mGBPDUVlIlmGBPD k 15 I JWVU GBPD 15m lf lBV lH lDlBlH lH W GBPiJfi f WGBPiJlB l1GBPilWfi IBrn IYUml1mrn clJl UJ 1J0 0 TEACHING COMPUTER SCIENCE TO Li GUISTS U o THE STORY OF MOSES U j ' 'i I o o o o o o o oo o 5 6 TRAFFIC ANALYSIS OF THE FUTURE eU oo o oo 10 HOW ARE YOUR STAMINA U Sydney FaUba9 i 11 NSA-CROSTIC NO 26 eU I 12 DATA STANDARDS WITHOUT TEETH eU ooo oo o 14 MR PATTIE REPLIES eU Mark T Pattie Jr 15 LETTERS TO THE EDITOR eU o ooo o oo oo o o o o ooo 16 THE BALTIC ENCODERS eU o oooo oo o o o Anthony Reiskis o o ooo 18 THE NAVAJO CODE TALKERS eU o o oo o o o o oo o 20 I I r eLASSlflEB BY NSA eSSM 123 2 9E6LASSIFY eN 1 JUNE icc Declassified and Approved for Release by NSA on '10-'1 2- 20'1 2 pursuant to E O '135 26 vl DR Case # 54778 DOCID 4019666 SHCRET Published Monthly by PI Techniques and Standards for the Personnel of Operations VOL VI No 6 JUNE 1979 PUBLISHER WILLIAM LUTWINIAK BOARD OF EDITORS Editor-in-Chief David H Williams 39575 COllection Cryptanalysis I 55s 1 1 4Q02 Cryptolinguistics oo 1 1 Information Science Language o 1 1 5Q lS 1 30345 81615 SOg4 Machine Support 1 Mathematics r -- 1 IC8S18s Special Research Vera R Filby 71195 Traffic Analysis Don Taurone 35735 Production Manager Harry Goff 52365 For individual subscriptions send name and organizational designator to CRYPTOLOG PI SECRB'f ' ' P L 86-36 DOCID 4019666 UNCLASSIFIED TEACHING COMPUTER SCIENCE ELCITRA ERITNE SIHT UPMOC YB TES SAW TO LINGUISTS byl e TES ooo UPMOCYBTES Consider the plight of the NSA linguist He entered his chosen field to ling i e to work with foreign language material to translate to transcribe or perhaps to analyze and report the significance of large amounts of text He chose to deal with the fuzzy world of ambiguous meanings of convoluted and unpredictable rules of grammar and with the imprecision inherent in the transference of an idea from one language into another He never cared much for the picayune rigor or the grubby technical details of engineering or the physical sciences - they just weren' t appealing to him He felt at home in that imprecise world of meaning that is so foreign to most Americans And then he came to NSA Here the linguist must deal on a daily basis with computers As if it weren't bad enough that the phone company and his insurance agent used the darn things In fact not only does he have to deal with computers he has to actually on them More often than not they weu r SPeNT THREE PAYS U vM' PtlTTIN6 II New PR06RRM oN 0 ell r rAPe Ai'll' NOW m lI ' 'r r CRN'T 6er mE rr I COHPf Tel TO BVTTON fK ePr 1 iJ3fr'r' MJrS l I P L MOCYBUPMOCYBUPMOCYB Ip16 ITl The worst experience of all awaited the rare brave linguist who got involved in the design of a new computer system for his office The project development people seemed to be a special breed of programmers whose incomprehensibility was matched only by their desire to document in a level of detail that baffled the minds of ordinary folk Even though it is considered almost axiomatic that projects which don't intimately involve the proposed end-users from the very beginning are doomed to failure the linguist finds participation in planning extremely difficult because of the computer-ese language barrier and because of the lack of understanding of language work by others Once someone even asked a linguist on such a planning team if he really needed all 32 letters of the Cyrillic alphabet and couldn't he get along with just 2S or so because of computer limitations I 'IIf1 Ilf1 lJ lJf IIIlilm Ilfqfl'llll - 'IlIlIfI Jllrilif Jl JlI J AAR G6 I ' '2 provide his daily material for translation and store the older material God forbid that he should have to actually enter his translation or transcription into them for he has seen the words ENTIRE FILE DELETED on more than one occasion after spending an entire day laboriously entering his work keystroke by keystroke But even worse than the computers themselves are their keepers programmers - people who really have no comprehension of language work and who are always muttering about saving bits or something equally obscure when all you really wanted to know was why you Credits a Doonesbury cartoons reproduced with permission Copyright 1972 G 8 Trudeau distributed by Universal Press Syndicate CANDE word b This article was prepared using the processing system and a text composition system being designed and implemented byl IPl6 a114thefinal output of the article was done on the SEACO 171l1l CRT phototypeseuerin SJ 877'' ' couldn't get the machine to print out your daily take separated in an ever so slightly different way Because of this culture shock in going from the language world to the so-called electronic office and because of the tremendous improvements that are possible when linguists i l l included in the planning for the computer support for their work the idea of an introductory and terminal course for linguists in computer applications to language processing was born Such a course was developed and subsequently tested on two groups of linguists The results of those experiences offer many interesting revelations about the nature of the linguistic point of view ys that of computer science Before some of these experiences can be detailed a brief explanation of the newly developed course is needed I For a similar account of a linguist's first experience with computers see Robert Wachal's article Humanities and Computers A Personal View North America Reyiew Spring 1971 June 79 CRYPTOLOG Page 1 UNCLASSIFIED P L 86-36 86-36 - -DOCID UNCLASSIFIED The course CL-200 Linguistic Applications of Computers is both an introduction to computer science in general and to NSA-specific language projects in particular The students experience the art of computer programming by learning enough about two quite different programming languages to write one small program in each language In addition they take tours of two computer operations areas in order to see some real machines in the flesh The work in computational linguistics that is discussed includes the Agency's efforts in computer lexicography computer scripting speech processing and automatic degarbling as well as talks on the academic fields of machine translation of natural languages and artificial intelligence The last portion of the course explains the Agency's project management system and the ways in which the proposed end-user can influence a new project to insure its success A detailed outline of the course can be found at the end of this article Without a doubt the most difficult tasks for the