I wish to thank all of you who took the time to let me know about the several software and documentation errors. Without your help TSXLIB could very well degenerate into something of a pitiful mess. Please keep up the very fine debugging and proof reading. I also want to apologize to those of you who disliked my experiment with 12 char/in and 8 line/in. You will be pleased that the document is back to the standard 10 char/in and 6 line/in. A major change to the document is the replacement of the old Appendix A with a key word in context (KWIK) index. This should help the user who is not intimately familiar with TSXLIB to find a routine to meet a given requirement. With the major change in the privilege structure of TSX V6.0, many of the older routines were affected. Hence, many changes have been made to both the code and the document. One of you suggested that I standardize the error return information in TSXLIB to be zero for no error or no exception. This would be great, but not very convenient to implement. However, as TSX evolves, so will TSXLIB. A negative (-1) return is becomming more common as the no error or no exception condition. Following the release of TSXLIB V5.1, I updated some of the FORTRAN test code for TSXLIB. It is once again in dire need of revision. The job is not trivial, and I no longer have access to many machines to be able to exercise all of the TSXLIB routines. I would appreciate some help from any of you. The FORTRAN test code is not written as sample applications, but as simple exercises to test the MACRO-11 FORTRAN callable implementations of the TSX EMTs. The test kit just fits on an RX01 floppy. Nick Bourgeois NAB Software Services, Inc. PO Box 20009 Albuquerque, NM 87154 27-Feb-86