Customer Reviews

Windows Assembly Language and Systems Programming: 16- and 32-Bit Low-Level Programming for the PC and Windows

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

 

 

This book is NOT for everybody. But if you need to extract maximum performance from Windows 95/98, it is indispensible eg, Games, SCSI i/face, real-time I/O, etc. In conjunction with Walter Oney's "System's Programming for Windows 95" it is unbeatable. The appendix describing the Microsoft DPMI extensions is worth the price. It is the only work that addresses: TSRs with Windows, Call gate thunking and shared VM memory areas. (If you don't understand what these mean, this book is not for you; if you do then you will appreciate their importance and indispensibility for high performance apps, particularly I/O oriented ones.) I found Kauler's description of Assembly implementation of OOP innards most illuminating. The 1st 2 book chapters could be culled substantially, to a summary of x86 architecture & Assembly, and the space better used for subsequent chapters where the descriptions are somewhat thin. Since the number of books on low level Assembly hacking into Windows is just 1, and Ring 0 Assembly is the only way to handle multiple CDRs, RAID array, multi-DAT or other high throughput I/O apps -- this book stands alone, regardless of its warts. Oney's book lays out VxDs, Kauler's fills in the key gaps for direct DPMI calls, fast thunking, VM sharing and working through DOS REAL Real Mode (not V86). Kauler's irreverent style is somewhat flippant for this serious a topic.

I have to admit I learned alot from this book. I think it could have been written a little "smoother". I agree with one reader when he says the first two chapters could have been put to better use. If you use Microsofts Masm assembler, then you may be somewhat lost. I know of at least two instances where he refers you to another section to configure your programs in Masm, but, there is no reference to Masm in these sections. The appendices are excellent. I must admit this is the only book of its kind I have been able to find but in most areas he just skims the surface. He doesn't even tell you the proper way to retrieve a key stroke using Windows routines. In all honesty, "Its better than Nothing"..! I know more now than I did...

Kauler has produced a book, published in 1997, that was last relevant in about 1994. Although he promises 32-bit Windows assembly programming (meaning Win9x), virtually everything is geared towards DOS or Win3.x. The included samples are missing pieces and won't assembly or link. I wish I could sent it back.

Assembly language is important - you really need it when
a) your compiler is producing instructions different to those you intended with your higher-level language
b) debugging system-level code
c) disassembling some binary file that has no source code
d) presented with a crash address alone.
e) You might even need to write a little assembly when what you are doing simply cannot be written in a higher level language.
There are some excellent books to help learn assembly, and chapters in books and articles by the likes of Pietrek and Robbins.

Kauler is different. He believes you should write your Windows GUI programs straight in assembler - dialogs, menus, windows and all. He thinks this is a good way to write Windows programs. Anybody who wants to follow his advice does not need this book, he needs to learn about modern tools. Modern compilers are really good, and it is ludicrous to suggest replacing their work with yours.

Having undermined the central premise of this book, it is worth commenting on the content. Firstly, it is very heavy going, and somehow clunky. I don't know if it is the font, page layout, or simply trying to cover too much too quickly, but I had to read each paragraph a few times to understand what was being said here. The book has clearly been rehashed from previous Kauler literature, even leaving in the same screen shots from 1992! Several chapters have rambling overviews of Windows architecture or the boot up process, and quite frankly, other books cover this far better. What this stuff has to do with assembly is not explained.

He also seems to be stuck in a time warp, by writing most of his code in 16-bit assembly. There was a time, when I was still young, when you had no choice - 32-bit Windows was still a pipedream. But already for several years this has been obsolete, and the only need to know it is when poking around in the 16-bit underworld of 95, but heaven forbid actually programming in it!

I have to admit that there are some gems here - Kauler has prised open some of the cracks in the Win95 OS, and revealed some amazing tricks. Among them are using DOS interrupts to gain access to low-level services and using CallGates to run Ring0 code from Ring3. However, even this gem is written cock-eyed, with the main program in 16-bit code, and the CallGate callback in 32-bits! Does he want nobody to understand him?

And all this to expose the Win9x OS! It still exists, but is becoming more and more obsolete. XP Home is already upon us, and I doubt any home PCs will be sold with the 9x family installed ever again. The kind of people who want to dig into the OS migrated to NT years ago, while Kauler is still stuck in the 9x days, blinking in denial as he emerges from a 16-bit slumber.

Not for me such stuff. Avoid this book, unless you are a real 9x underworld junkie, think in assembly, have more that just a dash of Windows 3.1 nostalgia, and yearn for the good old days of 16-bit programming.

This is interesting stuff. However it's old and a lot of it has a kind of rambling, meandering quality like the author's just mentioning some stuff off the top of his head. A lot of it is geared toward Win3.1 and 16bit vs 32bit, but old info is maybe better than no info, since this is one of the only books out there on this subject.

His first edition on 16-bit was very well written and covered nearly every subject on writing Windows programming in assembler. His second edition is just a rehash of the first edition with a few 32-bit tidbits tossed in among code that will not compile in either TASM or MASM.

Granted, it's the only book available on this subject, so you might was well buy it, but clearly the author didn't spend much time on this edition.

Being one of the very few books on this topic, this book is essential for anyone who's interested. Although there is a chapter about the basics of assembly language, I think the reader must have some basic assembly skills before starting with this book. I personally knew DOS assembly and wanted to learn Windows assembly. This book formed a GREAT introduction.

The only drawbacks I find are:

- the comparison Win16 <-> Win32 is made, I'd rather see the comparison Win9x <-> Winnt;

- the author doesn't always explain what he promised to explain. E.g. A chapter about Ring0 Code is about an example of *getting* ring0 from ring3;

- the source code is messy, and doesn't always compile.

But still, it's a great book.

Good book for the programmer who uses assembly and would like to cross over from DOS to Windows, however, the constant confusion between MASM and TASM with insufficient clarification and then writing a program using Borland's C++ and telling the user to use Microsoft's NMAKE.EXE instead of Borland's MAKE.EXE because Borland's is different without showing why is a little absurb. I have had TASM 4.0 for several years and enjoyed writing DOS assembly programs, but I do not wish to go out and buy MASM just to satisfy the author. In addition, having to buy anything else to supplement a book when you already have good programs should not be necessary.

it is. The main problem here is that its author, Barry Kauler, while a very competent engineer (and a college professor, I seem to remember?), is an inept writer. The book needs to be re-written to be easily understandable (if it were, it'd be worth reading even despite its age).

The other major flaw of this book is that this 1997 edition is a _perfunctory_ update of an even older (circa 1993?) book that was fully Win 3.1. "Perfunctory" is the key: it's almost all the old, 3.1 stuff. The book does have some neat 32-bit tricks in it, but the effort needed to extract them is too high. Also, the 32-bit element in this update (a perfunctory one, as I said) is all Windows-95 specific and will not work for the NT-based family of Windows, which is the only kind around these days.

The only thing you can learn from this disorganised and unclear book is some non-obvious uses of the assembly language. If still you want to check it out, make sure you get it for the price of postage. Maybe the seller should pay a half -- for the privilege of having it removed from the premises. It pains me to recall how I bought it in a B&N store when it came out for the full price. I have long since trashed it in one of my once-every-five-years house cleanups (and I never throw books away, 'cause who knows? maybe I'll need it... and I regularly do, but not this one).

I thought this book read like it was a patch to his earlier 3.1 book - not that I've read it so I can't say for sure. As far as I was concerned there were too many references to the 16-bit world. On the plus side, I like his informal 'chatty' writing style.