students were the programming assignments The reasons for this were not completely clear and often varied among the individual students although two problems were shared by all The first of these had to do with the choice of names for the variables in the programs In high school algebra for example one usually prefaces a discussion of a problem with an explanation like Let x be the number of apples that John bought This is often not easily done in computer science and even if it is it is not sufficient for a large complex computer program as one quickly forgets what x was supposed to represent or even what John was trying to do Most professional programmers tend to choose names for the variables in a program that are at least somewhat suggestive of the meaning those variables have in that program Hence a variable which denotes the position of a certain keyword within a section of running text might be named' KEYWORD_OFFSET' or 'SUBSTRING_POSITION' if one were programming in PUI a language which allows very long descriptive names or 'STRPOS' if one were using a more restrictive inflexible language like FORTRAN which limits the lengths of names to six letters Yet these expressions have no meaning to the computer It merely stores the names in a table for future reference and sets aside a certain amount of computer memory to hold the values associated with those names 2 Variable names are unlike for example programming language reserved words like' READ' 'DECLARE' 'FORMAT' 'PROCEDURE' etc which have fixed and definite meanings to the computer meanings which are reasonably suggested to English speakers by these particular words Yet the distinction between these two types of words was difficult for almost everyone to grasp This led one student to attempt to do a frequency count by just listing' the words 'NUM_OF_ONES' 'NUM_OF_TWOS' etc since he reasoned that the computer understood the English words 'READ' and 'END' so therefore it ought to also be able to understand something like 'NUM_OF_EIGHTS' It wasn't until all variable names in the course lectures and examples were changed to words that clearly had nothing to do with the semantics of the particular example e g variables that held the frequency counts for certain characters had names like' LION' and 'BEAR' as opposed to 'NUM_OF_A_S' and individual lines in the program were given labels like 'COW' 'DOG' etc that the students really caught on While such a practice is at best poor for a professional programmer it was almost mandatory for the linguist who would have read too much into the choice of a name otherwise I really knew that this notion had been mastered when one of my students presented me with the frequency-count program abstracted here KAZOE PROCEDURE OPTIONS MAIN DECLARE ICHI NI SAN FIXED BIN 15 O HAJIME READ FILE SYSIN INTO TEXT IF SUBSTR TEXT l 3 'END' THEN GO TO OWARI MODORI IF SUBSTR TEXT I l '3' THEN SAN SAN I GO TO HAJIME END KAZOE 2 The fact that the computer does not understand English can make it rather tolerant of the inadequacies of some programmers In a reasonably large program I wrote a few years ago that dealt with dictionary retrievals I was quite proud of the clear and descriptive names I had chosen and the fact that they made the program so much easier to understand It wasn't until the program was completely finished that I found out that there was something not quite right about some of the names I had chosen names like 'NUMBER_OF_RETREIVALS' and 'RETREIVAL T1ME' Since my misspellings were at least consistent they were perfectly 'understandable' to the machine June 79 CRYPTOLOG Page 2 UNCLASSIFIED DOCID 4019666 UNCLASSIFIED This program uses quite descriptive names but in Japanese Since clearly the computer could not read Japanese this student showed that he realized what the descriptive names were really for the people who will read the program not the computer 3 j I 1 I The other problem in programming that was shared by all linguists was the difficulty in the reduction of large difficult tasks to a number of small easily accomplished actions This type of problem-reduction approach to problem-solving is that often used in high school geometry If I can just prove that triangle ABC is congruent to triangle DBF then I can do the whole problem and in symbolic integration Let's try integration by parts and see if we can get two integrals that are easier to solve 4 This division of a complex problem into smaller more rudimentary pieces is imperative if one is to write a program to solve that problem One must reduce the given problem to small primitive problems that can be easily programmed often in a single line One student was so overwhelmed by the complexity of programming one problem and accounting for all the various possibilities that the first step of his solution was The programmer manually checks through the input and removes the following special cases Other students just couldn't get enough of a handle on a small part of the problem to even begin but when helped in that first step were able to complete it without too much difficulty It is possible that this is just a difficulty encountered by every beginning programmer though I don't remember this happening at all among the students in my first programming class a class made up entirely of students in engineering and the physical sciences It may be that this in fact points to a basic difference in the approach of a linguist They perhaps do not analytically dissect their work as the engineer must often do but rather approach the whole problem intact perhaps slightly more attentive to one particular portion at anyone time but nevertheless keeping the entire problem in focus This is sort of a gestalt approach as opposed to an analytic one but an approach that may be required for language work This idea that linguists may have a different way of attacking problems is also supported by their reactions to the two programming languages that are taught in the class Without exception the students felt much more 'comfortable with SNOBOL a language that is strange to many professional programmers SNOBOL is unlike the more standard programming languages like FORTRAN PLlI ALGOL etc and in fact was designed for the processing of text in complex ways SNOBOL allows one to look for complex patterns in running text and to easily manipUlate the text portions it finds For example the 3 It is not at all so clear that a computer could not 'read' Japanese To the computer there is no real difference between the string of letters' BEGIN' and