Computerizing Alphabets

http://www.nytimes.com/2003/09/25/technology/circuits/25code.html?8hpib

For the World's A B C's, He Makes 1's and 0's
By MICHAEL ERARD

MICHAEL EVERSON, a 40-year-old typographer who lives in Dublin, considers
himself blessed because he has found his life's work: to be an alphabetician
to all the peoples of the world. Mr. Everson's largest project to date - a
contribution to a new version of Unicode 4.0, an international standard for
computerizing text - is cementing his reputation.

His mission has taken him to Kabul, Afghanistan, and Helsinki, Finland; to
Beijing, Tokyo and Redmond, Wash. His Dublin house is a shrine to his
obsession with every writing system that humans are known to have created -
148 of which Mr. Everson says he can use for writing his name. In the hallway
is an icon of the saints Cyril and Methodius (Cyril is often credited with
inventing the Cyrillic alphabet) and a page from a Maghreb manuscript from
North Africa.

He keeps a photo of a stone inscribed with ogham, an ancient Irish alphabet
that looks like hash marks, in a silver frame. His office chair, parked in
front of a Macintosh G4 laptop named Cyril, is upholstered with dark blue
fabric dotted with Egyptian hieroglyphics. Surrounding his desk are shelves
heavy with books on the origins of cuneiform and other writing systems.

He remains fond of the Roman alphabet, however. "Of all the alphabets, it's
the best one," he said in a telephone interview.

For the last 10 years, Mr. Everson, who has American and Irish citizenship,
has played a crucial role in developing Unicode, which might be viewed as the
computer age's Rosetta stone. Mr. Everson explains Unicode as "a big, giant
font that is supposed to contain all the letters of all the alphabets of all
the languages in the world."

A more technical explanation of Unicode is this: When Mr. Everson sends e-mail
in ogham, his computer isn't sending ogham letters through the ether. Instead,
strings of 0's and 1's are transmitted, and when they arrive on a friend's
computer, they generate on its screen the same ogham letters that Mr. Everson
typed. Unicode is the master list that resides in both computers and
translates individual letters and symbols into strings of 0's and 1's and back
again. Most current software is Unicode-compliant, which means that this
master list of all the world's writing systems has been built into operating
systems, browsers and software.

The code assigned to all 96,000 characters is handled only by programmers in
its naked form, while computer users (and sometimes vendors) install the
specific fonts that represent a specific alphabet. A font renders a language
readable to humans; Unicode renders a font readable to computers.

Most people don't even realize Unicode is at work. "Unicode is like plumbing,"
said Rick McGowan, the vice president of the Unicode Consortium. "Yet it's the
most far-reaching and ambitious multilingual project in history."

It is because of Unicode that bloggers can muse in Arabic and domain names can
exist in Chinese, or that National Security Agency analysts can scour the
Internet for reports on the latest threats in East African newspapers.
"Because of Unicode," Mr. McGowan said, "you can plunk down a vanilla
off-the-shelf computer into a cafe anywhere in the world and have any user in
any language walk up to it and use it for accessing the Web."

Mr. McGowan was a member of the group of computer scientists and linguists who
set out to create the system in 1990 to solve an emerging problem.

As a growing number of users wanted to write in their own languages on their
machines, companies had developed methods for computerizing text that did not
appear in the Roman alphabet. With the rise of the Internet, the problem
became more complicated because there was no assurance that all those machines
would be able to share text data. Without a shared standard, manufacturers and
even governments were creating isolated islands of data, each with its own
standard, and each computer would have to be customized to the writing system
that the owner wanted to use. Many users could not write e-mail, build Web
sites or search databases in their own languages and alphabets.

The solution was Unicode, an international standard for character encoding.
(Character encoding is simply any system that transmits textual information;
Morse code is one example.) Last month the latest version of the standard,
Unicode Standard Version 4.0, was published. It contains encodings (that is,
unique strings of 0's and 1's) for some 96,000 letters and symbols.
Approximately 70,000 of them are Chinese characters. Unicode also contains
support for 54 other writing systems, from Mongolian to Thai to Gothic to
Cyrillic.

Mr. Everson said he had worked on about 5,000 of those characters. Version 4.0
includes characters for Linear B (for which he designed the font) and other
ancient Mediterranean alphabets that are used mainly by scholars.

As vast as Version 4.0 seems, it is still not complete, and nearly 100 writing
systems remain to be encoded. Mr. Everson is haunted by the prospect that
Unicode may never be finished. "Imagine how you would feel if your name was
François, but there was no ç available," Mr. Everson said. "You would be
irritated that your phone bill came addressed spelling your name wrong. Now
imagine if your language used a totally different alphabet and you couldn't
use computers at all because of it. It's a question of human rights, really."

