by Peter H. Salus

Peter H. Salus is a member of the ACM, the Early English Text Society, and the Trollope Society, and is a life member of the American Oriental Society. He has just become Chief Knowledge Officer (= Mr. Know-It-All) at Matrix.Net.

<peter@pedant.com>

BOOKS REVIEWED IN THIS COLUMN

THE UNICODE STANDARD, VERSION 3.0
The Unicode Consortium
Reading, MA: Addison-Wesley, 2000. Pp. 1,040 + CD-ROM. ISBN 0-201-61633-5.

[INCR TCL/TK] FROM THE GROUND UP
Chad Smith
Berkeley, CA: Osborne/McGraw Hill, 2000. Pp. 746. ISBN 0-07-212106-8.

PRINCIPLES OF NETWORK AND SYSTEM ADMINISTRATION
Mark Burgess Sussex, U.K: Wiley U.K., 2000. Pp. 400. ISBN 0-471-82303-1.

THE PROCESS OF NETWORK SECURITY
Thomas A. Wadlow
Reading, MA: Addison-Wesley, 2000. Pp. 283. ISBN 0-201-43317-6.

What If It Ain't English?

While a large number of people in the world speak and read a version of English (as a first, second, or nth language), only a very few cultures employ the 26 characters of our alphabet. Representing the various scripts becomes vital. Just how many we should be trying to accommodate is totally unclear.

In 1996, Oxford University Press published The World's Writing Systems, a long-awaited tome edited by Peter T. Daniels and William Bright. It's over 900 pages and costs about $150. I consider it a treasure. When Unicode Standard, Version 1.0 appeared in 1991 (vol. 1) and 1992 (vol. 2), I was pleased, but not overwhelmed: there were a number of gaps, and a large number of unattractive choices had been made.

2.0 (1996) was a vast improvement. 3.0, the 1,000-page, large format at hand, is yet better. But its choices are quite strange at times.

For example, under "European Alphabetic Scripts," there are descriptions for Runic and for Ogham. I presume that some elderly Scandinavians and old Irish use these in sufficient quantity. (The Unicoders don't seem to realize that there were non-Scandinavian runes.) On the other hand, I gather that the 25 million folks in Afghanistan carry too little weight for the Pashto version of Arabic to be treated.

Under Cyrillic, Byelorussian and Ukranian are treated as variants of Russian letter forms; Glagolithic is unsupported.

South Asia has its problems, too. It might have been helpful for the Unicoders to refer to Daniels and Bright. In Unicode 3.0 (p. 211), it claims that Santali is written in "Devanagari." In Daniels and Bright, Norman Zide tells us that Santali is written in Ol' Cemet as well as "four 'older' scripts: Devanagari, Bengali, Oriya, and Roman." (Other "tribal" groups have also adopted Ol' Cemet.)

I won't beat the Unicoders more than this. There are hundreds of millions of people on this planet whose languages/ scripts are not well done-by by the Unicoders. But I recognize that this is a work-in-progress, and that we won't require Pashto or Santali speakers to employ English when they get PCs.

For those who do scholarly work, I suggest getting a Mac and the font diskettes/CDs from Lloyd Anderson at Ecological Linguistics or the Egyptian hieroglyphic fonts from Cleo Huggins (based on the Gardiner OUP fonts).

I trust that someday even Akkadian, Babylonian, Hittite and Sumerian cuneiform, and Mayan hieroglyphs will find their way into Unicode.

But, after nearly a decade, we ain't quite there yet.

Tcl Me

The successor to C wasn't either P or D, but C++. Following this example, Tcl led to Michael McLennan's "extension," [incr Tcl/Tk]. [incr Tcl/Tk] is useful for developing large-scale applications and building extensible GUIs. Chad Smith has an engaging style and has really produced a volume that will teach and inform: you need not be a Tcl programmer to understand this book. As Tcl isn't O-O, if you want to employ O-O capabilities, you're going to need [incr Tcl/Tk]. And if you aren't an adept, you'll want Smith. This is a fine book.

Another Winner!

I keep Nemeth et al. and Aeleen Frisch at hand for referencing the system admin tasks we all need to do. Burgess's fine book is something "completely different." It is a well-articulated introduction to a corpus of guiding principles for system administrators.

And as we live in a world of heterogeneity, Burgess "covers" UNIX, UNIX-like, DOS, Windows, Mac, Amiga, and NT systems.

Burgess says that he wants to express a sound and logical way to approach networked systems. While I can find nits (that's a reviewer's job, isn't it?), I consider this an important book. More and more talk of certification can only lead to a body of knowledge and a set of tenets that are "required."

I think that Burgess will become part of the required reading for future (and current) system administrators.

Are You Secure?

I feel that I've been reading and writing about security for too long a time. But Tom Wadlow has turned out a fine, small (under 300 pages) volume for the network-security administrator. As Wadlow points out, "If you are a five-person real estate office, with individual Internet dialup connections for all your Windows 98 desktop PCs, this is not the book for you." The only real criticism I have of this book is that Wadlow gives no "further readings" nor a bibliography. I wish that Bill Cheswick, Dorothy Denning, Gene Spafford, Simson Garfinkel, and many others were placed appropriately. Even the citations from Sherlock Holmes are merely ascribed to Conan Doyle.

Collections

If you go to the IETF site, look up RFCs, and then download and print them, as I do, you end up with oceans of paper, much of which flutters like leaves in the wind, shedding all over the keyboard, desk, floor, etc. I also have about six feet of files. Morgan Kaufmann (a division of Harcourt Brace) is performing a good deed. They have gotten Pete Loshin to edit a series containing the full RFCs by subject.

The first volume was The Big Book of IPsec RFCs, the second was The Big Book of World Wide Web RFCs, and the third was The Big Book of Internet File Transfer RFCs. I have found them quite useful but cannot review them: I wrote the prefaces to the ftp volume and to the (forthcoming) Big Book of IPv6 Addressing RFCs — and have other involvement as well. But they are worth looking at, if only as a way to stop deforestation.