the string' HAJIME' It just so happens that most computers are built to recognize only one of these words u denotina the meaning 'start' There would be no real problem in theofY in programming the computer in German Japanese or even Arabic There m however some political problems For more details see Theodor Sterling's article 'Computers in Developilll Nations A Cautionary Tale' CommBDicatioDs of tbe ACM Volume 2' Number 12 December 1977 pp 971-972 Nilsson Nils 1 Problem-SohiDg Methods iD Artificial latenlaeace McGraW-Hill 1974 chapter 4 following one SNOBOL line formats arbitrarilary spaced English text and outputs it on the line printer 5 lOOPA TEXT POS O SPAN ' ' NUll SW I ARB' ' SW SPAN ' ' I lEN 60 OUTPUT NUll SW ' ' ARBNO BREAK ' ' ' OUTPUT ARBNO NOTANY ' ' SW POS 60 I LEN I ARB OUTPUT NULL SW RPOS O SW S lOOPA The linguists seemed to find SNOBOL much closer to their way of thinking than the more standard i e to most programmers programming language PLII Since it has been hypothesized that one's natural language limits or restricts one's thinking patterns 6 it should also be equally possible that given a choice between two artificial languages one would feel more comfortable using the one that least disturbed his natural thinking patterns To the engineer this might be one of the so-called algorithmic languages lik e PLII or ALGOL to the linguist it might be a totally different sort of language like SNOBOL But by far the most interesting observation about the linguists was the imagination and enthusiasm with which they approached a foreign field One principle which was expounded upon often was that of Oars Law a property of some programming languages A programming language is said to satisfy Day's Law if anything that a programmer would ever reasonably want to do can be done easily and directly in that language PLII is an example of a language that satisfies Day's Law and FORTRAN is an example of a language that clearly does not The students were told that PUI satisfied Day's Law and many were able to discover some actual PLII syntax by guessing One student theorized that PLII ought to have a way to assign the same value to two different variables simultaneously and guessed the syntax I J ' Correct 1 and another student felt that there must be a way to assign an initial value to a variable when it is declared and guessed the correct syntax of the PLII keyword INITIAL Unfortunately there were also a number of incorrect guesses though not incorrect applications of Day's Law most notably when one student decided that array references really shouldn't have to be numbers but ought to be arbitrary character strings 8 This belief led the student to try the syntax 5 This SNOBOL program was written by R61 in response to a challange by another SNOBOL programmer 6 This is the Sapir-Whorf hypothesis in linguistics and an entirely artificial 'natural' language has been designed to test it'$ validity For more information concerning this hypothesis and this new language see James C Brown's LOGLANI A Logical LaDgBage 3rd edition The Loglan Institute 1975 7 Days' s Law is named in honor of its proposer a former Agency employee 8 This is approximately true in PUI if one considers P JI structures heterogeneous arrays with non-numeric subscripts and is definitely true in SNOBOL with the SNOBOL table see T W Prall's ProgrlllmiDg LaDgBlges DesigD aDd ImplemeDtatioD PrenticeHall 1975 pp 64-67 June 79 CRYPTOLOG Page 3 UNCLASSIFIED I I P L 86-36 - DOCID 4019666 UNCLASSIFIED DECLARE COUNT ' ' '9' FIXED Course Outline to build an array which was indexed with the literals' 0' I' '9' Unfortunately such a statement has nothing lik e that effect Linguistic Applications of Computers CL-2 The enthusiasm and optimism of the students during the entire course and in the programming assignments in particular was indefatigable Two particular cases immediately come to mind One student's first attempt at one of the programming assignments resulted in a list of diagnostic errors that was about ten times as long as the total length of the program being tested Yet the student was only slightly daunted and completed the assignment with only a few additional runs Another student upon discovery of a small syntactic error in his program was then so confident of his work that he was then willing to wager that the program would definitely be completely successful on the very next run Much to the student's dismay it took man y more runs before the program accomplished anything even close to the required task Had I collected on all our 5' bets I would now have a small drawerful of nick els A INTRODUCTION TO COMPUTER SCIENCE Hardware Software PUI SNOBOL NSA Computer Complexes Computer Operations Current and Future Trends in Programming Languages B COMPUTERS AND TRANSLATION Computer Lexicography Machine Translation and MachineAided Translation Computer Scripting Speech Processing and Computer Aids to Transcription Artificial Intelligence C COMPUTERS AND CRYPTOLINGUISTICS Sorting KWIC Indices Automatic Degarbling Schemes Mathematical Modeling of Languages PTAH Dj LINGUISTIC PARTICIPATION IN THE PLANNING OF A FUTURE COMPUTER SYSTEM Required Agency documentation SCP Series Linguistic Input Case Study Project THISTLEDOWN Proposal Evaluation SOLUTION TO NSA-CROSTIC NO 25 CRYPTOLOG May 1979 by D H W Once some illiterate SOUl concludes that type means type of the step to this type thing is immediate Since it is well known that no error is stupid or vulgar enough to guarantee that it will not become respectable we refrain from rending our garments But we submit that at this period English this type writing is not appropriate to this sort journal Editorial NSA Technical Journal January 1958 by Sydney Fairbanks reprinted in CRYPTOLOG April 1979 International road signs are being seen more and more frequently in the United states For the benefit of readers who may not be familiar with them here are a few examp les SIGN LANGUAGE Reprinted from Q R L Umbrella Trouble When Windy You are on a runwa y You Arc NOli Entering Johnnie Applc5ccd COl6lrry JllireracC' nadves hereaboufs Snake Crossing Barber W tch out ror returnins boomerangs @OOO Caution Drunken Drivers Ahead SbJnt Drivers Practidne - Bookie Joint Sho ' B ce of purse Snatcher3r Only One Out of Six ill Make II Unfriendly Spca hrowina Watch Your NatlYU Hereabouts Step II June 79 CRYPTOLOG Page 4 UNCLASSIFIED DOCID 4019666 Gqr CONFID BNTIl b tnry Itf 1Illsrs I OJ Ul IG51 - r '1 tl P L 86-36 EO 1 4 c P L 86-36 f you should ever want to know when some