The MOL hosted a most useful day out for MCG members, to describe the definition and procurement of a new computer network and collections management software. The morning was taken up with presentations on the development of the museum’s IT strategy, outsourcing of IT services, and the procurement of Multi-MIMSY. The afternoon included a series of demonstrations of aspects of the package followed by a LASSI progress report and a short panel discussion.
Peter James, IT coordinator at the museum, started the proceedings with an overview of the museum’s IT strategy. The current process – which took 2 years to complete – began with an independent review by CCTA of the museum’s IT provision which uncovered a familiar story of lack of understanding by the existing IT section of the fundamental requirements of the museum users, no practical service level agreement, and a mish-mash of unrelated software and hardware. The CCTA recommended that the museum acquire and implement:
- Client-server architecture
- LAN and WAN
- Image capability
- Windows-type software
- Industry standard hardware, communications and operating systems
- A standard for support and maintenance.
The office and collections tools of choice were
- An office automation suite (e.g. Microsoft Office)
- Specialist collections management software
- Specialist Library Management software
- Internet connection
The museum then embarked on three separate procurements:
- A main service provider to supply project management, up-to-date IT knowledge and IT support.
- Collections information and management system
- Library and archives management system
The project team was composed of members from all business areas and issued a definition of service requirements in November 1994. There were 75 responses to the European Journal advertisement for a service provider and 33 serious follow-ups. Data Sciences UK was finally chosen from a shortlist of 6. They recommended Netware, Windows for Workgroups (“since Microsoft were not being entirely honest about the future of Windows 95”), and Microsoft Office. Hardware was supplied by Data Sciences on a lease-lend basis over three years. The service contract is for 3 years after which a review will result either in a new contract or a new procurement.
The software requirements were sent to 9 possible suppliers, with the final choice of collections information and management software being Willoughby Systems’ Multi-MIMSY (MM) via their UK agent Lusis (which has grown from the Heritage Services section of the Welsh Water Board’s IT department). Agreement was reached only 2 weeks prior to the MCG meeting. The Library and Archive system (of which the choice was Oracle Library) had to be postponed following budget cuts and the library will be catalogued using the bibliographical capability of MM instead.
Expected benefits of the new systems include:
- Improved communications and document transfer (e-mail, fax, Internet)
- Common training programs
- Improved information handling
- Help desk on site
- Access to trained experienced IT professionals
- A monitored service level agreement
IT is now managed by a centralized IT Steering Committee which ensures that business needs are met; that value for money is obtained (via a newly-centralized IT budget); agrees priorities and a common policy framework; and monitors development work. Future plans include an image archive, Internet and Intranet (a web page is to be launched in 2-3 months), public access facilities in galleries (including touch screen access), multimedia publishing, and data-voice integration.
Richard Avery of Data Sciences (a subsidiary of IBM) presented the case for outsourcing – a term they had invented. Major benefits are considered to include cost saving, allowing existing resources to focus on core business needs, and access to emerging technologies and knowledge of existing ones. (For more information on the company see http://www.DataSci.co.uk.) Paul Duggan, the Data Sciences MOL account manager, commented on the basics of a successful procurement: a good statement of requirements, and a thorough audit of the existing system. It is also essential to get high-level support for the project.
It was too early to say yet what major benefits the procurement had brought. E-mail was an instant hit, but users were still getting to grips with the other software. A questionnaire had been circulated to gauge user satisfaction. In answer to queries from the floor, Richard affirmed that out-sourcing was not to be feared and that the presence of the CCTA analysis was vital to the success of the project, however, those going down this path must be prepared to pay accordingly for such detailed analysis.
Andrew Roberts described the background and planning required for the project. The MOL is looking to catalogue and manage some 185,000 objects in its core collections, a large archaeological archive, the Docklands Museum Project, nearly 400,000 photographs and ephemera items and the museum library of over 25,0000 books. The MOL’s information systems and documentation procedures were reviewed in 1989 and 1992. Some of the latter shortcomings were addressed by the installation of a Documentation Officer and Registrar in 1993, but problems with the existing IT solutions continued to grow. The in-house system did not offer a long-term solution; furthermore the development of a new in-house system was seen as too expensive and risky. A review by the CCTA recommended separate collection and library applications, and standardised hardware and office software. In June-November 1994 a detailed Operational Requirement (OR) was drawn up by a working group of lead users who examined the museum’s documentation and collection management objectives in the light of the experience of other museums, the LASSI proposals, and national documentation standards. Of the final list of functional requirements, 111 were flagged as mandatory, 104 highly desirable, and 127 medium to low desirable. The available packages were reviewed and a preliminary assessment made. The resulting shortlist of 9 vendors were sent the OR in November 1994. The responses were carefully analysed and evaluated. Only two products unequivocally provided the mandatory requirements, 4 failed. The final shortlist included Multi MIMSY and Advanced Argus. After a full analysis and detailed demonstrations the former emerged as the option of choice. It is an Oracle-based client-server application running under Microsoft Windows.
A major part of the process was data migration. Not only were there the object records to transfer, but many thousands of collections care, management and conservation records needed to be converted. Terminology standards needed to be decided upon and data cleaned to match the new requirements. Syntax and terms had to be standardised, and errors reduced. For example the 8,000 object names latterly in use have been reduced to 4-5,000 names, all matched to the Art and Architecture Thesaurus. The process took over 18 months. Rationalization of the names of places and persons offer even more of a challenge. During the transfer process some data had to be reassigned to new fields, and coded entries were expanded to full text. A great deal of customization has been found necessary to accommodate MOL data fields. MIMSY has many more fields than are actually required and care is necessary to ensure that different fields are not being used for what is essentially the same data.
