3.1 The future of metadata
Read
The first two sections of chapter 9 (pp. 215-217), in Hider, P. (2018). Information resource description: Creating and managing metadata (2nd ed.). Facet Publishing.
Three Approaches to Information Access Provision
In the online world, information access can be provided using one or more of three general approaches. The first approach bypasses metadata entirely, using computers to analyse the content of information resources to find those most relevant to users’ needs. The second approach, Web 2.0, involves end-users, contributors, and authors providing the metadata. The third, traditional approach involves information professionals organising and describing resources in ways they consider most effective for their clients.
These approaches reflect different perspectives on the nature of information resources. Every information resource contains a message that can be viewed objectively, independently of the creator or recipient, and analysed by a third party. This perspective is adopted by content-based automated retrieval systems. Alternatively, the message can be analysed based on the creator’s intended meaning, reflected in metadata provided by authors and publishers. Lastly, the message can be analysed for its utility, often associated with user-generated metadata.
Information professionals might adopt any of these perspectives, ideally bridging the gap between author and user by identifying aspects of the message that the user may not be aware of and aspects the author may not have considered. While information retrieval increasingly focuses on computer processing of content, information organisation develops tools and standards based on user and literary warrant, approximating the author’s viewpoint. Professional indexers often provide entries that reflect the author’s language and the audience’s needs and interests while incorporating their third-party perspective. Library cataloguers build common ground with other cataloguers, authors, and users (Šauperl, 2004).
Three major movements in information retrieval were outlined in the first chapter. The first movement standardised library cataloguing and classification principles in the 19th century. The second was the documentation movement, evolving into content-based information retrieval in the 20th century. Another strand of this movement continued to emphasise controlled vocabularies and classification for subject access. Over time, these traditions merged with library and information science, creating a body of theory and practice that influences many information domains despite the dominance of content-based information retrieval.
The Rise of Google
Google and other major Internet search engines have profoundly influenced information organisation, challenging the traditional bibliographic approach in two key ways:
- Are traditional tools like library catalogues becoming too cumbersome compared to search engines?
- Is adding metadata to resources and their records necessary when content-based retrieval can work effectively in many cases?
While library catalogue interfaces are becoming more similar to Google, they still rely on both metadata and the resources they catalog. With many collections now containing more digital than hardcopy resources, it’s worth considering if the content of most collections will eventually be searchable, potentially reducing the significance of metadata. The answers to these questions are both practical and theoretical and depend on:
- The specific information content being searched
- The searcher’s needs
- The nature of the information resources being searched
For example, Trove can search the content of an increasing number of newspapers. This raises the question: will metadata become less significant, or does it still offer advantages that content-based retrieval does not?
Read
‘Content-based information retrieval’ in chapter 9 (pp. 217-219), in Hider, P. (2018). Information resource description: Creating and managing metadata (2nd ed.). Facet Publishing.
Content-based Information Retrieval
Computers have advantages in analysing the content of information resources and processing vast amounts of content quickly and consistently. Content-based retrieval systems are scalable and economical, often performing well in many search contexts through keyword searching.
However, their effectiveness has plateaued in basic text retrieval and is limited when dealing with non-verbal information, such as images or music, where interpretation is more complex. Automated systems also struggle to support selection and exploration tasks where users are uncertain about their needs. Despite progress in content-based retrieval, human expertise remains crucial for intellectual tasks like indexing, abstracting, and creating search strategies.
Computers excel in retrieval and mining but are less capable of organising information in a way that makes sense to users. Human effort in building knowledge structures to support browsing and searching can be valuable in closed information environments. The challenge remains to determine the appropriate balance between human indexing and machine information retrieval (Barnum et al., 2004).
Social Metadata
Non-professional contributions to metadata have significantly impacted the field of information organisation, though the extent of this impact is still unclear. Social tagging and folksonomies, similar to search engines, are now permanent fixtures. The “wisdom of the crowd” is a powerful force; however, it is important to recognise that, just as some cataloguers are more knowledgeable than others, the same applies to crowds.