Iblimportant event such as an international conference or meeting is to take place il'SES may well be able to help you Not the Old Testament prophet but the MUlti-national On-Line Scheduled Event Survey This data U In that form the MOSES extract was too base which contains information on upcoming events gleaned from a variety of open and unwieldy to use If anyone wanted to find out classified sources is an excellent example when a given meeting was going to start he of interaction between T12 Information Sermight have to read through an inch-thick stack vices and analytic elements producing a of paper before finding his answer usually on data base to meet operational needs It rethe last page Nevertheless the MOSES inforsides on TIPS Technical Information Procesmation was still useful so we asked for an sing System and can be accessed at NSA extract we wanted it in an easy-to-read through RYE outstations and at other agencies format sorted by host country and by the through the Community On-line Intelligence beginning date of the events System COINS The final format of the extract was U Moses at Work The folks in Tl22l quite usable even if not ideally so It was gather information on future events and put usable because it allowed people to look for it into the data base as they find it assigntheir desired information quickly After all ing a one-up number to each item The way we the main reason an analyst needs to know when used to get the MOSES information was to have a meeting is to start is to permit him to dean analyst submit queries through a RYE outcide whether or not to publish a piece of station at the end of each month Those information which has just become available queries would then produce listings of selectrelating to some event For instance if a ed items in item number order But these listings were hard to use because they were not organized either by country or by event - t h e 1 n s - t r 'u c - t 1 o n - s ' 'a r - e --lh 1 s - t - o r ' 'y' ' ' n - o lt 1 n lt - e 11 1 1 --To circumvent that problem the listings were gence and will not normally be published reduced by hand to a list of events arranged The event information usually availby beginning dates The list was then dupliable to the person making the publish-or-not cated and distributed throughout the division decision consists of its name and where it is Developing those chronological lists by to be held With those two items he can hand was quite a time-consuming task This then quickly thumb through the revised vercame to a halt after a reorganization shuffled sion of the extract to the proper country away the capability to compile them As a reand then scan down the first word of the sult the original computer listings were denarrative descriptions of the events and livered directly to the analysts who had became find the needed date used to getting the nice chronological listings U The problem then exposed was that all The first time the raw computer listings were too often the event we were interested in was brought into our area we found out how Phanot even on the list the format was usable raoh's army felt that day in the Red Sea but the data base was incomplete To remedy MOSES had just closed a sea of paper over this a series of tetragraphs was devised our heads I I June 79 CRYPTOLOG Page 5 CONFI9 BNTIA b DOCID 4019666 CONrU EN'fIAL to be used in the TAG Topics and Area Guide line of every piece of product discussing the date of any future meeting conference or similar event Now whenever an analyst gets advance' notice of an upcoming event the information ran be pread through u e of one of the -SKD U The MOSES folks can then use the SOLIS system and pull on the -SKD tetragraphs to get easy rapid access to that advance notice Once MOSES knows about it anyone in the intelligence community can easily find out Obviously one of the most important sources of MOSES inputs is the translators and the checkers who keep -SKD in mind as they write or review TAG lines U After a few months experience with the new easy-to-read printouts user analysts decided that calling T122l on extension 3258s was even better for a very good reason up-to-the-minute information was available rather than what had been on hand back on the first of the month U It's surprising how few linguists and analysts even know about MOSES Perhaps this article will help spread the word around other useful file also maintained I _---- ------- ------ ---- excellent source of expansions of acronyms and abbreviations P L June 79 CRYPTOLOG Page 6 CONFIBI3N'I'IAL EO 1 4 c P L 86-36 86-36 - DOCID EO 1 4 e 4019666 CONFIBRNtplAL F L o June 79 CRYPTOLOG Page 7 CONFIBEN IAL 86-36 DOCID 4019666 CONFIDENTIAL June 79 CR YPTOLOG Page 8 _ _ EO 1 4 c L 86-36 P'--' DOCID 4019666 EO 1 4 c P L 86-36 CONFIBBN'fIAL TO START A SUBSCRIPTION TO CRYPTOLOG OR TO CHANGE THE ADDRESS ON YOUR PRESENT ONE CALL 3957s June 79 CRYPTOLOG Page 9 - --------- -----_ - _ _ - - - - - - - - - - DOCID 4019666 SECRET P L 86-36 TRAFFIC ANALYSIS OF THE FUTURE-CU m L - lWGBP hat did you do in the war sir I was a traffic analyst Oh how _interesting What kind of cars did you analyze Various Agency analytic disciplines are on he decline Take a look at the table in the January 1978 CRYPTOLOG article entitled The Changing -Face of NSA It shows a sharp decline in numbers of people assigned in certain job fields over the previous five years Of the Agency's analytic fields traffic analysis showed the largest percentage drop Then take a look at the latest Quarterly Management Report traffic analysis has been reduced still further Where are we going What is happening U UI I'm going tostoPl1eI e and make my prediction a double-barrelled Qne There llJiZZ aome a time llJhen aU o what llJe nOllJ aaZZ traffia analysis llJiU be aentralized in a single organization in DDO 01' llJhatever it llJill then be aaZZed This step llJiZl be taken to bring together and husband the fellJ remaining TA skilled people and to develop under a single line management authority the unified teahniaal data system and software system to support TA tasks llJorldWide UI I don't know when this will happen ButI'm pretty sure that it will happen U Of course I was wrong about Dewey in '48 June 79 CRYPTOLOG Page 10 SBEURRJ3T IIANB ' B ViA eSPlHN'f SIlMI I B 'S SN 'Y EO 1 4 c P L 86-36 DOCID 4019666 UNCLASSIFIED I I I i How Are Your Stamina Sydney Fairbanks Reprinted from The NSA Technical Journal of January 1959 ne of our readers stung by some peculiarly noxious idiot'sidiom that had crept into an official