Talk:Operating System Design/Case studies

Linux already has reliable journalling filesystems, several graphical interfaces, massive amounts of written documentation, etc. etc. Why would anybody want to use a DOS type of OS?? r3m0t (cont) (talk) 07:33, 17 Oct 2004 (UTC)

I agree - anyone who would consider basing an OS on DOS (even back in the 80's and early 90's when Microsoft was doing this with Windows) needs to learn more about the topic. DOS doesnt allow for much expansion (more filesystems), is single tasking (that means no running mozilla at the same time as notepad, and even GUIs themselves are almost unusable in this state) unless this is done manually by almost replacing all of DOS. The command line is defianately an advantage, however anyone familiar with a DOS command line would notice its almost impossible to work in by today's standards. Any Unix implementation has all the desirable features of DOS, only they actually are desirable for Unix - the command line is useable. Also, DOS offers no security whatsoever, any application can delete anything and stop the system from functioning and prevent recovery of any data. Unix is very secure (much moreso than Windows, this is a topic of debate, however anyone who has experience with both systems rather than just believing what microsoft says would defianately agree). Unix's only real shortcoming is the GUI - X11 is seriously slow and lacking in features desirable by the common user. While X is certainly more than sufficient for web browsing, word processing and most other things, benchmarks show that it is slower than Windows, and the fact alone that many applications look and behave differently is a serious problem. No GUI for DOS even competes with X however. I have deleted the FreeDOS section, no offence to the author, but it is misinforming.

How is this a wikibook topic? - anonymous

I don't think this belongs. It acts as if a new os is going to be created in the book. This is not the case according to the preface.

I have to agree with the above. This sounds like the author is going to document writing their own OS from scratch. Much better for this type of textbook (imho) would be to discuss the issues, and various approaches to resolving them. Using real life examples (particularly open source ones like Linux or the BSDs) to point the user to implementations that match specific approaches. For example most OSes have some kind of memory management, some kind of task scheduler, and some kind of file system. The actual implementations can vary massively, and there are many ways of dealing with the associated problems. In the spirit of this I'll try modifying this page to state more explicitly what could make this a good text book.