified catalogue, which is a single data catalogue that gives the user unrestricted access to all stored resources. Trove uses OAI-PMH, which is a web-based protocol that defines a standard way to move records between resources. It allows Trove to request records from a system in batches, the results are returned in XML format. A wealth of digital content is then produced and catalogued using a standard set of vocabulary terms defined by Dublin Core. From the digital metadata tags, material is separated into eight different zones; all view, books, photos, newspapers, sound, diaries, maps and organisations. Currently, the site holds 134,283,381 Journals and articles as well as 63,693,069 digitalised newspapers.[1] Trove has an excellent interface, which is aesthetically pleasing and laid out in a logical user-friendly fashion.

Fig 1: Homepage of Trove, http://trove.nla.gov.au 

 

When searching for books, Trove will group together multiple items and different editions together instead of displaying only one record. Trove achieves this by using FRBR (Functional Requirements for Bibliographical Records), an approach to cataloguing, which takes into account the searching for different copies and specific editions. Figure 2 demonstrates how the different versions/editions of the work are displayed.[2]

                   Fig 2: The versions of the work. Found on http://www.dlib.org/dlib/march11/holley/03holley.html

 Trove’s material is derived from other organisations and museums, which may have their own access restrictions. Although Trove can take some steps to improve the interface usability, a large percentage of the responsibility falls on the contributing organisations. Some material advises the user to contact the publisher directly or purchase the item.  Trove uses the Australian Libraries Gateway (ALG) to source the library names. It also added ‘short library names’ (Fig 3) to the ALG which received positive feedback from users [3] However, when Trove undertook usability testing, it showed that “users expected to be able to borrow from any library listed in Trove,” even though many libraries have access restrictions.[4]

Fig 3: Short library names, found on http://www.dlib.org/dlib/march11/holley/03holley.html

One part of the project involved the digital cataloguing of scanned documents such as newspaper articles. These have been translated electronically into text using OCR (Optical Character Recognition) and although OCR is usually very accurate, the small font and uneven printing makes conversion difficult. Due to these conversion difficulties, more than 5,000 online users have manually corrected the text themselves.[5] This is an excellent example of crowdsourcing. The manager of Trove, Rose Holley, states that over two million lines of text are corrected every month.[6] However, work still needs to be carried out on some of the older newspapers and archives that OCR has failed to correctly read. 

Trove is a great example of how history can be presented online and the website has a number of great successes. However, it needs to improve its ability to obtain the material from its contributing sites. Furthermore, a percentage of its photos that are available are of poor quality, presumably to discourage users from using the image for their own use.

---

---

[1] Statistics taken from http://trove.nla.gov.au/system/counts <accessed, 09/03/12>
[2] Rose Holley, ‘Resource sharing in Australia: Find and get in Trove – making “getting” better, information and diagram found on http://www.dlib.org/dlib/march11/holley/03holley.html <accessed 10/03/12>

[3] Ibid

[4] Ibid

[5] National Library of Australia, ‘Online text correctors to be honoured’ posted 27^th January 2010 on http://www.nla.gov.au/media-releases/online-text-correctors-to-be-honoured <accessed 09/03/12>

[6] Rose Holley, ‘Crowdsourcing and social engagement in libraries: the state of play’ found on http://eprints.rclis.org/bitstream/10760/16385/1/Crowdsourcing%20State%20of%20Play%20June%202011.pdf <accessed 10/03/12>

Photo Credit: http://www.nla.gov.au/pub/gateways/issues/104/trove.html