communication called up the other day to ask almost tearfully if we thought he could volunteer to write all the memos put out in the Agency No we told him the suggestion would probably not be well received but we too we confessed had had daydreams of a similar czardom lightened in our own case by the imposition of a scale of penalties Omissions of the definite article subject memorandum is reprinted in referenced document would call merely for confinement to barracks references to the overall picture would involve a substantial fine statements as to the capability of the facility to become operational transmission-wise on a continuing basis necessitate a painless beheading and naturally anyone writing the reason why this is so is because of the fact that will be hanged drawn and quartered Beyond these we progress to actual errors We hope you shall like he did this is a new one and which oil the bearings such as we did yesterday and of course our friends COIlllla however this type thing and they are writing and notify the contractor We are sorry but we cannot tell you the penalties for these There are also the people who write this phenomena is noted in more than one media and the discoverer is worthy of several kudos -but we are becoming a common scold Actually the matter of foreign plurals is not quite so simple as the purists would have us believe Data which started life as a proper little plural is rapidly becoming a collective singular and anyone adopting a hoI1er-than-thou attitude about it should be asked how his stamina are this morning Back formations of singulars are even more confused A tactic or a statistic has no more right to exist than a mathematic or a calisthenic but it does A man joining the commandos should no more become a commando than a man joining the troops becomes a troop but he has And sO we suppose from a purely scientific point of view one has to admit the possibility that a time may come when something called a kudo can exist naked and unashamed but not we hope until we are dust before the doors of friends or radioactive matter a-blowing down the night We too can be scientific on occasion A short while ago we lamented the absence of a good portmanteau word of spontaneous rather than deliberate' generation In accordance with Somebody's Law which says that as soon as you say there is no such animal a perfect specimen trots around the corner we came for the first time on the word meld embedded in a very distinguished matrix Not the verb used in pinochle which is of course German melden to declare but one meaning something like amalgamate Webster's International knew it not and we were tempted to throw it out but we have a weakness for monosyllables To make a long story short Webster's New World Dictionary of the American Language gives meld v t and v i merging of melt and weld to blend merge unite Sinking the classicist in the collector we left it in the copy where we found it and tiptoed away But this does not mean that we are prepared to accept irregardless O June 79 CRYPTOLOG Page 11 UNCLASSIFIED I l DOClD 4019666 UNCLASSIFIED NSA -C fi N 26 8g D H W WORDS DEFINITIONS A Site of baseball's Hall of Fame B Dog star comp C Impending D Act of investiture of a monarch E NY town center of winemaking activities F Uncultivated uncultivated person G Diana Rigg's role in The Avengeps 2 wds H my dear Watson Holmes I reply when asked how one should refer to a saffron-colored devilfish 3 wds I Person shunned for social or moral reasons J Any of more than 100 fundamental substances 201 sr l7S lSI 253 243 20S K Ringing of bells L Why at the the skin of much as one calfskin 5 Sioux bride sale the maiden on the African river beast cost as 140 104 25S- 146 ISO 36 -2- 131 S5 126 122 on a bearskin plus on on a wds followed by Word M 9I 10 244 f 204 156 215 163 255 99 30 M See Word L 11 wds N o 242 13 66 3T 44 264 221 llS 159 214 -8- 246 Room Volunteer State P Just underway racetrack jargon 3 wds Q Candy made by boiling sugar and butter June 79 CRYPTOLOG Page 12 UNCLASSIFIED DocrD 4019666 UNCLASSIFIED R Women of a Moslem household S Overture by Beethoven op 84 1810 r T Compass point U Tanzanian farewell to Miss Day 2 wds V Resourcefulness fantasy W Filling to the brim with a liquid 2 wds X Assault crime Brit sp Y Uppermost part of a building I ' t ' ' _ noll _ 69 S e 51' '54 550' 71 U 11 0 72 E DB B LJ B 1 _ 103 D 34 P 50 v e 51 K e 90 56 B 73 x e 74 14 1 L II e 121 t l I L l T 47 48 L 4914 57 Q 58 14 590 60 61 L 62 G 63 P 64 0 65 A 6614 67 V 68B 75 U 76 A 78 0 79 Q 80 G 81 14 82 U 83 B 84 V 85 L e 77 P e e e e e t4 N 11 l l lb 1 O K 137 E11 N 1 M 14U I 141 Xl14 l U e 66 H 167 M 68 P 182 F 183 M 84 B 52 K 15314154 V 155 C e e 169 G 170M 111 E 185 Y 186 II 98 II 199 Q 00 G 201 J 202 X 203 N e e 0 156 L 157 Y I tl e e X e 143 v 144 33 A 2411 l1249 U ' II II 34 M 235 E 236 It 237 R 238 H til 11 t' e 11 1 l' l X e 114 L lUU' 172 H 173 R 174 M 175 C 187 A 188 II 189 C 190 M 191 T a4 t'lll L Ub M e 113U 14 239 Q 240 II 50 V f l ll G 252 H 2 13 J 2 14 II 255 L _ e 131 L IIUI 1 IW l e 13 l 1 Q 14b L 147 II 148 R 149 e e e e e 176 II 177 14 178 J 192 0 193 S 117 F 133 N He e 241 E 242 M 243 J 228 M 229 N e 1179 D 180 L 181 194 C 195 H 196 II 197 U 204 L 205 M 206 U 207 K 20S J 209 C 210 A 211 H 212 0 e e 86 H 11 18 R 159 160 D _'1161 1162 E 163 Lll64 U 165 A 215 L 216 A 17 X 1118 II 219 0 220 N 221 M 22 V 223 F 224 II 225 H 226 B 227 G e 51 J 45 e 46A SO 4 P 43 V 4I II 041 1105 L 106-14 107 T e l O S M l09 ElllO l e l a va l 1 a t' 18 M I' I 1 lU l o _ e e 244 L 245 J 213 C 214 M e 230 M 231 E 232 C e 246 M 247 E 256 N 1257 A 12 18 L 259 II 1260 B lOl t' lO H 263 E 264 M _ o June 79 CRYPTOLOG Page 13 UNCLASSIFIED o dhw DOCID 4019666 UNCLASSIFIED eo T lEElt --- '-----_I tis clear from Ma r1cPattie's and I I al'ticles that the data standards program has encountered major problems Mr Pattie cites several important areas where the implementation of standards has been less than successful and he says that the NSA Data Standards Center is just scratching the surface Mr Murchison in his final paragraph says that we sometimes feel that our job is impossible The root of the problem is not in the actual development of standards but in their implementation Mr Pattie blames the general user at one point appealing for cooperation and at another point ' thundering Until we are authorized to deny computer use to those who refuse to accept Agency standards we can have little effect is I P L 86-36 To illustrate my point let us assume that NDSC is given teeth and can