Read
‘Social metadata’ (pp. 219-222), in Hider, P. (2018). Information resource description: Creating and managing metadata (2nd ed.). Facet Publishing.
Social Metadata
The second approach to information retrieval involves users, contributors, and other non-professionals providing metadata, referred to as “social metadata.” The contemporary online environment allows the general public to contribute both resources and resource descriptions. The primary advantage of social metadata is its low cost, though it requires management and monitoring. It is also potentially far more scalable than professional descriptions. The key question is whether social metadata is good enough.
The quality and nature of social metadata have been analysed in various studies, yielding mixed results. Advocates for user-centric metadata argue that end-users provide the most relevant insights. Users can offer valuable contextual information about a resource’s utility, leading to richer indexing (Rorissa, 2010). However, research shows that user-generated tags often include personal meanings with limited application for document retrieval. Consistency is another issue; different users have different perspectives and needs, which can lead to greater recall but lower precision.
Many retrieval systems using social tagging deal with subjective content, like images, videos, and creative works, where multiple interpretations exist. In such cases, recall may be more important than precision (Hughes and Rafferty, 2011). Social tagging may not always represent the views of all users, as those who provide metadata might be biased. Professional metadata might be more representative of the audience as a whole.
Another practical concern is whether users will provide rich and accurate metadata without compensation. Users with a personal connection to a resource are more likely to tag it but might provide misleading metadata. To represent the “wisdom of the crowd,” a broader range of users needs to be encouraged to tag, perhaps by making the activity enjoyable or gamified (Konkova et al., 2014; Flanagan and Carini, 2012; Mandel and Ellis, 2008).
Libraries have had limited success in getting patrons to tag or rate loaned items, despite the potential for social engagement (Tarulli and Spiteri, 2012). Non-professional metadata can suffer from carelessness and lack of technical expertise. However, folksonomies may keep up with linguistic and conceptual developments better than controlled vocabularies, which often fail to cover all subject queries (Lund, Ng and Pera, 2009; Mering and Nowick, 2003; Carlyle, 1989).
Efforts to make folksonomies more structured may miss the point, as their value lies in their lack of control (Chan and Yi, 2009; Hunter, 2009). Folksonomies can supplement but not replace controlled indexing (Rolla, 2009). Instead, social tagging and professionally applied controlled vocabularies should be seen as complementary (Stvilia, Jørgensen and Wu, 2012). Social metadata can cover elements not addressed by professional indexing, including resource quality through ratings and comments. The challenge is to integrate professional and social metadata effectively, with various approaches being tried and no definitive solution yet clear.
Information Professionals and Metadata
Many people still make a living by describing and analysing information resources to make them accessible to others. It is difficult to determine whether their numbers are increasing or decreasing overall. However, it is evident that some areas have seen a decline, while others have experienced growth. For instance, fewer people are paid to write abstracts nowadays, whereas more people are employed to organise websites than twenty years ago. Library cataloguing remains the largest pool of metadata specialists, but the nature of their work is evolving. For professional description to continue thriving, end-users must appreciate and use the tools created by cataloguers and indexers. Additionally, managers of agencies employing cataloguers and indexers need to recognise their value. In today’s world, everyone needs to promote their skills, which metadata specialists have not always excelled at doing. Information organisation skills are challenging to market due to their intangible results, but they need to be promoted nonetheless.
The Semantic Web
There has been speculation that the Semantic Web will herald a new golden age for cataloguers, with RDA paving the way for this by enabling cataloguers to create metadata in structures that can be processed by Semantic Web ‘agents’. These agents are computer applications that scour the Internet, linking data from different sources to create a coherent picture of the information world and report back to humans.
The Semantic Web is best understood as a vision or goal. It aims to link data between systems or entities, allowing rich connections and relationships between the data available on the Internet to be realised. This marks a shift from publishing data in human-readable HTML documents to machine-readable documents, enabling machines to perform more background processing.
Currently, much of the data on the Internet is delivered as web pages—HTML documents linked by hyperlinks. While humans or machines can read these documents, machines struggle to extract meaningful information beyond finding keywords.
