Hardware: Networked PCs linked to server.
Operating System: Microsoft DOS version 5.0 in 1998. By 2000, Windows 95.
Application Software: Microsoft FoxPro version 2.5 for DOS. The Core Database project in 1998-1999 saw the transfer of both databases to an Oracle 7 database. This was used by the MGC and Resource in conjunction with a report writing package, Arpeggio.
User Interface: Originally based around a series of windows (referred to as 'screens'), which allowed staff to move through the data relating to an individual museum. They included:
- 'Address' screen giving basic details for each museum (DOMUS number, name, address, ownership, DOMUS contact).
- 'Facilities' screen giving details of the facilities and types of collections held by a museum.
- 'Museum Info' screen which repeated address information, and gave the type of museum, its status under the MGC's registration scheme, its registration number, the museum service, the museum's Registered Charity Number, Registered Company Number and VAT Number, the local authority, year of the museum's foundation, date of the last visit by the MGC's museum security adviser, and general comments.
- 'Museum Statistics' screen which displayed the 'Annual' information collected in DOMUS.
Following the suspension of the Core Database project a subsequent user manual indicates that users of the merged DOMUS and Registration Scheme databases were provided with a typical Windows-style GUI interface consisting of windows, icons, menus, forms and dialogue boxes. As with the earlier version DOMUS, the interface was built around a series of windows ('screens').
Logical structure and schema: Until the 1998-1999 Core Database project, the DOMUS database had a relatively simple logical structure, consisting of eight core tables plus tables which were added for the one-off supplementary questionnaires. This structure is reflected in the first two DOMUS datasets. The structure of the final DOMUS dataset is more complex. The dataset consists of 42 tables. Fourteen tables appear in an entity-relationship diagram which was produced by the contractor in April 1999.
It is more difficult to understand the intended function of the other tables, which include:
(1) Tables which appear to be designed to facilitate 'fuzzy' searching (e.g. T_FUZZY_INST, T_ADMIN_DSCT).
(2) Tables whose precise function is unclear (e.g. EVT_PROFILE, EVT_PROFILE_EVENTS, AQ$_QUEUE_TABLES).
(3) The table relating to the Designation Scheme (DESIGNATED_COLNS).
(4) Tables which cover the same data as other tables, but present it in a different format.
It is possible that some of the tables in the third dataset are intermediate tables that were not intended to appear in the final database: e.g. T_INST_FROM_INST, T_INST_FROM_DOMUS_AND_BRANCHES, SUPPLEMENTAL_OLD.
How data was originally captured and validated: DOMUS data was gathered through paper questionnaires which were sent to museums which were part of the Registration Scheme. The response rate was fairly high: 86% in 1994, 77% in 1995, 85% in 1997. From 1997 questionnaires were sent to museums in June of each year and were returned directly to the MGC. Prior to that date questionnaires were sent to museums at staggered dates throughout the year.
The main DOMUS questionnaire appears to have been relatively stable over the life of the survey. The MGC tended to add new questions or categories rather than to withdraw topics which had previously been asked.
These changes in the questionnaire are reflected in the datasets, particularly in the tables recording 'annual' information. In some records, certain fields may have missing or default values because the field relates to a question which was not added until a later sweep of DOMUS.
The inputting of data from the forms was done in-house by the MGC in all DOMUS surveys except for the 1997 survey, when an external contractor was employed to enter data onto the database. The MGC returned to in-house inputting in 1998. The MGC would contact museums to resolve any ambiguities in returns. A telephone survey involving 45 interviews with DOMUS respondents was conducted by the Museum Documentation Association in 1996 to evaluate the accuracy of data gathered in the survey and attitudes towards DOMUS.
As part of the UK Museums Retrospective Statistics Projects, the Library and Information Statistics Unit (LISU) of Loughborough University attempted to resolve a number of anomalies in the DOMUS data. Data was input from 80 questionnaires from the 1999 survey which had been returned but not entered onto the system. LISU did not input data which was regarded as not being relevant to statistical analysis, such as details of individual contacts. A few records were deleted which were thought to contain 'dummy' data entered during initial testing of the database.
The DOMUS database exhibited a mixture of dynamic and 'closed' (i.e. static) elements. The static side was represented by the tables which held the data from the 'annual' section of the main questionnaire, and the tables for the supplementary questionnaires. The tables in the final dataset relating to the Registration Scheme also appear to have been 'closed', in that data from past meetings was preserved in the system and not overwritten. The other tables in the DOMUS system were dynamic in that they were updated whenever new information was received.
Constraints on the reliability of the data: The MGC acknowledged that many museums (especially those in the local authority and university sectors) had difficulty supplying accurate financial information because their accounts could not be distinguished from those of their parent organisation. Museums operating as part of a group could usually give individual figures for visitors; however, according to the MGC, it was possible that some returns might include combined visitor figures. Museums operating as part of a group could usually give individual figures for visitors; however, according to the MGC, it was possible that some returns might include combined visitor figures.
The report of the UK Museums Retrospective Statistics Project contained a critique of DOMUS, which highlighted the following limitations:
- DOMUS always focussed on museums which had entered the MGC's voluntary Registration Scheme, despite long-standing plans by the MGC to extend it to museums outside the scheme. Consequently DOMUS misrepresented the total number of museums in the UK.
- Museums appear to have frequently misunderstood questions or were incapable of answering certain questions.
- There was a lack of proper data verification before the data was input.
- The data was always out of date, because of when it was collected and the time taken to process the data.
- Many AMCs questioned the accuracy of DOMUS for regional purposes, and tended to use their own data instead of DOMUS.
Validation performed after transfer: Details of the content and transformation validation checks performed by NDAD staff on each DOMUS dataset are contained in the catalogues of individual datasets.