actually deny a user access to the computer if he does not adopt standards To start with the idea of enforcing standards only on new applications is doomed to failure because there are few really new applications Everything we do is built on what we have done before The new is tied inextricably to the old and must therefore conform to the old conventions To be successful enforcement will have to address the whole world at once The first problem is to discover who is using standards and who is not There is no easy way of doing this and the whole plan would probably founder at this point But let us assume that a lengthy study is able to ferret out the facts It would undoubtedly reveal that the whole Agency even NDSC re I e t lo e c t J i gtliJtyoGBPincoIIlPl#t c t J ndardization Ibacksmhimmup Usually a sponsor cries 'unacceptable' just because he does not want to go to the trouble of re-programming This analysis of the cause of the problem is totally wrong Most people suffer to a greater or lesser degree from lack of standardization and would do anything reasonable to avoid this But standards are only justified if they bring tangible benefits If the pain caused by implementation exceeds a reasonable threshold then any benefit to the user will be cancelled out and we will be reduced to the untenable position of trying to enforce standards for their own sake The rejection of operational necessity is the rejection of reality NSA Data Standards Center by Mark T Pattie Jr CRYFTOLOG November 1978 and Comment on 'Data Standards Without Tears ' byl I CRYFTOLOG March 1979 Letter to the Editor CRYPTOLOG October 1978 P L 86-36 The next problem is to determine exactly What we are going to enforre A Gropp has at least 1000 data elements L the NDSC centralized file of data elements and data field definitions lists only 200 standards Shall we re-program to cater for the 200 and then every few months when a new standard comes along re-program allover again Or shall we suspend all operations until the other 800 are standardiied In the latter case calculating from the fact that it took seven years to develop 200 standards we will be waiting until the year 2006 Both Mr Murchison and Mr Pattie are therefore wrong in suggesting that the general user is the villain There are facts of life which make enforcement impossible in a large complex organization irrespective of the attitude of the users I am personally in favor of standardization yet when I had the job of establishing standardsfor a project I was forced by operational necessity to be satisfied with partial standardization If a standard exists I will adopt it But if it does not exist what do I do How much time can I spend tramping around trying to find out what June 79 CRYPTOLOG Page 14 UNCLASSIFIED P L 86-36 1 UNCLASSIFIED DOCID 4019666 other people and projects are using Inevitably I invent my own names abbreviations and codes just to get my job done before the deadline If we cannot blame the general user can L j s D d e tn st d d' -S - i -S---f inevitably a long and rigorous process NDSC is staffed by able people who are dedi cated to their work to an extent rarely found No one could do better The truth of the matter is tha once non-standardization has reached a certain extent in a complex organization it is not possible to enforce standardization without causing unacceptable damage Non-standardization is self-perpetuating During the time it takes to standardize one data element a dozen new non-standard uses can evolve so that the standardization process never catches up It is a vicious circle which can be expressed thus Laak of stan- dardization aauses an inareased Zaak of standardization Once the truths of the situation are recognized then we can open our minds to reasonable alternatives I firmly believe that the Data Element Dictionary Directory is the best alternative we have My personal experience of a DED D is that it mitigates the adverse effects of existing non-standardization by tying different naming conventions into a single data definition But it could do much more than this If all applications are suitably described in the Directories then everyone will have easy access to the conventions used by other people I Ide cribes it very aptly The Directories will show what the current usage of data fields is along a wide spectrum of different Agency applications Exposure to this usage will gradually lead us towards the necessary standardization If the Directories tell me that a certain data element has six different names and no standard then I will surely adopt one of the six thus preventing a seventh from appearing The process will not of itself bring about standardization but it will certainly slow down the everincreasing tendency to non-standardj zation It will reverse the iCiolisCirCle I Itook me to task for using the word magic in my first article and indeed I ought to withdraw it I am tempted to let it remain because it daes help to express the extent of the breakthrough in a very difficult area NDSC will still have the labor of establishing standards No machine can do that But the DED D will fill the gap where there are no standards and will act as a positive force preventing the proliferation of non-standard uses as well as providing NDSC with a mine of information as input into their processes And all this without any additional sweat tailor tears on behalf of the NDSC There is of course labor involved in the acquisition and loading of data into the Directories but this is not NDSC's work It is spread over the whole work force and does not bear very hard on any particular indi yidual or office I have sat on a terminal loading such data and can vouch for the fact that itisJlQtyeryexcitiTlg u t n ither is it very hard Large complex' P L 86-36 flIes can be descrIbed and loaded within a few days not much compared to the length of time it takes to standardize one data element Other things must be done such as making the DED D well-known and easily available to everyone and encouraging its use by providing a variety of services It is all well within the realm of the possible The fact is that the DED D is a new tool which benefits everyone without any additional effort documentation has to be done with or without it and in fact your data elements may be loaded by someone else with overlapping interests It is not even necessary for all data from all areas to be loaded before the benefits are felt There is no enforcement no need for teeth There are no adverse side effects Nobody can lose least of all