The Semantic Web addresses this issue by:
- Opening up the web of data to artificial intelligence, enabling the Internet to process information for us.
- Encouraging companies, organisations, and individuals to publish their data freely in an open standard format.
- Promoting the use of available Internet data by businesses (data give/take).
Read
‘Linked data and the Semantic Web’ (pp. 229-232), in Hider, P. (2018). Information resource description: Creating and managing metadata (2nd ed.). Facet Publishing.
Linked Data and the Semantic Web
The future of cataloguing and professional information resource description is argued to lie with linked data and the Semantic Web. Although human searchers can often manage without controlled vocabularies, these vocabularies provide the context and meaning on which Semantic Web technologies rely. Structured metadata aids humans in navigating collections and helps computers make sense of content, enhancing the utility of information resources.
The Semantic Web differs from the original Web (1.0), which is a vast collection of hyperlinked resources, and the Social Web (2.0), which allows for more user input and interaction. The Semantic Web, a mostly unrealised vision, aims to transform the web from a collection of human-handled resources into a data collection processable by computers. This does not replace the existing web but makes its content meaningful to both humans and computers.
Semantic Web technologies could revolutionise information by allowing computers to handle much of the synthesis and comparison of online data. Linked data can be used to analyse various resources, with schemas and vocabularies being developed to describe a wide range of entities, including people. These structures represent different knowledge domains in a way that computers can process, linked through devices such as namespaces and standards like RDF.
The information profession has been keen on the Semantic Web’s promise, seeing linked cataloguing data as key to fully exploiting library content. Leading organisations like LC and OCLC have aligned their metadata standards with Semantic Web standards and published schemas and vocabularies as linked data. The concept of sharing metadata for the community’s benefit is not new to libraries, which have been sharing bibliographic data for over a century.
The future importance of professionally created metadata and its schemas and vocabularies depends on the success of the Semantic Web. The term “Semantic Web” has been partially replaced by “linked data,” a term easier for the layperson to understand and reflecting a push to popularise the concept with simpler, more accessible standards.
The success of Semantic Web technologies relies on the creation and publication of data, similar to how Web 1.0 and 2.0 succeeded through website creation and social media usage. The process must be scalable and sustainable. Challenges include inferential tractability, logical consistency, the ever-changing web, and the need for political will and resources. Existing vocabularies need cleaning and updating to be effectively used by Semantic Web technologies, requiring significant effort.
While some specific Semantic Web applications may meet real-life needs, it is unlikely they will become pervasive soon. Currently, most Semantic Web work is conducted in research labs. For instance, the Linked Jazz Project experiments with linked data in a music context.
Semantic Web technology has not yet become a “disruptive innovation” with an established user base, spontaneous use-case scenarios, or widespread market buy-in. Information professionals should maintain a watching brief, recognising that controlled vocabularies provide context for both computers and people.
Outline of how Semantic Web technologies could revolutionise our information environment:
3.2 The future of metadata in school libraries
Current and Future Issues for Metadata in School Libraries
1. The “We Don’t Need a School Library Now That We Have the Internet” Argument
The notion that school libraries are obsolete due to the abundance of online resources has been a contentious issue, causing disruption in some schools. Schools that have eliminated or downgraded their libraries often find themselves at a disadvantage, as the tailored library collections and effective retrieval methods they provide are lost. Metadata is crucial in countering the push to abolish school libraries, as it organises information for effective and timely retrieval. While the importance of online resources is undeniable, school libraries should adapt by offering better access to these resources and educating users on how to retrieve them. The future school library might be more virtual, with teacher librarians at its core, identifying relevant resources and ensuring efficient access to them.
2. The Evolution of OPACs in School Libraries
Teacher librarians are enhancing their catalogues with advanced functionality and interfaces. Next-generation catalogues offer features like faceted navigation, relevance ranking, and hierarchical presentations. The ideal would be for school library OPACs to operate like Trove, with new features available either as system components or as add-on “discovery layers.” However, this shift sometimes sacrifices the stand-alone catalogue, favouring federated search interfaces, such as Primo. While this greater interoperability is beneficial, it can also result in data and metadata loss, raising concerns about whether essential elements are being overlooked.