An incomplete Unicode is a looming possibility, however. Now that the writing
systems of the major computer markets are encoded, the computer companies that
once backed the Unicode project are beginning to question the expense. To
ensure that the remaining writing systems are included, a project named the
Script Encoding Initiative has been set up at the University of California at
Berkeley to enlist scholars and apply for funds from private foundations to
hire Mr. Everson full time.

One result of the dwindling interest from the private sector is to put
pressure on Mr. Everson to complete large projects. "They say, 'Here, Michael,
can you do Egyptian?' It's like, no. Egyptian is on my list, Egyptian is hard,
and it's big."

To pay the bills, Mr. Everson works as a typesetter. He is currently setting
type for "Gargantua," by Rabelais, in Irish. Other notable projects include
the first publication of the entire New Testament in Cornish, as well as an
English-Cornish dictionary.

But Mr. Everson admits that he is most drawn to the encoding work. "It's best
for me in my life to be consumed by an obsession in writing systems, because I
am extraordinarily well suited to dealing with it," he said.

Mr. Everson was first attracted to far-off places and languages by the books
of J. R. R. Tolkien, which he first read as a 13-year-old living in Tucson.
(Mr. Everson said he still has a "soft spot" for Tengwar, one of the alphabets
that Tolkein invented for his made-up languages of Sindarin and Quenya.) "The
Lord of the Rings" led him to Anglo-Saxon and the epic poem "Beowulf," which
he decided to translate from Old English at the age of 14. From his copy of
the "Beowulf" manuscript, he practiced copying Anglo-Saxon letters with a
calligraphy pen.

Then he graduated to designing fonts on his Macintosh, tackling Georgian and
Cyrillic, then Devanagari. After feeling dissatisfied with graduate school at
U.C.L.A., he moved to Ireland in 1989 and began typesetting for a living while
designing exotic fonts on the side for writing systems including Cherokee,
ogham and Sinhala.

In 1993, he saw a request from the Unicode Consortium for revisions involving
some archaic scripts, one of which was ogham. "It was like, ooh, this is
Irish, let's look into it," Mr. Everson recalled. He also sent comments on
Burmese, Ethiopic, Yi and Sinhala. "I started in early," he said. "I just
plunged right in."

Meanwhile, at the Unicode offices in Silicon Valley, people were impressed
with the work by this relative unknown in Ireland. Mr. McGowan remembers the
first proposal he received from Mr. Everson, on a particular character in
ogham. The first time they met, Mr. McGowan was so captivated by Mr. Everson's
charm and erudition that he saved his name tag. "Michael is a pretty special
guy," he said. "Also, he wrote the month with a Roman numeral. I thought that
was amusing."

Mr. Everson's knowledge of the world's writing systems has made him
indispensable to Unicode. "At this point, Michael is probably the world's
leading expert in the computer encoding of scripts," Mr. McGowan said. "Nobody
else comes close to having his detailed knowledge about so many scripts and
how they are, or should be, encoded."

Deborah Anderson, a researcher in the linguistics department at Berkeley who
heads the Script Encoding Initiative, credits Mr. Everson with getting most of
the lesser-known writing systems into Unicode. His 220 proposals or technical
documents, she said, make him "without question the single most prolific
Unicode proposal author around."

As exotic as the subject matter is, the work itself is fairly dry. It involves
finding authoritative texts, assembling examples and seeking out experts and
then working with them to determine how many characters there should be and
how they should look.

"It's one thing to be a specialist who reads Ugaritic," Mr. Everson said.
"It's another to be a person who can figure out the essential bits of the
writing system in terms of the way Unicode works."

He takes a remarkably long view of the impact of his work. "There's
satisfaction in knowing that the work of analyzing and encoding these
languages, once done, will never need to be done again," he said. "This will
be used for the next thousand years."

Mr. Everson also seems to enjoy the human interactions. He is proud of working
with the grandson of Osman Yusuf Kaynandid, who invented the Osmanian script
in Somalia in 1922. He also likes to tell about how he met the president of
the Tibetan Calligraphy Society at a Unicode meeting in Copenhagen. Mr.
Everson had helped the organization ensure that Tibetan was included in the
standard. The president showed Mr. Everson how to write his name in Tibetan
with a highlighter pen.

"He thanked me," Mr. Everson said with reverence. "I couldn't believe that,
because his organization has been in existence for over a thousand years."