Training for lead-users and initial training of end-users was still going on and it is anticipated that 6-9 months will be required to get to grips with the software. Target date for live use of the museum MIMSY database was July 1996.
After lunch, delegates were shown various aspects of the package by different members of staff. Delays in formalising the contract meant that the system had only been installed at the MOL the day before the MCG visit (!) and, owing to the so recent implementation of the package, and the staff’s understandable unfamiliarity with the software, these demonstrations often asked more questions than they answered. Nevertheless, the day was a very successful and interesting one, the venue and hospitality excellent, and MCG members will be very interested to hear more of this major development.
The final presentation was an update of the LASSI project by Alice Grant, Head of Collections Information at the Science Museum. The complexities of the LASSI Framework and Enabling Agreements were described, whereby the MGC and Willoughby contract together to provide an umbrella for consortium members who contract with the MCG to become an Authorized Demander and then contract direct with Willoughby via LUSIS, their UK agents, to acquire the product and arrange installs, data migration and training. Readers will not be surprised to learn that the development of these contacts took a long time and engaged the attentions of 8 museum services (each with their own legal departments), hosts of acronyms â€“ including MGC, DNH and CCTA â€“ Willoughby, Lusis, 16 solicitors, lots of money, time, and effort. This long-drawn-out process was illuminated by a wonderful edited photograph of the “other” Lassie with a gun to her head and the caption “Sign the contract or the dog gets it.” How we laughed…
Much work remained to be done on upgrading the product to the LASSI standard. The LASSI-compliant version of Multi MIMSY (version 2) will be available in the summer of this year.
Future plans at the NMSI were outlined, including exciting projects on public access for both real and virtual visitors with the aim of producing a multiple-use knowledge base, a £14m “imaging frontiers” project at the National Museum of Photography, Film and Television, the “Challenge of Materials” gallery, database and CD-ROM and virtual exhibition on the World Wide Web. Less sexy, but vital, background work included revised collections management procedures (with cross-fertilisation to Spectrum), and the development of a thorough names authority file.
User Friendly Terminology Control
You will be familiar with the term GIGO: garbage in – garbage out. A database will only be useful if its content is accurate, coherent and consistent. Sometimes non-specialists may enter information which leaves something to be desired. Even among period specialists, there can be a variety of terms used to denote the same object type. It is hardly surprising that confusion and frustration can ensue. The logic of the approach given below is intended to provide an example where terminology control can be helpful, rather than ogreish. Any implementation of terminology control that its worth its salt should not require detailed knowledge of the system or of its codes. Users should be able to enter natural language terms which enable searches.
Having inherited and laundered a stream of consciousness database of some 10,000 archaeological objects (described in ACN 34 and 36), it made sense to ensure that some form of checking at source was introduced wherever feasible. This task was facilitated (for specific item names) by the Bedfordshire Artefacts Typology (BAT) and generally (e.g. for period) by the National Archaeological Record (NAR) compiled by the Royal Commission on the Historical Monuments in England (RCHME).
An example BAT record from Luton Museum is shown below. The underlines show “Symbolic Data.”
|Prehistoric Flint Tools|
|15.14 axe (stone)|
|Use for hafted axes only, not handaxes, see also adze (flint)|
|Full name(s) Note|
|edge-ground chisel lne|
|Related BAT records Function|
|15.31 adze (flint)|
“Symbolic data” is provided by the system rather than keyed in. For example Prehistoric Flint Tools is derived from the first element of the BAT code (15); adze (flint) is culled from a related record. In this example, the note field is used to remind users of period (using NAR abbreviations). With use, the BAT continues to expand as more objects are categorised and added. In time, objects may be listed under more than one category.
Every archaeological artefact on Luton Museum’s database is cross-checked by a BAT record. Data entry is both more structured and simplified. Use of standard terms, expansion of approved (NAR) and user-defined abbreviations have also been implemented. Thus if a user enters ne or neo into the Period field, it is automatically expanded to neolithic. If required the process can be reversed. If we need to target or group together types of artefact, the BAT makes it much simpler to do so. The BAT is implemented using Microsoft Access at Bedfordshire County Council, and Advanced Revelation (version 2.12) at Luton Museum.
Having implemented a synthesis of the BAT with the NAR at Luton, it was considered worthwhile to put into effect some form of consistency checking. To avoid anachronisms, a program was written to check period against simple and full name entries. As the BAT is a developing system, it was thought unwise to make the period checking mandatory. The program advises the user, rather than changing any information. It should be relatively straightforward to apply the same logic using any other programming language. Copies of the code are available upon request.
Listing of BAT for flint artefacts
|Key||Simple name||Full name|
|15.1||arrowhead (flint)||barbed and tanged|
|hollow based point|
|15.13||fire fractured flint|
|15.16||core rejuvenation flake|
Bedfordshire County Council Archaeology Service, The Bedfordshire Artefacts Typology, (unpublished paper).
Royal Commission on the Historical Monuments in England, Recording England’s Past, A Data Standard for the Extended National Archaeological Record, 1993.
Thanks are due to Holly Duncan (Bedfordshire County Council Archaeology Service) for proof reading (and half rewriting) this article. The implementation of the program (to check names against period) is based upon information compiled by Dr Robin Holgate (Curator, Luton Museum Service), Holly Duncan and Carolyn Wingfield (formerly of Bedford Museum, now at Doncaster Museum). Any errors remain the responsibility of the author.
Mark McCall, Documentation Office, Luton Museum Service
Wardown Park, Luton LU2 7HA
(01582) 746750, Facsimile (01582) 746763