Dataset details: CRDA/71/DS/8

The dataset is a database record of conservation work carried out for the years 1995-2003. The database design allowed for greater flexibility in data entry than the 'flat' MS Excel spreadsheets previously used, and enabled such things as automated calculations for dates and time spent on work, and offered various reporting functions. Other than this, the content of the dataset is broadly the same as the previous datasets in this series.

The Conservation department treated every document as a 'piece', that is a discrete physical object requiring work. This was broadly true even when several pieces belonged to a single record with the same reference. The database worked on this same basis, and assigned a unique reference - the 'Conservation_Ref' field - to each piece. This key field occurs in all the tables which describe any repair and conservation work done. The workflow was further managed by breaking down the work into one of 15 'Aspects', to classify the application used.

The principal data table is 'PIECE', which contains most of the contextual detail on each conserved piece, where it came from, when it was received, etc. The table 'Applications U' identifies the application(s) used in conservation, and contains a free text description of work done. Further descriptive text on work done can be found in the table 'Condition'. The table 'Historical description' contains a descriptive text of the historical context or significance of the piece. The table 'Processes U' details the technical processes used in conservation and repair, and the amount of time spent on the job.

Occasionally documents were repaired especially to be loaned to exhibitions, either in The Public Record Office / TNA or elsewhere. The tables 'Exhibitions' and 'Exhibition Documents' were used in this capacity. There are also tables used for the internal storage and retrieval of related registered files, and of electronic documents; the pathnames listed in the tables 'Images' and 'Exhibition Documents' refer to the fileplan of the internal electronic document management system. The 'Image' table also presumably relates to an internal resource or collection, although there are no images stored within this database or transferred with the dataset. The data referred to above are unique to this dataset, and will not be found in the previous datasets in this series.

The table NADFAS describes work on a single series of Admiralty records, reference ADM 36 (Admiralty: Ships' Musters Series I 1688-1808), and the table may have recorded a conservation project using workers from the National Association of Decorative and Fine Arts Societies.

The remaining tables in the dataset are mostly lookup tables. See the individual scope notes with each table for additional detail.

This dataset was transferred to NDAD as a Microsoft Access 2000 database (original file name oldConservationV21.mdb). The tables were exported by NDAD to comma-delimited text (CSV) files using a Visual Basic procedure created by NDAD. An additional four external tables, linked from the Access database, were later transferred by TNA in CSV format.

Included in the Access database were two empty tables (SUBSTANCES_A and SUBSTANCES_U); these tables were omitted from the NDAD dataset as rendered, but are included in the preservation copy. Also included in the database were several reports, queries, and forms.

The external tables, transferred in CSV format, included field names in the first record of the files; these were removed by NDAD. The records of two of the tables (Extract_Register_main and NADFAS), were truncated after the last non-empty field; each record was padded with commas to provide the correct number of fields per record.

All DOS record delimiters were replaced with the Unix equivalent. MS Access adds a superfluous timestamp (00:00:00) to date fields during the export process; these characters were removed from all date fields except where a field contained records with non-null timestamps.

The MS Access database transferred to NDAD contained 33 tables. Of these, four were links to tables in external databases (CO 700, Extract lettercode, Extract Register main, and NADFAS), and two tables were empty (SUBSTANCES_A and SUBSTANCES_U); the two empty tables have been omitted from the NDAD dataset as rendered. Details of all tables can be found in the Access Documenter output file in the Dataset Documentation Catalogue, reference CRDA/71/DD/2/1.

All field labels and descriptions were derived from the field labels in the Access database, and supplemented with information from the dataset documentation. Any embedded spaces in field names were replaced with underscores.

A relationships diagram, extracted from the Access database and showing links between key fields, can be found in the Dataset Documentation Catalogue, reference CRDA/71/DD/2/14. The relationships defined in this diagram have been recreated without modification in the NDAD table relationships, with some exceptions. The link between PIECE.Return_Dept and PRO_DEPTS.Department was omitted because PIECE.Origin_Dept is also linked to PRO_DEPTS.Department and currently the NDAD software cannot accommodate dual links between two tables in the same dataset. Additional links between fields in the PIECE table and two of the external tables (External_Register_main and NADFAS) have been deduced by NDAD. The link between the Exhibits and tblExhibitions tables was defined incorrectly in the relationship diagram as 1:1; this has been changed to m:1 (Exhibits:tblExhibitions).

The dataset comprises the following table(s):

The database was principally used internally by Conservation Department staff for recording their own conservation actions, although evidence shows that the database connected with the wider information environment in The National Archives (TNA). As noted above in Scope and content, certain exhibition tables either pointed or linked automatically to internal paper and electronic file collections. In addition, four of the tables in this database linked to other TNA databases. The tables CO 700, Extract Lettercode and Extract Register Main linked to a database called Extracts Register. The table NADFAS linked to a database called ADM-36, which may have contained more detail on this series of Admiralty records. NDAD do not currently hold copies of these two other 'external' databases.

The contents of the dataset were checked to ensure that the data matched the datatypes specified in the field descriptions of the original system definition; no discrepancies were found.

In the EXHIBITS table, the field Reg_File_Numb has been defined as a key field in Access, but no other tables have an appropriate field for cross-linking. In the Access Relationships diagram (Dataset Documentation Catalogue, reference CRDA/71/DD/2/14), the Con_Depts table is linked to the PERSONS table, but none of the fields in these tables are appropriate for cross-linking; it is possible that the field PERSONS.Obsolete_field field was originally a key field.

The entry of identifying data in the PIECE table appears to be inconsistent, with some reference and class data included in the Title field.

All records in the PRO_DEPTS table have an empty Dept_Description field.

In the external table NADFAS, the Date_conserved field in four records has dates with 2-digit years (in contrast to the 4-digit years used in all other fields). In the external table Extract_Register_Main, the Date_of_extraction field is defined in month-day-year order, in contrast to all other date fields in this dataset, which are defined in day-month-year order.

The Exhibitions table contains several duplicated Exhib_Title fields, often with minor variations in spelling or format.

It is likely that some tables were used for internal purposes, including testing (eg tblAdditionalTreatment).

Following the export of the data from the MS Access database, the number of records and fields in the exported files were compared with those in the Access tables, and no inconsistencies were found.

A number of random queries were made on the original data file in MS Access 2000. After exporting the tables and transferring them to the NDAD system, the same queries were made using the NDAD query pages. No discrepancies were found.