Each term has its own record in the thesaurus. This holds all hierarchical and semantic relationships with other terms. You can input the following data for a preferred term:
On starting Adlib, you have to select the database in which you want to search, input or modify records. You to input term records directly into the thesaurus, and also by ‘forcing’ in terms when you are entering data in other linked databases. In the Library catalogue, for instance, the Material type field is linked to the thesaurus. This not only means that every entry in to the Material type field is automatically validated against the thesaurus, but that you can also use the field to add new terms to the thesaurus without having to close the catalogue record. This functionality works in the same way for all fields validated against the thesaurus in all Adlib databases – and there are many of those.
The validation process described above takes place when leaving the field, though can also be activated by the user prior to inputting. A window can be opened to browse for terms that already exists in the domain for this field and indeed all other domains. This allows you to check whether an appropriate term is already available in the thesaurus; you can then select that term or add a new term. If you decide to do the latter, this can be done in the same already open window. Validation also occurs automatically during searches: non-preferred terms will be automatically replaced by their preferred terms, and generic searches can be used to refine your search query.
All necessary reciprocal records are created when thesaurus records with hierarchical relationships are saved. This is needed to include all the relationships in an index. Every broader, narrower, related, equivalent and non-preferred term is thus stored its own record, along with the link to reciprocal records. Thankfully, Adlib does this automatically.
It is possible to use Adlib’s hierarchy browser function to provide a detailed display of records with a hierarchical structure. This appears on-screen as a tree structure showing all linked records. You can also interactively change the structure by simply dragging and dropping ‘leaves’ or branches of the tree.
There are reasons why you may want to use more than one thesaurus in your Adlib applications. For example, you may wish to use specialised thesauri for cataloguing different collection types, or in Adlib Museum, for example, it may be appropriate to use different thesauri for specific fields within an object description. Plus, there are excellent published terminologies for specific professional areas which are ready to use, such as the Getty AAT. It obviously saves you a lot of time and effort if you are able to employ published thesauri alongside your own in-house thesaurus. Adlib applications make it possible to use external thesauri in searching, to validate specific fields on data entry or to modify your own thesaurus. During use, standard thesauri can be on a CD-ROM, on the intranet or simply somewhere on your hard drive. You therefore do not need to perform a data conversion to get an update, and only have to use a new CD or validate against another thesaurus elsewhere on the system. The only requirement for the thesaurus is that it must be available as an indexed Adlib database.
If your Adlib application runs on an MS-SQL or Oracle database, then the settings of the Term field in the Adlib thesaurus can be configured so that one or more translations can be associated with the term. Translations will be stored together in the same field. If you enter translations for each term in this manner, you can make your thesaurus completely language-independent. The Data language option in the Adlib menu bar will only be active if a field (in the Adlib data dictionary) has been made multi-lingual. The translation that you will see displayed in the field is determined by the language selected from the menu. If a translation has not yet been entered for a given language, the input box will not remain empty: in this case, the term from the first language entered (termed the ‘invariant language’) will be displayed; this assists you to do the translation more effectively. You can manually change the invariant language, of course.