This appendix describes the current IRIS InSight book-building architecture and process.
Figure A-1 shows the general process flow for building the textual content of an IRIS InSight book. Following the figure is a more detailed text description of the process.
Authors compose a book using FrameMaker, then “compile” the book into an InSight viewable book:
The FrameMaker files (.doc) are converted to Maker Interchange Format (.mif) using FrameMaker's fmbatch utility. A special file, conditional.doc, is used during this process to force the proper FrameMaker conditionals to be applied.
The next phase is to translate the MIF files (.mif) into SGML files (.sgm) using a MIF to SGML translator such as mif2sgml, a translator from SGI. After each file is translated, the .sgm file is scanned to see whether it contains any CGM figure references. CGM figures are generated from FrameMaker line art. If the .sgm file contains CGM figure references then it is run through Inso's miftocgm utility and every anchored frame is converted to a CGM image. The file is scanned again to convert the CGM images to GIF images for use by IRIS InSight.
Once all the SGML files have been generated for the book, they're concatenated and the indexgen utility is run over them to create a back-of-the-book index. This index is then concatenated to the rest of the SGML files, and a document bounding tag pair, such as <SGIDOCBK> </SGIDOCBK>, is wrapped around the result to create the master SGML file that is fed into Inso's mksgidoc or mksgidocbk command to create an IRIS InSight viewable book.
Note: Supported DTDs include SGIDOCBK and SGIDOC. Arbitrary DTDs are supported to a limited degree; for example, CDOCBK will work with the mkbook command, but the indexgen utilty is not supported.