1. Introduction

The intended audience of this article includes:

• anyone who wants to audit the steps that were taken in this migration project
• anyone in the ILO STAT section who was used to the old MVS system and is trying to get their work done on the new Unix system
• anyone trying to implement a similar migration, perhaps in preparation for the Year 2000

I worked on mainframes a long time ago, but this was my first effort at migrating programs and applications from the MVS environment to a Unix environment. The first thing that one must understand is the vastly different philosophies upon which the two operating systems were designed. This difference had a significant influence on the task of migrating job scripts, but little effect on the task of migrating SAS application-specific material. To help the reader who doesn't comprehend one or the other of these systems, here are the two philosophies in a nutshell:

1.1 Unix philosophy

Unix was created as a research operating system in a place where getting more hardware was rarely a problem (AT&T Bell Labs). It was assumed that the users knew what they were doing and that they a had a right to use whatever resources were available. For this reason, space quota systems are minimalist and one is encouraged to start as many processes as one needs if it will get your work done faster or in a more elegant manner.

1.2 MVS philosophy

MVS was created during a time when disk space, tapes, and memory were extremely expensive resources. For this reason, the vast majority of JCL job scripting statements are allocation statements of one type or another (usually DD statements). The user expends a major effort filling in on-line requisition forms to tell the operating system (often in very great detail) how much disk or tape space one wants (or will want) and what format that space will take.