3. The Future of Subject Access in School Libraries
The future of subject access in school libraries raises questions about whether subject heading lists like SCIS Subject Headings will be replaced by thesauri like ScOT (Schools Online Thesaurus). Developments in keyword searching could reduce the value of the controlled vocabulary approach. Subject access is needed for users to search for resources, making these developments significant. Karen Calhoun’s 2006 report for the Library of Congress questioned the relevance of Library of Congress Subject Headings in the age of Google. Content-based retrieval has excelled in subject access, offering relevant results with simple keyword searches. However, library catalogues can frustrate users who fail to enter the correct authorised headings, resulting in ‘zero hits.’
Controlled vocabularies still add value in many search situations, especially when users cannot guess the terms, cannot think of a term, or are unsure of their needs. For catalogues to assist users in these situations, users need to understand how to use the controlled vocabulary. Unfortunately, users are often unaware of the existence of the vocabulary or how to access its structure. The vocabulary must be current and intelligible to users.
While content-based retrieval challenges controlled subject indexing, library classification schemes face challenges from the web itself. Users do not browse virtual shelves as they do in physical libraries, possibly due to the convenience of search boxes or the absence of browsable virtual shelves. Schemes like DDC work well in physical libraries and might also succeed in the future’s digital libraries.
4. The Future of SCIS
The future of SCIS (Schools Catalogue Information Service) is another critical issue. While many Australian teacher librarians might take SCIS for granted, its continuity cannot be assumed. SCIS provides suitable catalogue records for school libraries, freeing teacher librarians for other roles. However, some SCIS cataloguing agencies have disappeared, with contract cataloguers replacing government agencies. Tasmanian and Northern Territory secondary schools share catalogue records with state or territory library services. Given the era of government service rationalisation, teacher librarians might need to use broader cataloguing agencies that focus less on school library contexts. SCIS needs active support and promotion by teacher librarians to maintain its role and value within the school community.
These scenarios highlight present and potential future issues in information organisation within school libraries. Staying aware of changes in the field is crucial to maximising users’ benefits.
Online read
Tarulli, L., & Spiteri, L. (2012). Library catalogues of the future: A social space and collaborative tool? Library trends, 61(1), 107-131. Retrieved from https://dalspace.library.dal.ca/bitstream/handle/10222/15571/06_61_1_tarulli_107-131.pdf?sequence=1
Library Catalogues of the Future: A Social Space and Collaborative Tool?
Overview
Next-generation library catalogues are evolving into social and collaborative spaces that enhance traditional library functions. These catalogues provide opportunities for user interaction, creating an online environment that mirrors the physical library experience. The integration of technology, innovation, and creativity is key to this transformation, allowing library professionals to engage with users in new ways.
Key Concepts
- Enhanced Interaction: Modern catalogues facilitate patron interaction with library staff and other users, promoting a seamless and enjoyable experience.
- Collaboration and Customisation: These catalogues allow for data manipulation and personalisation across various devices, offering users a flexible and user-friendly interface.
- Extended Library Services: They extend the physical library’s reach, offering services remotely and integrating social features such as tagging, reviews, and user-generated content.
- Innovative Opportunities: The potential for catalogues to serve as tools for readers’ advisory services, collection development, and youth services is significant. They provide a platform for community engagement and information sharing.
Impact on Library Services
- Readers’ Advisory Services: Next-generation catalogues can transform traditional readers’ advisory models by enabling remote interaction and personalised recommendations.
- Traditional Library Services: They enhance the capabilities of reference, children’s services, and collection development by offering sophisticated search functionalities and integrating data from multiple sources.
- Youth Services: These catalogues engage younger patrons through social features and customisable platforms accessible via computers, tablets, and smartphones.
- Book Clubs and Adult Services: They facilitate book discussions and community-generated content, providing adults with convenient access to library resources and services.
