The Deseret Alphabet is a phonetic alphabet for the spelling of English promulgated by The Church of Jesus Christ of Latter-day Saints ("Mormon") in the 1850's and 1860's. Its last significant use was in 1869, when the Church published an edition of the Book of Mormon using the Deseret Alphabet, but it's remained of interest to Latter-day Saints since and continues to be used in private documents.
In 1869, The Church of Jesus Christ of Latter-day Saints published an edition of the Book of Mormon using the Deseret Alphabet which has since become a collector's item commanding prices of over $2000. The text was prepared by Orson Pratt at Brigham Young's direction.
To celebrate the sesquecentennial of the settlement of Salt Lake City–and because I have the technology to do it–I've matched Brother Orson's work and done him two better by also producing Deseret Alphabet versions of the Doctrine and Covenants and Pearl of Great Price. The whole shebang is available as a pdf file which can be viewed on a whole slew of platforms using Adobe's free pdf viewer software, and it doesn't require any fonts to be installed on your system (in theory).
One caveat: My edition was produced by computer, and not by hand. This means three things.
One is that the current software doesn't handle cases where a spelling has two or more pronunciations ("read", "lead", "bow") at all well. Someday I'm going to add a bit more sophistication to handle that.
Another is that the computer sometimes had to figure out how to pronounce a certain word on its own. This is particularly true of proper nouns. I used two databases of English phonetic data available on the Internet for the basic source of the conversions, and I supplemented these with a couple of hundred pronunciations of my own. There are still a couple of hundred words where the computer—a Macintosh running Mac OS 8 and using Apple's built-in text-to-speech software—had to come up with its own pronunciations. It does a fairly good job on the whole but there are some surprises.
Finally, a linguist would distinguish between a phonemic and a phonetic spelling of a word. The former is a reflection of the abstract sounds we theoretically make when we talk, and the latter the sounds we actually make. A good example is the "-ty" suffix of the number words 20, 30, 40, 50, 60, 70, 80, and 90. All these words end with the same phonemes (which is why we spell them all the same way), but in actual fact, the "t" sometimes becomes a "d" when we actually speak them (listen closely when you talk). My software is based on phonetics and not phonemics, which means that the words are spelled in the Deseret Alphabet the way they are actually pronounced, and not in accordance with their theoretical sounds.
Bottom line: Some of the spellings you'll see look strange. That's OK. Some of these strange spellings are mistakes, it's true, but some aren't. (Meanwhile, I'm still trying to figure out why Orson Pratt chose some of the spellings he used, such as for the word "Deseret" itself, so I'm in good company.)
The up side is that it only took me a couple of hours to produce the actual text. In fact, I spent a lot more time editing the text to make it look pretty on the page than I spent actually generating it.
There are not many computer typefaces available for the Deseret Alphabet. Most of the ones which are available still overlay the Deseret Alphabet glyphs over ASCII (or 8859-1) characters. This is a dangerous practice and one which I discourage.
Inasmuch as the Deseret Alphabet is included in Unicode, and has been since Unicode 3.1 was released in March 2001, production and exchange of text data for the Deseret Alphabet should be done using Unicode. Unicode is generally supported by most operating systems and major applications (as of September 2006), so this should not be an unreasonable burden.
Nonetheless, for the sake of people running older Macintoshes, I still make my Deseret Language Kit available, which can be used with Mac OS 7.1 through 9.2.
For cross-platform use, two fonts are listed at David McCreedy's Gallery of Unicode Fonts: James Kass’ Code2001 and MPH 2B Damase. Neither, thankfully, is based on the 19th century originals, which are not elegant by current typographic standards.
Mac OS users need not install additional fonts; Mac OS X 10.4 (Tiger) includes a set of Deseret Alphabet glyphs in the Apple Symbols font. These are, unfortunately, based on the 19th century originals.
