SEARCHING HELP: HOW IT WORKS

The database

Our database consists of a series of records which represent each musical work and individual recordings of that work. Each record consists of a number of database fields which aim do describe the people (such as composers) and musical properties (such as name, genre and instrumentation) related to each and any recordings we have of each work.

What is searching?

Searching consists of matching 'Search Terms' (the words you enter into the search field) to the words and phrases in various database fields in each record. For instance, we have a field called 'Real Name' which is the title given to a work by its composer. If you type in the search term "symphony", then this word will be compared against all of the 'Real Name' fields. If any of them contain the word "symphony" then that work is considered to be relevant to your search.

Relevance score

For each record we calculate a 'Relevance Score'. The score is proportional to how many matches occur, and is also affected by which particular database fields the match occurs in.

Result set

Once a match has been found and a relevance score calculated, the record is added into the 'Result Set'. Once all the records have been processed, the Result Set is ordered by relevance score (highest first) and then converted into the graphical representation you see as search results.

The 3 phases of searching:

  1. Pre-processing - Search terms are pre-processed before being searched in order to correct common spelling mistakes and to convert alternate formatting to our cataloguing standards.  For instance, "chikofski" is corrected to "Tchaikovsky" and "g#-major" is converted into "G sharp major".

  2. Specific Term Search - Each search term is compared against a special list of terms in our database known as 'Specific Terms'. 'Specific Terms' are words with specific musical significance, such as the genre of a piece of music. For instance, the terms 'String Quartet', would be matched against the specific term of the same name, which is a musical genre. We can therefore identify that you are searching for recordings of string quartets.  When Search Terms match Specific Terms, they score relatively high relevance.

  3. Text Search - Each search term is used in a simple text search of all our database fields. In this phase, the search terms "string" and "quartet" would be matched against work names and words in programme notes and biographies, as well as nicknames, librettists and all the other database fields. When Search Terms match words in a database field, they score relatively low relevance.

Conclusion

It is a combination of these three phases which allows us to prioritise the results in order of their relevance to your search terms. Specific terms are a powerful way of making sure that the musical work you were interested in finding is the first result to be displayed in the result set as it allows us to clearly distinguish works that ARE string quartets from those that just happen to have the words 'string' or  'quartet' somwhere in one of the fields in their database records.