Three boxed “golden references”--a quote fromEinstein and descriptions of the concepts of “text” and“tool”--add a humorous touch to this book’sintroductory material. Part 1, “User’s Guide,” is a tutorialthat describes the user interface in general and presents six commandmodules--edit, system, compiler, backup, net, and miscellaneous.Part 2, “Reference,” is a manual of 11 modules of the Oberonsystem proper, developed in a bottom-up fashion. Reiser provides both analphabetical summary and a detailed description, in logical order, ofthe procedures and data structures of each module. Part 3,“Programming Guide,” puts it all together, describing rules,tips, and tricks for working with the entities that play a prominentrole in Oberon. Appendix A shows how to write another system module thatimplements a new data structure. The book contains only two or threeexamples, which are described during stepwise refinements; this approachdoes not fit Part 2, with its bottom-up hierarchy. Some of the modulespresented in the text seem to be part of a standard library. The lastpart of this book contains an ASCII table and a bibliography, not citedin the text, listing 11 reports and “not yet published”books and six articles about Oberon. Separate references to many of thesame articles appear in footnotes in the preface and chapter 1. Aglossary follows, which should be consulted if a search in the indexfails. The only exercise is dropped into the text on page 252; the bookis not intended for students.

Oberon does not differ substantially from other programminglanguages, so prior knowledge of Oberon is not a prerequisite for thisbook; Pountain’s article provides sufficient background [1]. The bookcontains both repetitions and forward references. None of the proceduresfrom Part 3 are listed in the index; some are used before they aredescribed, and they are hard to find. The length of the book is justright, although many questions arise that are not answered. The physicalform highlights the material, with only a few flaws, the major one beingthe lack of picture captions and numbers. Only a half dozen typosappear, although the author has made over 40 careless errors. Almost twodozen logical errors of varying severity raise the suspicion that Oberonhas never run.

According to the author, this book is intended as a reference forusers of the Oberon system. Why should I sit in front of a Ceresterminal with a three-key mouse in my hand? Oberon has no spreadsheet,no database, no AutoCAD, and no Pascal, and it never will have them. Theaim is to learn programming--the author confesses that Oberon isuseless for non-programmers. It could be the right vehicle. I remembertrying to teach programming using extracts from Dijkstra’s papers buthaving only Algol available. Although Oberon is not overcomplicated likeModula-2, it will not supersede Pascal, even in schools. It is a pureactor programming language, without control flow, reduction, or otherconstructs of object-oriented languages based on the class concept. Ageneral Oberon rule is that data structures are extended retaining alltheir fields, whereas methods (procedures) are not inherited and must beexplicitly stated--Oberon is a poor man’s object-orientedprogramming system. Oberon modules should be less forgetful and shouldsave their states on mass memory, as databases do, instead of learningeverything anew at each invocation. Oberon suffers overconsequence andoversimplification: new semantics is badly needed for active objectcreation and message passing, and I missed my favorite syntax,“for I while B do S.” A combination of Oberon with anapplicative language might be best. Oberon is yet another Ada, born inan ivory tower as a would-be new and elegant UNIX.

The standard we use to judge Niklaus Wirth and his co-workers is ofcourse much higher than that applied to ordinary writers. Two reproachesstated over a decade ago still hold. Oberon is intolerant, in that itrequires absolutely correct input data to function, and impersonal, inthat a broad general knowledge of everything in the language is requiredbefore it is possible for me to do what I want. Oberon fits into 200kilobytes, which is great, but the system’s functionality is notcomparable to that of the one-megabyte Windows system. It fallssomewhere between DOS (80 kbytes) and the Brief editor(120–500 kbytes). Modula’slimitations remain in Oberon. For example, no true multitasking ispossible--the concept of an interrupt, which is crucial to modernoperating systems, is never mentioned in this book. The author does notaddress batch files or macroparameter substitutions. Oberon iserror-prone. No clear boundary exists between Oberon and itsapplications; it is impossible to teach students using such anenvironment. The file system is very primitive, resembling that of IBMmainframes. I cannot imagine the process of debugging a system with somany up-calls, and in which several handlers are called recursively. Thewhole concept of showing currently loaded modules (p. 56)should be abandoned soon after disposing of the command line--RAMand CPU should be private resources of the operating system with whichit executes modules from mass storage, and these modules should all beaccessible, as if they were located in virtual memory. Oberon is sosmall because it lacks all the features an operating system must have.These omissions are why the author never compares Oberon to popularoperating systems--it is a lost cause.