NDSC I vote we give it a try Mr Pattie oo plloo J s uppos e t h at goi fonthethrowing Editor would it we could barbs allow at one another for months or even years but there is l'eallYTlothing to be gained from that I would propoSe that we bring a hal t to the exchange and I would like to apologize i f I have hurt any feelings or stepped on any toes P L 86-36 _ A l' What really bot 1 ersme is not the name calling oT1 heobfuscation but the apparent failure 6f Agency personnel to understand what we have been trying to say about our work in the first place I fjust happened to put on paper some of the t ings we have been hearing all along In spite of our letters and articles on the subject people still do not seem to understand what the role of the NDSC is We do not and I repeat not impose data standards on anyone When higher authorities DoD NBS and the like agree to a standard then NSA has to abide by that or request an exemption We help to coordinate action on those terms Our main job with the Agency is to work with those who want to propose a cryptologic standard coordinate the proposal with others and get Agency agreement to call it a standard June 79 CRYPTOLOG Page 15 UNCLASSIFIED - - ----- - - -----_ DOCID 4019666 UNCLASSIFIED It often takes a long time to do this partly tempt to incorporate them into a single one because the people we deal with are involved for all of DDO to use It appears to be a with data standards only as an add-on to repeat of the problems encountered by users their normal duties We appreciate the supof COINS files By not standardizing we port we do get but we recognize that data force users to make costly searches of all standards may not have as high a priority in possible files directories to make sure their lives as we would like nothing is overlooked We do not expect to standardize everyPerhaps it might be useful to close thing in the Agency nor do we think that with an illustration of the difficulty we necessary Where programs have limited have in the area of data standardization applicability and use it is enough that even when we are talking about something they work When the applications or uses that has long been standardized This past are more widespread standardization becomes March we received two annexes to a USSID a worthy goal for comment in each there is a line with a non-standard date DDMMY instead of the It is costly to standardize and it is standard YYMMD which was implemented in 1971 fairly easy to cite such costs as a reason How could to keep going our someone eight individual ways years later put I wonder if anysomething like that one has ever given any thought to into a program Why wasn't this noticed what it costs when by someone else long we do not stanbefore it got to the dardize I ACREE NOTHING 10RE ON DATA NDSC I can only We have nevSTANDARDS OUR I 'G THE REST OF THE assume that no matter YEAR UNLESS IT TURNS OUT THAT THEY er spoken out HELP CURE THE CO 1 -ION COLD how many data stanagainst the use of dards exist there Data Dictionaries will always be peoDirectories Our ple who will want to own PANDOWDY file go their own ways is a small but as before to be and we will continue useful example of one Our main concern is accused of getting in the way of operational that these dictionaries seem to be prolifernecessity ating when there ought to be more of an at- Letters wthe Ediwr In the May issue of CRYP1 LOG we asked if anyone could identifY a sixth language of the Soviet Union which is not written in Cyrillic the other five being Armenian Georgian and the languages of the three Baltic republics To the Editor CRYPTOLOG I believe that the sixth language you are looking for is German There are almost a million Soviet citizens whose primary language is German They are the survivors and descendants of the people of the Volga German Republic who were forcibly resettled to remote parts of Soviet Central Asia in 1941 where they still live in village communities and cooperatives speaking their own language running their own schools and despite odds maintaining a cultural integrity Suspected by Stalin of being potential fifth columnists these people were segregated by sex age and family status and shipped off to the Asian boondocks They were not put into prison camps butwel'e simply set down on the steppes and semi-deserts and left to fend for themselves Those who survived did so through the assistance of the local tribesmen who taught them how to build sod shelters and surivive the first shocks of resettlement In 1964 these Germans were officially rehabilitated which restored to them certain rights of citizenship including the right to serve in the armed forces and to hold membership in the Communist Party With regard to the language used by these people in his book National Languages in the USSR Problems and Solutions the author M I Isaev states -that Soviet Germans use the same literary language that is employed in the German Democratic Republic and West Germany o I June 79 CRYPTOLOG Page 16 UNCLASSIFIED I P L 86-36 DOCID 4019666 SHCRE'f U Gradually we can change the location of particular books and the appropriate records U I would like to complain about the recent as use shows where the demand is Further changes in the library Basically what has new_books will be purchased for both colbeen done is to move all mathematics and engi- lections regardless of class if the need is neering vol es to FANX and leave all computer there and the budget and space permit We science entries at Ft Meade This inconveare already doing this for books in computer niences many people since there are people science interested in computer science at FANX and many mathematicians at Ft Meade At a small C h i-ef ' T I 2 1-cost in effort the library could at least have left some of the duplicate mathematical P L 86-36 items of which there are many at Ft Meade and sent some of the duplicate computer science works to FANX U It simply is inadequate to be told that AND FROM AN OVERSEAS READER ooo we at Ft Meade can always order from FANX and vice versa For example I often need an immediate answer to a question in say linear algebra or statistics The time lost To the Editor CRYPTOLOG by many people will certainly reduce the article Let's Not efficiency of the Agency's technical effort Lose Our TA Skills CRYPTOLOG arch 1979 EO 1 4 c U Even more helpful than merely splitting made an untimely appearance hereL__- --- _ P L 86-36 the duplicates would have been to try to The week before it arrived several of the cover the content of various