Abundantia Verborum

4. Virtual corpora and corpus linguistics

4.4 Summary

We conclude the chapter with an enumeration of the most important features of Abundantia Verborum with respect to searching corpora.

• Abundantia Verborum requires corpora to be collections of plain texts files, possibly enriched with markup. The orientation towards this format is motivated by common practice and by the argument that all use of proprietary formats will eventually pose serious limits to the accessibility of data.
• The virtual corpus mechanism, and the embedded virtual corpus view mechanism, support the recognition of several markup styles, including TEI-compliant SGML. The choice to support many different styles was motivated by the heterogeneity of currently available data.
• The corpus preparation tool was constructed to reduce the need for complication in the design and time-intensive computation in the use of the virtual corpus mechanism. Moreover, it is a useful general purpose tool.
• The query language of the program is based on regular expressions and word-based searches. These components were chosen for they are widely known and because in combination with the virtual corpus mechanism they form a solid basis for, respectively, general and linguistic queries. Due to its typed nature, the query language can easily be expanded.
• Abundantia Verborum corpus exploration uses neither indices nor data encryption. This is because the typical use of Abundantia Verborum is to linguistically explore locally available corpora. Locally available linguistic corpora typically are fast changing and freely accessible. Moreover, linguistic exploration typically involves sophisticated ad hoc questions rather than high-frequent stereotypical behaviour. This current orientation, however, is not a fundamental limitation. Both the use of indices and the use of data encryption/decryption could be built in.