Integrating TEX into a Document Imaging System

Texas Life Insurance Company, Waco, TX

Abstract: Modern computer hardware and software has made possible the construction of ``document-imaging'' systems. These systems maintain large repositories of documents in electronic form. In production environments of many large companies and in particular the life insurance industry, a significant percentage of printed documents are produced electronically in an automated fashion usually by merging variable data into an existing document with some fixed structure. Storing scanned images of these electronically produced print documents wastes time, computing resources, and disk storage space. It is useful to address the problem of document storage along with the related problem of electronically formatting and producing printed documents. Then the choice of document formatter can be made such that the formatting engine used to produce printed documents may be reused to display those same documents in a document imaging environment.

This paper examines how TEX has been used as an important component in building a document production and imaging system at Texas Life Insurance Company. TEX's macro facilities, conditional typesetting, text-based source files, a robust page formatting mechanism, and pre-compiled format files allow it to play a central role in the system. TEXmerge, a C-language API, was developed to allow variable data to be merged, under program control, with static TEX source documents containing special merge tags to produce a final output document. This API is used to prepare policy contracts, produce automated client correspondence, as well as in interactive document preparation and in application-specific document production. Documents produced via the TEXmerge API are filed in the imaging system using a minimalist approach. TEX form files are stored once and separately from all document instances. Variable data along with a pointer to its associated TEX form file is all that comprise a stored document instance. When the document is displayed a ``just-in-time'' compile technique is used to reconstruct the document's .dvi file which is converted to PostScript for display purposes.

TEX has additional attributes that make it an excellent choice as document formatting engine. The ability to convert raster bitmaps to TEX fonts allow complex letter-head/footer macros to be developed and easily used in a fashion that lends itself to effective revision management. Incorporation of a Code 2-of-5 scalable barcode font has enabled printed forms and documents that are returned to the company from external individuals to be recognized by the document imaging scanner and automatically filed in the imaging system.