From - Thu Oct 02 10:42:16 1997 Path: news.mitre.org!blanket.mitre.org!bone.think.com!newsswitch.lcs.mit.edu!eecs-usenet-02.mit.edu!nntprelay.mathworks.com!logbridge.uoregon.edu!su-news-hub1.bbnplanet.com!news.bbnplanet.com!zdc!szdc!newsp.zippo.com!sdrn From: shelton@despam.metronet.com Newsgroups: comp.os.ms-windows.programmer.nt.kernel-mode Subject: Re: Which NT Driver book do you recommend ? Date: 1 Oct 1997 09:05:39 -0700 Organization: None Lines: 101 Message-ID: <60tscj$o1@sdrn.zippo.com> References: <342F5826.C57@ibm.net> <60ou95$k48$1@news.real3d.com> <60s41c$q8n$1@news.fta.com> <34321DFA.6604@oneysoft.com> NNTP-Posting-Host: p-103.newsdawg.com In article <34321DFA.6604@oneysoft.com>, Walter says... > >bill a wrote: >> Actually, I heard this book got slammed in the OSR newsletter ? I >> thought there were a couple more books in the works. > >No-one has had the courage to post here any description of the supposed >errors in Art's book. If no-one is willing to be specific, I would >conclude that the rumors are unfounded. Art Baker doesn't deny there are errors, as there surely will be in a new book on a difficult subject. What I find incomprehensible is the lack of an errata page at Prentice Hall or anywhere else. I've seen the book on the best-seller list of at least one net bookstore, and I conclude that they are selling a lot of these books and can afford to support it -- which they are not. Walter's book is well supported on his site: thank you very much! Dick Shelton P.S Here is Art Baker's reply about the subject in this news group: ` Subject: OSR's review of the NT Driver Book From: cydonix@clark.net (Art Baker) Date: 1997/04/07 Message-Id: <5iaut4$mul@clarknet.clark.net> Newsgroups: comp.os.ms-windows.programmer.nt.kernel-mode In the most recent issue of the "NT Insider," Peter Viscarola of OSR reviewed my work, "The Windows NT Device Driver Book." I must say, I was both surprised and confused by Mr.Viscarola's strident (almost hysterical) tone until I remembered that the "NT Insider" is essentially a piece of advertising material for his company. This leads me to wonder how many of his comments were motivated by a simple desire to present his competition in a bad light. It seems rather disingenuous of him to make no mention of the fact that his company and mine are direct competitors in the same market-space. Cydonix Corporation has been offering NT driver training and consulting since shortly after the arrival of Windows NT 3.1. Some of our regular customers have included such companies as Intel, Compaq, Hewlett Packard, and Lucent. I believe a somewhat more professional and objective review would have included this information and pointed out the reviewer's potential conflict of interest. Also, while he continually questions my understanding of basic driver issues, he carefully neglects saying anything about my background. The book itself points out that I've been writing various kinds of device control software for nearly 25 years and teaching driver programming for 12. Mr. Viscarola's references to the "...lack of pragmatic experience on the part of the author," are simply incorrect. Mr. Viscarola seems disappointed by the fact that the book covers only basic NT driver development. However, the book's introduction clearly identifies the work as being a general introduction to the subject. Some of his unhappiness with the book seems to be because it isn't the book that he wanted to read. Clearly, not every book is going to be suitable for every reader. Someone of Mr.Viscarola's level of skill isn't going to find anything new here. Perhaps his disappointment is another contributing factor in the extreme negativity of his review. However, other reviewers, and people who have taken my driver classes, have found great value in the book. These readers have commented that the Windows NT Device Driver Book has given them a clearer understanding rather than, as Mr Viscarola claims, "adding to the obfuscation of the DDK." As to the technical errors in the book, several of them are due to the publisher's production staff not incorporating some of my corrections into the final draft version of the book. The incorrect description of DPCs on SMP machines is one such case in point. Yes, I know there's one DPC queue per CPU in an SMP machine, and the book would have said that, had the publisher not filed my changes away in a dusty drawer and forgotten them. While an author is ultimately responsible for the contents of his book, the most I can do here is to apologize and make the corrections as widely available as possible. Some of Mr. Viscarola's other comments puzzle me. He claims I never point out that the location of a device's control registers is a function of the device and not the CPU architecture. Yet on page 27 of my book I say: "Even on CPUs with a separate I/O space, some peripherals memory-map their control registers anyway. This improves the performance of high-speed devices with large register sets, since I/O instructions are typically much slower than memory-access instructions." To keep this note from becoming too long, I'm not going to respond point-by- point to the rest of Mr. Viscarola's review. While he raises a number of technical issues that need to be addressed in future revisions of the book, I disagree with his conclusion that the Windows NT Device Driver Book is without value. Regards -- Art Baker Cydonix Corporation Training that makes a difference cydonix@clark.net NT programming, tuning, drivers