Technological Integration
- Mobile Applications: The development of mobile-friendly catalogues is essential, as users increasingly access library services via smartphones and tablets. These applications should be intuitive, offering quick and easy access to the most popular features.
- Customisation and Flexibility: Next-generation catalogues must be adaptable to various devices and platforms, ensuring they meet the evolving needs of users.
Challenges and Considerations
- Collaboration between Staff: These catalogues foster a collaborative environment among library staff, breaking down traditional barriers between cataloguers and frontline staff.
- User-Generated Content: While inviting users to contribute content, it is essential to maintain the integrity of bibliographic records and ensure consistency.
- Adapting to Change: Libraries must continuously adapt to technological advancements and changing user expectations, leveraging next-generation catalogues to remain relevant and effective.
Conclusion
Next-generation library catalogues have the potential to revolutionise the way libraries interact with their communities. By embracing these technologies, libraries can offer a more engaging, collaborative, and personalised experience, extending their services beyond physical boundaries and into the digital realm. These catalogues not only enhance access to collections but also serve as vital tools for community building and information sharing.
3.3 What about AI?
Online read
International Federation of Library Associations and Institutions (IFLA) (2020). IFLA Statement on Libraries and Artificial Intelligence. Retrieved from https://repository.ifla.org/handle/123456789/1646 Link currently not working
Online read
Tanzi, Nick (2023). How can AI be used in libraries? The Digital Librarian, https://the-digital-librarian.com/2023/01/23/artificial-intelligence-in-the-library/
Artificial Intelligence in the Library
AI technology is increasingly impacting libraries, enhancing various services and operations. Key applications include:
- AI-Powered Patron Assistance: Chatbots provide 24/7 support, improving user experience.
- Marketing and Design Tools: AI-generated images offer cost-effective, customisable options for libraries with limited budgets.
- Library Programs and Services: Libraries educate patrons on AI, offering hands-on experiences with AI tools.
- AI in Collection Development: AI tools aid in analysing materials, improving accessibility, and diversifying collections.
AI is poised to revolutionise library services, offering both challenges and opportunities for the future.
As artificial intelligence continues to grow by leaps and bounds, how is AI technology manifesting itself in our public, academic, and school libraries?
Watch
Generative AI in a Nutshell – how to survive and thrive in the age of AI
The video explores the transformative impact of generative AI, contrasting it with traditional computing, which has functioned as a glorified calculator executing predefined instructions. Generative AI, demonstrated by models like GPT, enables computers to learn, think, and communicate similarly to humans, undertaking creative and intellectual tasks previously exclusive to humans. This technology marks a significant shift, providing intelligence as a service that is rapidly improving and accessible to individuals and companies globally.
A helpful analogy in the video is “Einstein in your basement,” symbolising AI’s collective knowledge and expertise. However, generative AI’s effectiveness is limited by the user’s imagination and communication skills, known as prompt engineering. Mastering this skill is needed in the AI era to unlock generative models’ full potential.
The video clarifies that while AI is not new, generative AI specifically refers to systems creating original content rather than merely sorting existing information. ChatGPT by OpenAI is an example of a large language model (LLM), generating human-like text using advanced architecture.
LLMs function as artificial neural networks processing numerical representations of language. Extensive training in vast text amounts enables these models to predict and generate coherent responses. Human feedback during training ensures the models produce appropriate and responsible content.
The emergence of generative AI models has resulted in diverse applications, from text generation to image creation and video production. These advancements present challenges and opportunities for individuals and businesses, fostering workplace adaptability.
A central theme of the video is the importance of a balanced mindset towards AI. While some fear job displacement, the speaker argues that partnering with AI can boost efficiency and creativity. Human expertise remains crucial in guiding AI applications, ensuring outputs are evaluated and contextualised.
The video concludes by highlighting generative AI as a powerful productivity tool. Understanding and mastering prompt engineering is vital for effectively leveraging AI. Viewers are encouraged to experiment with AI, recognising that initial challenges can improve communication skills and lead to innovative applications.
How do you anticipate AI will be used in school libraries in the future?