How about this,
Is it ever ok for code in the "View" subsystem to cause a coredump
or problems in the "Model" subsystem. Since this architecture is
compiled into one program it coexists in one process/task space.
Is it ever "OK"? Of course not? What kind of question is that? I guess it is OK if that was the intent of the program, like a virus of some kind.
What about any client-server relationship?
What about it? Is it OK for a client to crash the server? If that's the question, then obviously not. Not getting your point yet...
What about any software(Word = client) causing an error on the
operating system( Dos = server).
Of course this isn't OK! I guess I still don't understand what you are asking.
Are you suggesting that a program (Word) should not be able to cause a fault in the OS (DOS)? Perhaps this is what you mean. Of course, this should not be able to happen. Any program could make a system level call with bad parameters. The OS should handle this gracefully, which DOS and Windows 3.1 didn't do very well. But any code running in the same process space with other code is unprotected against "harmful" calls - I don't think you can stop it. DOS and Windows 3.1 allowed this, which is why it crashed a lot. A bunch of other older OS'es allowed this also, which made them unstable.
So, should Word crash DOS? Of course it shouldn't be able to, but it does. Can Word crash Windows 3.1? Of course. Can Word crash other programs running on Dos/Windows 3.1? Sure.
Can Word crash Windows 95? It can, but it's a lot harder. Can Word crash NT? I doubt it, unless there was a bug in the kernel, which is possible. Can Word crash another program running on NT? It could, but not by messing with something in it's process space. |
