"Observations on the Development of an Operating System" page 33:Hugh C. Lauer, Xerox Corporation, Palo Alto, California In SOSP '81 Proceedings of the eighth ACM symposium on Operating systems principles, ACM SIGOPS Operating Systems Review, Volume 15 Issue 5, December 1981, pages 30-3. ISBN:0-89791-062-1,doi: 10.1145/800216.806588, PDF . . . I found it useful to enumerate some of the other operating systems I have known, either from direct contact or from study of the literature or from contact with others. These systems seem to fall into five categories. which I shall first enumerate and then describe. 1. The Alto system. UNIX. 2. IBM's OS/360. MLUTICS, Pilot, etc. 3. MTS (the Michigan Terminal System). TE~EX. CP-67 4. CAL-TSS. Project SUE. HYDRA. etc. 5. DOS/360, RS-11, etc. These categories are the result of my personal observations. nor of a systematic study, and hence many systems are not listed because I don't know enough about them to classify them. The ordering of the categories is not significant. An important pan of the classification is the maturity or success of a system-i.e .. acceptance by its clients as a useful. economic tool for helping to get work done, for · supporting applications, or for fulfilling other goals. A characteristic of a successful system is that it is accepted by a nontrivial community of users outside its developing organization as a matter of choice and that this community contributes. directly or indirectly, to its further development and growth. . . . . . . Systems of the third kind. These systems borrow much of their supporting software from an existmg system but represent a fundamental change in the way of life. The Michigan Terminal System, for example, provides a paging, terminal-oriented. time· sharing system especially suited for university use on the IBM 360/67 and IBM 370 systems. Most of its compilers, run-time support, subroutine libraries, program development tools. etc., were taken and convened directly from OS/360, but its operating system kernel is dramatically different from OS/360 and it supports new applications that OS/360 never could. (Of course, there are also many OS/360 applications that MTS cannot support.) The obvious motivation for building a system of this kind is to avoid the time and expense of designing. implementing, and maintaining all new supporting software for the operating system when it is desired only to implement an operating system kernel and some basic functions. . . . [I think the above paragraph overstates how much of MTS was "borrowed". It is certainly true that many of the programming languages compilers, a number of other application programs, and some subroutine libraries in MTS were borrowed. It is not the case that most of the MTS run-time support and program development tools were borrowed. The MTS Editor, the Symbolic Debugging System (SDS), and the MTS file system are a few examples of components developed for MTS. And not all of the components that were borrowed came from OS/360, although many did. -Jeff] |
Discussions > Published information >