Issue 95

SCIS cataloguers use a combination of SCIS Subject Headings (SCISSHL)(1) and Schools Online Thesaurus (ScOT)(2) when adding resources to the SCIS database. ScOT is a relatively later addition to the SCIS arsenal of cataloguing tools, introduced and explained in Connections Issue 60(3).

This article evaluates this two-vocabulary approach with a view to distinguishing the strengths and roles of these vocabularies within education metadata.

Subject Access

MARC21 standards define a number of 'subject access' fields – the 6XX fields – some or all of which are used in each SCIS record. As a rule, every SCIS record includes entries in the topical terms (650) field. The term 'topic' will vary from one standard to another; for example, 'categories', 'keywords', or 'concept terms' are used interchangeably with 'topic'. Somewhat confusingly, 'subject' is commonly used to denote a similar purpose (the 'DC.Subject' element of the Dublin Core Metadata Initiative is effectively a topic field).

Topics are flexible and robust; they simply let the cataloguer (or user-tagger) define what a work is about. Entries are typically single words or multi-word terms, but not clauses or sentences. A topic could be 'physics' or 'nuclear physics', but not 'this is an article about nuclear physics'.

Article may not itself be a suitable topical term, if the work referred to is an article, and not about an article. For example, a work may be about some kind of genre or form, as in 'history of 19th century science fiction', where it is about science fiction, but is not itself science fiction.

Genre is another subject access field that provides a useful contrast to Topics. The MARC 655 field is reserved for genres such as science fiction or horror, or forms such as directories or diaries. In fact the genre versus form distinction is somewhat blurry.

Geographical terms (MARC 651) have a similar purpose to Topics but differ in scope. As the name suggests, this field is for storing subjects concerning places, environments, built structures, regions, and continents.

The 600, 610 and 611 fields deal with names that are the subject of a work: personal, corporate and meeting names, respectively. Proper names are a special challenge for any kind of authority control. Indeed, it may not be feasible to control a vocabulary of names, but rather to establish rules for their construction as the need arises. This is standard practice in cataloguing and in various metadata standards. Name authority files are a reflection of actual names stored in a bibliographic database, rather than a predetermined list that constrains which names may be used. Name authorities improve retrieval by ensuring that all works about a given person, corporation, or meeting are clustered around the exact same name, making it easier to identify those resources within a single set of results.

SCIS Subject Headings

Each of these subject access fields is hand-crafted by SCIS cataloguers using the SCISSHL as a controlled vocabulary and system of rules. SCISSHL provides a controlled vocabulary for the topics field, as well as rules around how and when to construct sub-divisions, for example 'Nuclear energy – Economic aspects'.

SCIS Standards for Cataloguing and Data Entry(4) provide detailed instructions on creating genres, geographical names, and each of the names-as-subjects fields. Each of the 6XX fields use scisshl as the source of the heading or term in subfield $2.

As well as providing guidelines and authority for creating a range of subject access fields, SCISSHL is a faithful reflection of what is in the SCIS database. All 1.4 million items added to the SCIS database are candidate sources of warrant for the SCISSHL. That means that each time a new topic, geographical feature, or personal name is identified in a new work, the SCISSHL are updated, making them a thorough description of what SCIS records are about, and of what kind of resources make up the SCIS database.

SCISSHL also provide see and see also references, which help users navigate subjects and negotiate synonyms and name-form variants. This provides users with useful search features such as 'did you mean' and 'you might also like' search tips.

Schools Online Thesaurus (ScOT)

In contrast, ScOT is unfaithful.

ScOT has provided SCIS with a controlled vocabulary for subject access – and especially for topic access – since 2006. But ScOT has also been busy elsewhere, serving cataloguing needs in other education repositories. A significant program with a strong recognisable brand was The Le@rning Federation (TLF). I started working on ScOT soon after TLF commenced. TLF at first sourced high-quality interactive learning objects and delivered tailored 'learning object metadata'(5) records into systems such as TALE (NSW), The Learning Place (QLD), FUSE (VIC) and Digistore (NZ). Today, Scootle (various sectors in Australia) is the most common portal through which teachers access the rich repository of learning content arising out of TLF and its successors(6).

TLF broadened its scope, creating metadata for cultural resources from galleries, libraries, archives, and museums (sometimes called the GLAM sector). ScOT did a lot of its growing up in amongst this vibrant and somewhat ambitious content program. ScOT provided topic access to resources that varied in levels of interactivity, learning design, aggregation level, content and media types, and intended user levels.

