The main focus of HOTBED is on implementation into L&T and there is no professional cataloguing expertise within the HOTBED team. (This is because the main body of material to be digitised is from the School of Scottish Studies and is already well documented.) There is capacity in the budget to buy in consultancy and we can also expect advice from the School of Scottish Studies, the RSAMD Librarians and the Archives Consultant. Inevitably, there is tension between HOTBED as an implementation project and the extent of its responsibilities (if any) as an archive and the more sophisticated level of resource description that might be expected from an archival resource. These issues will be covered in greater detail in an evolving Accessions Policy document with advice and input from various quarters. Although the project proposal is clear that L&T issues are the main focus, a reasonable level of metadata is necessary for users to locate materials. Also the context of the DNER and expectations of RSAMD staff as potential donors need to be taken into account.

Stevie Moffat has expertise in both audio digitisation and Scottish traditional music and we are fortunate that he has the musical knowledge to make cataloguing decisions. However, we should be cautious in our expectations of what Stevie can achieve and not underestimate the level of professional skill that cataloguing entails.

In this first, prototype phase we are encoding and mounting materials provided by staff from RSAMD and the School of Scottish Studies (SoSS). As we are reliant on the goodwill and co-operation of staff we cannot be sure of the amount of metadata to be provided but it could be minimal. We have to decide whether to use Stevie’s expertise in enhancing the basic information provided (which may be, for example just title/artist) or whether to leave it in its minimal state.

One of the areas for experimentation and review in the first stage of encoding will be to monitor Stevie’s workflow in both scenarios so that we can better predict the resource necessary for future batches of encoding work. Encoding is envisioned in the project plan as a relatively straightforward task which can be carried out by student labour. Once we know more about the nature of the first batch of materials we can isolate a portion for mid- to high-level description and monitor both Stevie’s workrate and any problems he encounters. We can also monitor how much rich resource description adds value for both student and staff users and make decisions on future levels of metadata description accordingly.

In accordance with good practice, HOTBED resources should be described using a standard scheme and Dublin Core will be adopted. Although the eventual incorporation of sound materials into the DNER cannot not guaranteed due to possible rights restrictions, there are unlikely to be restrictions on the metadata itself and thus DC compliance should ensure that HOTBED records can be retrieved by searches of the DNER.

We will take advice on the current DC implementation issues, although it’s likely that those aired in the early AHDS implementations are still valid (Discovering Online resources Across the Humanities; a Practical Implementation of the Dublin Core, ed P. Miller and D. Greenstein, UKOLN, 1997). Among these issues are:

• overloading a scheme that is not intended for rich resource description (e.g. not describing non-searching items like 78 rpm source)
• tendency for fields to be more appropriate for textual resources than sound/visual resources (e.g. Contributor field)
• lack of consistency in use of fields (date, coverage).

What follows is a suggestion for 3 levels of DC metadata for Stevie to work with in the first phase. These are

• Level 1: mandatory
• Level 2: desirable: if info easily to hand, or easily determined, then it should go in
• Level 3: deluxe: likely to require Stevie to make decisions

We should strive for standardisation wherever possible, whether from Library of Congress Subject Headings (LCSH) or School of Scottish Studies classifications.

Fields marked with a * will be constant (e.g. Type = Audio), italics refer to queries later.

Queries/Issues (from version 2):

1. The 3 suggested levels and philosophy behind metadata. At crudest level, metadata primarily describes the piece of data (e.g. encoded sound data) not what it represents (e.g. a recording made in Orkney in the 1950s). So date of encoding is mandatory, date of recording isn’t.
2. There are plenty of tunes without titles which, in practice, are given working titles.
3. Need to look at LCSH for general headings and more particular, for Subject field.
4. Is there an archive standard for Scottish/other regions?
5. Need DC type advice on Source and crossovers with Publisher for archival materials, LCSH – for now follow Patron and use Source_format.
6. Overall rights statement – CD to agree with lawyers.