The Unicode Standard Version 3.0, w. CD-ROM: Worldwide Character Encoding [Englisch] [Gebundene Ausgabe]

Central to the book, taking up the larger part of it, are the tables of the characters themselves, printed large with annotations and cross-references. If you enjoy the lure of strange symbols and curious writing systems then browsing these will occupy delightful hours.

For the Latin alphabet alone there are pages of accented letters and extended Latin alphabet characters used in particular languages or places or traditions: Pan-Turkic "oi", African clicks and other African sounds, obsolete letters from Old English and Old Norse, an "ou" digraph used only in Huron/Algonquin languages in Quebec, and many others, particularly those used for phonetic/phonemic transcriptions.

The Greek character set includes archaic letters and additional letters used in Coptic.

Character sets carried over from previous editions with additions and corrections are Cyrillic (with many national characters), Armenian, Georgian, Hebrew, Arabic (again many national and dialect characters), the most common Hindu scripts (Devanagari, Bengali, Gurmukhi, Gujarati, Oriya, Tamil, Telugu, Kannada and Malayalam), Tibetan, Thai, Lao, Hangul, Bopomofo, Japanese Katakana and Hiragana, capped by the enormous Han character set containing over 27,000 of the most commonly used ideographs in Chinese/Japanese/Korean writing. Then there are the symbols: mathematical/logical (including lots of arrows), technical, geometrical, and pictographic. You'll find astrological/zodiacal signs, chess pieces, I-Ching trigrams, Roman numerals not commonly known, and much more.

Scripts appearing for the first time this release are Syriac, Ethiopic, Unified Canadian Aboriginal Syllabics, Cherookee, Runes, Ogham, Yi, Mongolian, Sinhala, Thaana, Khmer, Myanmar, complete Braille patterns, and keyboard character sets. And yes, there are public domain/shareware fonts available on the web that support these with their new Unicode values.

There are very good (and not always brief) descriptions of the various scripts and of the special symbol sets. Rounding out the book are some involved, turgid (necessarily so) technical articles on composition, character properties, implementation guidelines, and combining characters, providing rules to use the character properties tables on the CD that accompanies the book. After all, this is the complete official, definitive Unicode standard.

Of course this version, 3.0, is already out-of-date. But updates and corrections are easily available from the official Unicode website where data for 3.1 Beta appears as I write this. My book bulges with interleaved additions and changes. And that's very good. Many standards have died or been superceded because the organizations behind them did not keep up with users' needs or the information was not easily accessible.

Caveats?

The notes on actual uses of the characters could be more extensive, particularly on Latin extended characters. More variants of some glyphs should be shown, as in previous editions, if only in the notations.

Some character names are clumsy or inaccurate (occasionly noted in the book), because of necessity to be compatible with ISO/IEC 10646 and with earlier versions of the Unicode standard. For example, many character names begin with "LEFT" rather than "OPENING" or "RIGHT" rather than "CLOSING" though the same character code is to be used for a mirrored version of the character in right-to-left scripts where "LEFT" and "RIGHT" then become incorrect. And sample this humorous quotation from page 298: "Despite its name, U+0043 SCRIPT CAPITAL LETTER P is neither script nor capital--it is uniquely the Weierstrass elliptic function derived from a calligraphic lowercase p."

This book is essential for software engineers, at least for the next ten years or so. All programmers should understand characters, and UNICODE is the best we have for now. Even if you don't need it in your personal library, you need it in your company or school library.

The standard is flawed, as all real standards are, but it is a functioning standard, and it should be sufficient for many purposes for the near future.

The book itself is fairly well laid out, contains an introduction to character handling problems and methods for most of the major languages in use in our present world as well as tables of basic images for all code points. Be aware that these are _only_ basic images. For most internationalization purposes, be prepared for more research. (And please share your results.)

**** Finally, UNICODE is _not_ a 16 bit code. ****

(This is well explained in the book.) It just turned out that there really are over 50,000 Han characters. (Mojikyo records more than 90,000.) UNICODE can be encoded in an eight-bit or 16-bit expanding method or a 32-bit non-expanding method. The expanding methods can be _cleanly_ parsed, frontwards, backwards, and from the middle, which is a significant improvement over previous methods.

Some of the material in the book is available at the UNICODE consortium's site, but the book is easier to read anyway. One complaint I have about the included CD is that the music track gets in the way of reading the transform files on my iBook.

At 1040 large (8.5 x 11) pages it is the ultimate guide to unicode. With information on scripts and glyphs I had no idea even existed.

However if you are just getting started with Unicode I would recomend you get Unicode a Primer written by Tony Graham from M&T books. If you understand or feel you are starting to understand Unicode then The Unicode Standard Version 3.0 is the best comprehensive reference on the subject out today.