perm filename ROY.PUB[ROY,LCS] blob
sn#214115 filedate 1976-05-06 generic text, type T, neo UTF8
00100 .DEVICE XGP
00200 .FONT a "BDR40"
00300 .FONT 1 "BASL30"
00400 .FONT 2 "BASI30"
00500 .FONT 3 "BDR25"
00600 .FONT 4 "BDI25"
00700 .FONT 5 "NGR20"
00800 .SPACING 25 MILLS
00900 .TURN ON "%";
00950 .!XGPCOMMAND"/TMAR=50,PMAR=2100,BMAR=50/
01000 .PAGE FRAME 56 HIGH 70 WIDE
01100 .SELECT 1
01200 .NOJUST
01300 .NOFILL
01400 .NEXT PAGE
01500 .NARROW 4,0
00100 .begin center
00200 %aCenter for Computer Research in Music and Acoustics
00300 %1Artificial Intelligence Laboratory, Stanford University
00400 Stanford, California 94305
00500 %5Co-directors: John M. Chowning, John M. Grey, James A. Moorer, Loren Rush, Leland C. Smith
00600 .end
00700 .group skip 4
00800 .begin flush right
00900 %1May 5,1976
00950 .!XGPCOMMANDS←"/TMAR=50/PMAR=2000/BMAR=150"
01000 .end
01100 .group skip 2
01200 Dear Mr. Anderson,
01300 .begin fill adjust
01400 Enclosed you will find a few copies of the first page of your
01500 INTERMEZZO as printed by my music graphics program. This turned out
01600 to be a somewhat more difficult page to set up than most I've dealt
01700 with. This is because of the very active inner parts. (Orchestra
01800 scores are usually easier to set up.) You may note that measure 10 is
01900 the same as measure 6. In this case I simply copied the data for
02000 measure 6 into the new position. Thus only the new material on the
02100 third line had to be typed in. Actually I could have done the same
02200 kind of thing for some of the other measures too.
02300
02400 I have enclosed listings of the basic input data for each of
02500 the lines. These lists probably aren't too comprehensible without a
02600 detailed explanation. Also I have enclosed the parameter data for
02700 line 2 of the music which the program generated internally. In
02800 general the list of numbers after each word gives information about
02900 the vertical and horizantal position of the item, its size and all
03000 sorts of things like the number of tails on notes, the accidentals,
03100 dots, stem direction and length, etc., etc. For the most part the
03200 operator doesn't have to bother about all these numbers except when
03300 making some of the finer adjustments of positions.
03400
03500 The actual program that processes all this data is written in
03600 a combination of FORTRAN and PDP10 machine language and occupies over
03700 a 100 pages. The program is "modular" so if it is to be translated
03800 for use on a mini-computer system, such as a PDP10 or NOVA system, it
03900 could be broken down into reasonably sized separate units. I
04000 believe that I could make all this work on a mini-system with 32000
04100 words of memory and a good graphics display system. About two years
04200 ago my cost estimate was around $150,000 for a complete system which
04300 would allow a certain amount of time-sharing. At this date I think
04400 the system could be put together for under $100,000. Of course this
04500 is still a lot of money, but perhaps a group of publishers could
04600 finance it.
04700
04800 I hope this information will be of use to you.
04900
07700 .end
07800 .group skip 1
07900 .center
08000 Very sincerely,
08100
08200
08300
08400 Leland Smith