A company has one or more large collections of documents that it wishes to make available via the web and also publish on DVD for archival and detached use. These documents are of varying detail and structure, but all include some forms of metadata. They wish to be able to combine database-like querying in the metadata with full-text searching in the document content.

Their problems are:

The structure of the documents is quite variable, so that a row-and-colum-oriented database such as Sybase or Microsoft SQL Server database would require a large investment in data reformatting and normalization, and the full-text search capabilities of such databases are weak. Licensing and installing desktop versions of those products can be very costly in terms of both fees and support. Full-text oriented databases are typically weak in querying structured fields, such as this company's metadata, and particularly in combining structured searches with full-text ones. The company has not been able to find a solution which performs well both as a large server and as a desktop application. Our product could solve these problems through:

Its flexible record storage model stores only the actual structure and content of each individual document, without reformatting. The software is easy to incorporate into a desktop application, and licensing costs are appropriate for an embedded component. The software has strong numeric and range search capabilities that can be combined seamlessly with full-text querying. The software is optimized for interactive performance on commodity hardware, such as a desktop PC.