BKUNXMNF.RVW 970502 "UNIX for the Mainframer", David Horvath, 1997, 0-13-632837-7, U$49.00 %A David Horvath %C One Lake St., Upper Saddle River, NJ 07458 %D 1997 %E Michael E. Meehan 0005955361@mcimail.com %E Jane Bonnell mjb@prenhall.com %G 0-13-632837-7 %I Prentice Hall %O U$49.00 +1-201-236-7139 fax: +1-201-236-7131 beth_hespe@prenhall.com %P 413 %T "UNIX for the Mainframer" First of all, there is no point in even pretending to be unbiased in this review. I am acquainted with David Horvath, and we corresponded all the time he was writing this book. He even mentions me (and plugs *my* book!) in the acknowledgements. I've reviewed the book twice, once in an earlier draft so that I could make suggestions. Still, it would have been disappointing, having encouraged David, if he had then produced a bad book. He hasn't: this is a work of quality with a good reason for its existence. The intended audience is the experienced and trained (qv. "brainwashed" :-) IBM mainframe systems programmer, and that I am not. Therefore it is hard for me to say with finality that the book is completely suitable. However, from my time spent with IBM systems, I feel that it is. It has aspects of the "look and feel" of IBM documentation, which makes for a good comfort level. The book presents the common tasks for management of a mainframe, and then explains the UNIX means of accomplishing the same job. There is a great deal about UNIX that the book does not cover, but the systems management tasks are dealt with pretty fully. Other aspects of UNIX are not necessarily within the interest or scope of the applications programmer, although the UNIX user can also find much of interest and value within these pages, and, even though mainframers are the intended readers, any determined novice could learn UNIX from this work. The book is a fair size, but management is a sizable task. Horvath does not waste any space introducing concepts that the intended audience should already be familiar with, only the new ways of dealing with them. Because it is aimed at technically experienced people, some detail and internals are also included. One thing struck me about the section on shell scripts: while the example scripts are useful, often complete replacements for mainframe utilities, perhaps simpler and smaller scripts might illustrate some of the functions more clearly. The glossary is an important part of the book. Generally, computing concepts are the same anywhere: only the terminology changes. (For example, the fact that a systems programmer is what everyone else calls a sysadmin.) The current glossary is a good start, with IBM terms referring to the UNIX counterparts. However, I think it both could and should be expanded. Or, perhaps an additional appendix could provide a quick reference for common mainframe functions and the UNIX equivalent. (This is, of course, the body of the book, but a quick reference would be helpful.) I know of only one other UNIX book intended for this target audience, "UNIX for MVS Programmers" (cf. BKUNMVPR.RVW), which doesn't really approach the topic from the mainframe perspective that Horvath uses. For those wanting to go further with UNIX, the bibliography that Horvath provides is first rate. With the ideas behind such buzzwords as "distributed computing", "client/server", and "data warehousing", integration of mainframes into networks is becoming more, rather than less, important. UNIX is the de facto "glue" behind multi-vendor networks, so a "meeting of minds" between the IBM and UNIX worlds is vital. Appendix F, on TCP/IP, is fairly important. In keeping with the idea of UNIX as a link between mainframes and networks, so TCP/IP is important to SNA (System Network Architecture). David Horvath has written an excellent introduction, tutorial, and reference for those mainframe programmers and administrators who need or want to become part of the larger computing and network world. It is solidly structured, technically sound, and more than passably readable. copyright Robert M. Slade, 1996, 1997 BKUNXMNF.RVW 970502