Customer Reviews

FreeDOS Kernel; An MS-DOS Emulator for Platform Independence and Embedded Systems Development

5 star:		(2)
4 star:		(0)
3 star:		(0)
2 star:		(2)
1 star:		(0)

 

 

The author really made the text flow by providing a high level view and then peeling back each layer to reveal the working of the kernel. In addition, he judiciously intermixed, text (historical, theoretical, practical, opinions) with diagrams, and code segments, which made the book easy to read, and concepts easy to follow.

The code was written in a clean and consistant style, amply annotated with comments which explain what the code does, as opposed to just paraphrasing it. A student or inexperienced "C" coder will see the practical uses of portability techniques, ifdef debugs, type defs, etc., which are frequently ignored in academic works.

Lastly, I'm glad the author resisted the current trend of dumping everything but the kitchen sink into a text just to see how much shelf space he could take up (most professionals are to busy for that nonsense). This is a nuts-and-bolts approach; what are the concepts, why is done, how is done, here's the code (and its all on a companion disk).

If your looking to learn about DOS, kernel implementations, or some good "C" code examples/idioms, this book is highly recommended.

This is my second pass on this book.And each pass gives more pleasure.Excellently designed while obeying ms-dos compatability.Excellent explanations.Highly commented programs.I have got the overall picture.But the f-node related functions are somehow harder to understand and needs more explanation.I have made pass 1 on Podanoffsky's dissecting dos.This book is much better than the latter.If there had been a table showing the functions and the files they are in it would have been better. I am looking for a similar book on a multiprogramming OS and as easy to grasp as this one.

I really regret purchasing this book. Here are my main complaints:

1.) The book was written in 1995, half a decade after ANSI C90 was established, and yet I see K&R style constructs. This leads to problems when actually compiling on modern compilers that have options to generate 16-bit code. Speaking of compilers, this leads to my second complaint.
2.) The tools that the author suggests to use (Microsoft C or Borland C) are completely dead. When I asked my professor for help, he literally laughed at me and told me to give up. Well, I didn't listen to him, and used Borland C++/Open Watcom to build the project, but that still failed. The compiler gave me a crapload of errors. When I checked the lines that caused the errors, I saw some awkward structure definitions that made no sense to me. Now, I understand C90/99, and I am pretty sure that things like

struct blah
{
struct huh
int what;
int dunno;
};

(There are supposed to be tabs before "int what" and "int dunno." Code is supposed to look like declaring a struct inside another struct, but does it in an erroneous way that doesn't compile.)
are not supposed to compile anyway.

I bought this book because I wanted to learn about OS development by studying a simplistic OS. If you are like me, I recommend purchasing Tanenbaum's "Operating Systems Design and Implementation." It is modern and building the project is way simpler, and there are lots of FAQ's about building the project. The content of this book is good. The explanations are clear, and the source code in the book is commented and explained thoroughly. The author should consider writing a second edition for this book to make the source code up to date, as well as completely rewriting the section on what software to use so that the reader can build the kernel.

This book was ok, but you need borland C++ unless you want to do some major porting... The back said that you only need borland OR microsoft c, but that is a lie. The only reasion I gave it a 4 was becouse of this. You could tell that a program wrote it and not an author. If you want a book on os programing get it