I apologize for not updating this page for several months. I've been so busy this year, I haven't been able to prepare a well-organized presentation.
Basically, the good news is that the new Volume 3 is done (see below), and so is my new 700-page book Digital Typography (about which I'm real excited right now, you will love it when you see it next January!). At present I am working full time on MMIX. Meanwhile this year I've come up for air every now and then to make a few Public Appearances, some of which are now past and others are still pending:
Beware of bugs in the above code; I have only proved it correct, not tried it.
More than a hundred webpages ascribe that quote to me, and it sounds like something I might well have said. But lately people have been asking me for the authentic source, and I've totally forgotten where I wrote it. Probably not in a published paper, since I've written few papers in the first person that include untried code.
My best guess is that it was in a letter I sent to somebody. So if you were that somebody, or if you have any other clues about the source of that line, please let me know.
Computer scientists need an English word for 16-bit quantities, just as ``nybble'' and ``byte'' are used for 4-bit and 8-bit things. So I have decided to coin the word ``wyde'' for this purpose: Two nybbles is one byte, two bytes is one wyde.
I tried this idea on several colleagues and found reasonably enthusiastic agreement. The pluses for this word are that it trips easily off the tongue, and it has suitable connections with the ``wide characters'' of C/Java languages (type wchar_t), also with the 16-bit ``word'' of oldtime computers like the PDP11 and Intel 8086, but pronounced with sortofa Brooklyn accent. The new C library functions that use strings of 16-bit characters are called wprintf, wgetc, etc.; we can think of them as abbreviations for wyde-print-formatted, wyde-get-char, and so on.
I certainly don't want to call 16-bit things ``words''---as Intel and DEC engineers amazingly do even as they build machines with word lengths of 64 bits---because most RISC machine designers know that a word is 32 bits long, just as IBM has said for a long time. (MIPS people speak of 64-bit things as doublewords, but Cray people think there are 64 bits to a word; clearly the word ``word'' is highly ambiguous.)
The only other competitors known to me are the terms ``parcel,'' which has been used for 16-bit components of instructions by people at Cray Research and also in the AT&T Hobbit group; and ``rune,'' used by the developers of Plan 9. About twenty years ago, some people at the University of Florida proposed the name ``chawmp'' for ``half of a machine word,'' which at that time meant either 16 or 18 bits. That term seems to have died out; in any case it is as ambiguous as ``word.'' Both parcel and chawmp lack the merits of ``wyde.''
I searched the Web and found WYDE, Inc., a French company that does innovative computer work, so they should be happy with further uses of their name. Then there's radio station WYDE in Birmingham, Alabama. I also found Canada WYDE-CASA, a nonprofit charity that publishes a magazine; I doubt if they will mind my usage either. There are people whose surname is Wyde, but there are also people named Short, Char, etc. The only prior use of ``wyde'' I found with all lowercase letters was in quotations from Chaucer.
Thus, I encourage everybody to begin world-wyde usage of this term.
Already a new online journal, THINK/WYDE --- planned as the successor to BYTE magazine for the new millennium --- has been announced. (Its URL is soon to be www.thinkwyde.org.)
What should I call 32-bit and 64-bit quantities? At the moment I'm having good luck with the terms ``tetrabyte'' and ``octabyte,'' also called ``tetra'' and ``octa'' for short, since these names make it clear how to advance pointers past such items in a byte-addressed machine. (The slightly nicer words ``quartet'' and ``octet'' have already been reserved, long ago, as coding theorists' technical names for 4-bit and 8-bit quantities. There's a group of programmers, graphic artists, and musicians in Germany that calls itself ``Tetrabyte.'') But if anybody has a significantly better idea, I'm happy to listen --- modulo the situation discussed in don't send me email.
On September 30, members of Stanford's Computer Science Department gathered to bid farewell to Phyllis Astrid Benson Winkler, on the occasion of her retirement after 32 years of service. During those years she was one of the key reasons for our department's successes; thus she indirectly had a substantial influence on the progress of computer science as a whole. We appreciated her intelligence, her efficiency, her world-class expertise at producing beautiful technical documents, her team spirit, her willingness to go the extra mile, and her contagious laughter.
I was fortunate to have had Phyllis as my secretary and essential co-pilot during the past 28 years; without her I could not have accomplished nearly as much. She typed more than 200 of my papers, most of which required several rounds of revisions. She buffered all of my mail and telephone messages. She administered the editorial work of more than a dozen technical journals, and helped out with numerous research projects. She made online indexes of all the correspondence in our files. She did all of the initial keyboarding for the new editions of The Art of Computer Programming, Volumes 1 and 3 (see below) --- amounting to more than 1500 printed pages of what printers used to call ``penalty copy'' because it is so hard to do. And so on and so on; what a team we made! And she was simultaneously also serving as secretary for several other faculty members.
When I originally wrote the TeX typesetting system, I intended it to be a tool just for Phyllis and me, but mostly for Phyllis. Soon other people decided to use it too, but Phyllis's influence on the TeX project has nevertheless been enormous. One of the events at her retirement party was the reading of a resolution recently passed by the board of directors of the TeX Users Group, expressing their appreciation for all of her contributions during the past 20 years. I'm sure people all over the world are sad that they will no longer be communicating with her at Stanford, yet wishing her happiness as she changes to a life of voluntary community service.
I can't express in words the enormous debt of gratitude I feel, but I have tried to do that in part by dedicating the book Literate Programming to her. I certainly wish her a long and productive life in retirement.
I recently completed a project that I started in February, 1994, namely to compile electronic files of updates to Volumes 1, 2, and 3 of The Art of Computer Programming, based on all the marks I had made in my personal copies of those books since 1975. (See the TAOCP page for further details of that work.) When Silvio Levy saw the first draft of those lists, he volunteered to retypeset all three volumes with the new material included, and his massive task is now completed. Furthermore, Jeffrey Oldham volunteered to enhance the illustrations by putting them in MetaPost form. This has made it possible for new editions of Volumes 1--3 to be prepared with a minimum of disruption to my work on Volume 4.
On 5 July 97 I received my copy of the glorious 3rd edition of Volume 1, and exactly four months later I received the even more glorious 3rd edition of Volume 2. The still more stupendous Volume 3 (2nd edition) arrived on the 23rd of April, 1998---thereby completing the first new editions of these books for more than 15 years. A specially-priced boxed set containing all three volumes was published in October.
(Further details can be found in an Addison-Wesley interview about the new editions --- including some pictures. See also the live Q&A interview hosted by ComputerLiteracy books with RealAudio!) And an Amazon.com interview.)