It is one of the chief glories of the Linkage Editor and now the Binder that it can accept and reprocess its outputs, generating new outputs (in whole or in part) from them. The scheme Edward Jaffe outlined is thus not just feasible, it is the only available one; but since Le bon Dieu est dan le détail, the devil, who keeps Him in sight, will be found there too: o NCAL needs to be specified for the smaller output objects, and o greater care needs to be taken in specifying the content of these output objects. It is possible to know reliably that a load module or program object the elements of which were written entirely in assembly language was constructed only of object modules A, B, and C; but if instead all or part of it was written in some statement-level procedural language--C, COBOL, FORTRAN, or PL/I say--it is all but certain that it contains not just A, B, C, but a number of library routines--call them e,f,g,h--pulled into the load module or program object by the Binder in response to compiler-generated directives. If this is the case every one of the broken out components of the original load module or program object will contain its own copies of e,f,g,h. This is ugly and inefficient; worse, it is a maintenance nightmare. Although they have gone unmentioned, this project would also seem to provide an opportunity to replace at best obsolescent load modules with program objects.
John Gilmore Ashland, MA 01721-1817 USA ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