For input, the situation is somewhat worse. The Deseret Language Kit includes a keyboard. Both Ken Beeseley and I are working on keyboard layouts for Mac OS X. I'm not aware of any keyboard layouts for Windows. In any case, both Mac OS X and Windows allow the input of arbitrary Unicode characters. On Mac OS X, this is most easily achieved using the Character Palette.
The Unicode Standard is a new 16-bit character set under development by various movers and shakers in the computer industry such as Microsoft, Digital, HP, and my own Apple Computer, Inc. Its goal is to provide a common way of representing all human writing systems in current or past use. As such, it's quickly becoming the lingua franca for text on the Internet.
The original design of Unicode allowed the inclusion of some 65,534 characters, with some six thousand of those "private use characters" which basically anybody can use any way they like. When it became clear that this simply wasn't enough to accomodate all the large, rare character sets such as Egyptian hieroglyphics and uncommon Han ideographs, an extension mechanism (surrogates) was added to the standard which allow the addition of another million characters.
The problem was, a vicious circle quickly arose. Nobody started to implement surrogates because there were no characters encoded using them, and nobody wanted their characters to be encoded using surrogates because nobody was implementing them.
To break the cycle, someone needed to propose for encoding a writing system which was at once real and at the same time so rare and/or dead that nobody would object to its being encoded with surrogates. I proposed the Deseret Alphabet in late 1996 as such as script, and it was quickly accepted.
Being approved for adoption and being adopted are two different things, however. It was March 2001 before the Deseret Alphabet was formally added to Unicode, with the release of Unicode 3.1. Other scripts, such as Shavian, the Cypriot syllabary, Linear B, Gothic, and Old Italic have also been added to Unicode using surrogates and dozens of additional scripts have been earmarked for encoding that way. The Unicode encoding also includes two rare ligatures which were included in some of the formal descriptions of the Deseret Alphabet.
As part of Unicode, the letters of the Deseret Alphabet are the scalar values from 0x00010400 through 0x0001044F.
Characters in Unicode are assigned scalar values, a 32-bit number that uniquely identifies them. There are three basic schemes for mapping these scalar values into actual sequences of bytes used on a computer.
UTF-32 has been recently adopted. This is a thirty-two bit form of Unicode. Each scalar is represented by itself.
UTF-16 is the oldest form of Unicode. It uses one or two sixteen-bit words to represent a scalar value. If the scalar is less than 0x00010000, then the scalar value is packed into sixteen bits; otherwise, it is split into a two surrogate values. Interconversion between the scalar value (SV) and the high and low surrogates (HS and LS) is accomplished by the following algorithms:
SV = 0x00010000 + ( HS - 0xD800 ) * 0x0400 + ( LS - 0xDC00 )
HS = ( SV - 0x00010000 ) / 0x0400 + 0xD800
LS = ( SV - 0x00010000 ) % 0x0400 + 0xDC00
UTF-8 is a variety of Unicode which is intended for using Unicode in situations or applications which are designed for text that comes in eight-bit chunks. Many older operating systems and transmission protocols are like this. (C programmers will be familiar with the idea of strings being represented as arrays of characters ending with a 0 byte which serves as an end-of-string flag. Since many Unicode code points have one byte or the other equal to zero, these characters will break existing C libraries. UTF-8 is one solution to this problem.)
UTF-8 also has the advantage that characters from the ASCII character set in their traditional representation of 0x00 through 0x7F are represented by the single bytes 0x00 through 0x7F. This means that English text in UTF-8 looks pretty much as it has for the past thirty-five years.
The transformation between UTF-16 or UTF-32 and UTF-8 is relatively straightforward and described in The Unicode Standard, Version 3.0, pp. 45–47.
(The whole issue of encoding forms of Unicode is complicated by the fact that different computer architectures don't agree on how to combine two eight-bit bytes to form a single sixteen-bit word.)
It is strongly recommended that any implementations of the Deseret Alphabet conform to Unicode/10646, if at all possible.
It is most decidedly improper to hijack the standard ASCII code points and assign various Deseret Alphabet letters to them so as to make typing easy. This is considered a very Bad Thing™ in the world of text encodings.