ScOT vocabulary developed in response to diverse metadata sources, but also in response to curriculum frameworks. In an attempt to align the taxonomy structure with State and Territory portals, ScOT incorporated terminology and structure from many curricula. This 'top-down' pressure on ScOT has resulted in a neat triangular shape of terms, with ten terms at the top, about 100 at the second level and about 1,000 at the third. ScOT still holds this hierarchy shape today.

The timing was right for ScOT when the Australian Curriculum was initially released in 2011. ScOT was identified by Australian Curriculum, Assessment and Reporting Authority (ACARA) as a suitable vocabulary with which to index, or 'catalogue' the Australian Curriculum Content Descriptions. This was a path less travelled, as ScOT was going places that subject vocabularies do not ordinarily go. As well as being used to describe information resources, ScOT has been used to describe the source of requirements for those resources: the educational objectives that we call 'curriculum'.

So not only has ScOT been used to catalogue records outside of the SCIS database, ScOT has been used to catalogue objects not usually considered information resources. These unique circumstances have positioned ScOT as a powerful curriculum-alignment tool. ScOT has a relatively one-track mind: aligning information resources used in schools with curriculum frameworks. ScOT plays this role in SCIS records, but also wherever curriculum alignment is needed. ScOT can be used in many different metadata standards such as MARC21, Dublin Core, ANZ-LOM and Achievement Standards Network (ASN), to name a few that are commonly used in the education sector.

ScOT can also be further adapted to any curriculum or teaching framework. Work has recently completed to align ScOT with the New Zealand Curriculum for Mathematics and Statistics(7).

Complementary approaches

ScOT does not and never will provide a basis for constructing subdivisions or for authorising names; nor will it ever be reasonable to expect that ScOT will absorb every specific topic taught or learned in schools. ScOT is involved in too many environments for that to be viable. The purpose and strength of ScOT is in providing a vocabulary with which to indicate topics, with a vocabulary immersed in the language and structure of curriculum; and the learning resources circulating within the Australian and New Zealand context since 2004. ScOT effectively provides a bridge between collections described with different metadata standards, and between curriculum and resources.

SCISSHL is closely adapted to the requirements of the SCIS database. There is nothing in the SCIS collection that SCISSHL cannot describe. It is by its very nature self-adapting as items are added to an ever growing collection of 1.4 million resources. SCISSHL provides the services that ScOT cannot: proper nouns, subdivisions, geographical, and genre headings.

Therefore, ScOT and SCISSHL have complementary roles. Together they ensure that SCIS records are thoroughly described and thoroughly connected to the wider school education environment. I often come back to a simple model published by Peter Morville in 2005(8) where metadata elements are plotted on a continuum between 'description' and 'identification'. We often refer to metadata as something that is plotted in the centre, but of course it is all metadata from the far left (eg an abstract) to the far right (eg an ISBN).

Subject vocabularies are primarily tools for supporting robust, consistent and intuitive descriptions of resources. Plotted on this continuum, I'd put SCISSHL closer to the description side and ScOT closer to the middle. ScOT goes some way to identifying resources within formally recognised frameworks. We can make simple assertions about resources that have ScOT terms: the terms may indicate that the resource is relevant to this curriculum objective or that teaching standard. Classification systems like Dewey Decimal go even further in this direction, and their descriptive power is narrower, more precise and far more selective and arbitrary.

A metadata strategy should consider the relative strengths of vocabularies and other encoding systems. SCIS exploits the richly-faceted subject access fields to create neutral, 'agnostic' descriptions about school resources with a view to faithfully representing content and membership within a significant collection. SCIS also links its records outside of its immediate environment with topics and classifications. The SCIS strategy requires significant investment, which is well warranted within the school education sector.

References

1. The "L" is for List – sometimes part of the acronym.
2. Both SCIS Subject Headings and ScOT are registered with the MARC Standards Office: http://www.loc.gov/standards/sourcelist/subject.html
3. http://www2.curriculum.edu.au/scis/connections/issue_60/scot_in_scis_-_more_of_the_same__or_different.html
4. Registered with the MARC standards office: http://www.loc.gov/standards/sourcelist/descriptive-conventions.html
5. IEEE-LOM, later adapted for the regional context as ANZ-LOM Metadata Application Profile: http://ndlrn.edu.au/metadata
6. TLF was superseded by National Digital Learning Resource Network.
7. http://schoolsonlinethesaurus.edublogs.org/2015/07/24/maths-refresher/
8. Ambient Findability.O'Reilly.ISBN 978-0-596-00765-2.