Book Review:Upgrading and Repairing PCs, 16/e, Scott Mueller

An old companion becomes less necessary, October 17, 2004


I have purchased three of the last five revisions of this book, and
bought the latest as a matter of course when it first became
available. It's the first one where I have wished that I had looked at
it physically before purchase, because I would have not purchased it
had I done so.

For many years, this volume has been an industry standard, on the
desktops of tech support personnel, PC purchasing agents, vocational
educators, and hobbyists. It has grown like Topsy with each revision,
steadily becoming larger and larger as more and more detail regarding
newer releases in x86-compatible desktop and mobile CPUs, bus types,
form factors, and compatibility issues evolved. However, previous
standards are described often in mind-numbing and wordy detail as
well. While some editing has taken place, it's not nearly enough.

If you have one of the last two or three editions of this book, don't
buy this one. It's just more paper weighing down your already sagging
bookshelf. If you are starting from scratch, this may still be a
useful book, but be warned: it's a colossal aggregation of data, much
of it of historical interest only.

Much of the "legacy" data is of primary interest to embedded systems
designers and developers, but Mueller is strictly a desktop person:
embedded people will be frustrated with this volume. Also, Mueller is
wholly Microsoft-centric, meaning that there is a lot of DOS and NT
stuff in here but no Unix, Unix-like (Linux or Free/Open/NetBSD), or
embeddable (VxWorks or QNX) information whatsoever. Considering the
average Linux user is going to be far more likely to need this type of
knowledge than the average Windows user, one gets the idea that
Mueller is just not comfortable with or even conversant with other
environments than Microsoft and has elected to dodge the issue.
Previously there was a "Linux Edition" of the book, where an outside
team simply edited out the Microsoft discussion and put in some
coverage of Linux video and sound configuration and some generic
boilerplate. I find it a little disturbing because one simply can't
consider oneself a computing professional in 2004 without a fair level
of Unix competency:it just isn't credible.


Although the text is often shovelware, the accompanying DVD is a
disappointment as the bundled software for disk testing and
partitioning included in previous editions' disks is no longer
included. Some previous editions had some very useful stuff including
a fully functioning commercial partitioning program.

In my opinion the book should be blue-pencilled down to about 700
pages with much of the remainder put in .pdf format and put on the
DVD. Very classy would be providing a DVD and a bootable CD with a
live OS enabling network and web access, disk utilities and so forth.
We used to highly value the QNX Demo Disk floppies, the images for
which are sadly no longer on the excellent QNX website, and perhaps
QNX would allow their OS to be used for a flavorful and functional
live CD to enable access of the outside world, M$ file systems on hard
drives for copying or repair, and formatting and partitioning of hard
disks and rewritable media. A DOS-based bootable CD might be less
featureful but would at least enable disk repair and file editing and
could be made with FreeDOS. Also, given the nature of the work,
perhaps printing the book on "Bible paper" with a ruggedized but
flexible cover-such as the Machinery's Handbook and other professional
reference works-would be a better choice than the current
consumer-quality binding used on most Que hardbacks.