topics at both people here due to return to NSA in the ear branches Thus if we have 25 works on dynamic future had attended reassimilation briefIngs programming none of which are duplicates it during' which the overstrength skil S problem surely would have been possible to divide was not only raised but was also gIven as a them such that the basics of the field could rationale for the reassignment of some to a be found at both locations For this option different career field Some of the moves of course the library would have needed tech- were in the out-of-TA direction It's easy nical advice As I have pointed out to the enough to fallout of touch when headquarters library's management a number of times there is an ocean away but it's even easier to be are mathematicians in many parts of the Agency confused by the apparent contradiction ready to lend a hand between the M3 view of TA as an overage SkIll and the view shared by land I am afraid that the library has taken that TA is a field with an the easiest to them route without fully P-i-m-p-e-n 'd 'i-n-g-s 'h-o-r- tagec i sis Perh ps some of considering the needs of the users the 50-ode traffic analysts now 1n excess could be used to head off the coming shortage The cotlcernabout the effect oftr Ll1sfers out of TA and the lackQfnew blood is shllred by many This feeling was expressed here recently by visiting managers who statedaneedfora a The Chie f 0 f the Library rep lies lytic talent to work in the api ly expandIng U I can understand the frustration felt by P L 86-36 world of mul tichannel commumcatlol1s the people in A who return to Ft Meade to dis U What all of this mayboil down to i the cover that many services they hoped to regain ever-present problem of individual h ld1ng have been moved to FANX to make room for them the title but not doing the analytIc Job and U Dividing a library collection is not a the apparent inability of management to cope simple task The card catalog show LtheToca- with this issue GOod traffic analysts are tion of books If the collection had been dihard to find and so may be talent in ot er vided asl Iproposed there would have overstrength skills the key may be qualIty been no way to know where anything was since vs quantity another problem that's always changing all records immediately would have with us If nothing else' _ r- been out of the question Under the system article may make some managers realize that we chose users can be informed of the locaskill balance by the numbers is not an end-totion of a particular class of books regardless all-your-troubles elixir for reducing the of what the catalog indicates Moreover in workforce and that talent returning f om recent years very few duplicates have been overseas should not be regarded as a magic purchased and since most requests are for ingredient for such a brew current books dividing duplicates would not have provided a great service Kathy Bjorklund F8313 To the Editor CRYPTOLOG I June 79 CRYPTOLOG Page 17 SHCRH'f' IhltNBhFJ JPYhIt eS HN't' enAW flhS SNhY DOCID 4019666 SI3CRI3T SPOKI3 In this article which originally appeared in Q R L Quarterly Review for Linguists for February 1974 the late Mr Reiskis postulated a P L 86-36 EO 1 4 c June 79 CRYPTOLOG Page 18 SI3CRI3T SPOKI3 DOCID 4019666 EO 1 4 Te Il CRIl T 8POKIl P L 86-36 Tnere 1S a V1aOle alternat1ve wn1cn WOUIC1 in keeping with the off-the-shelf approach and which would thereby be more U h more _----------------------' June 79 CRYPTOLOG Page 19 SECRE'f SPOKE P L 86-36 EO 1 4 c DOCI-D P 1 4019666- 86-36 EO 1 4 c 8ECRE'F SPOKR economical in dealing with the type of threat postulated here The following article appeared in The WaLL Street Jo L for December 1974 and was later reprinted in various NSA pUblications It seems appropriate to publish it once again as a postscript to Mr Reiskis' article THENNAVAJO CODE TALKERS ---' Navajos weren't the first Indians used by the U S military to confuse foreign enemies Choctaws transmitted orders by telephone for the Army infantry in World War I and early in World War II Comanches were employed in similar activity in the European combat zone But the Choctaws and Comanches conversed in their native tongues The Navajos on the other hand developed a special coded alphabet of 38 symbols plus an auxiliary vocabulary of 41 other terms It's been described by anthropologists Henry Dobyns and Robert Euler as absol ltely unbreakable --- ------- --- - - -- Skilled as the Japanese cryptographers ere it's doubtful whether they would have understood Navajo even if there had been no attempt to disguise it At that time it was virtually an unwritten language and even today few non-Navajos have succeeded in mastering its complex glottal sounds and vowel tones But rather than take a chance the Navajo code talkers improvised a system substituting clan names for military units the names of birds for airplanes and fish for ships plus a double alphabet when it was necessary to spell out proper names The idea originated with a Navajo-speaking white man Philip Johnston an engineer with the city of Los Angeles who was raised on the Navajo Reservation where his father had been a missionary During the first few months of the war he suggested his plan to a high ranking Marine Corps officer It was approved after five Navajos demonstrated its possibilities to Marine brass By April 1942 Marine Corps recruiters arrived at the reservation searching for Navajos who were physically fit as well as fluent in Navajo and English The first group of volunteers 29 youngsters from various boarding schools in Aritona and New Mexico were sent to boot camp at San Diego FFom there they were transferred to the Field Signal Battalion at Camp Pendleton then assigned to Marine combat divisions throughout the Pacific Philip Johnston joined the Marines in the fall of 1942 and was put in charge of the code talker training program Eventually some 320 Navajos served in combat under the program The code talkers served in many campaigns usually in two-man teams conversing by field telephone and walkie-talkie to call in air strikes and direct artillery bo bardment Marine Corps archives contain ringing praise for the Navajos from commanders in the field The idea for a formal association of code talkers grew out of the 1969 annual reunion of the Fourth Marine Division Association which honored several of the Navajos TWo years later the Navajo Tribal Museum the repository for Philip Johnston's papers and other code talker memorabilia sponsored a two-day reunion Now the Navajo Code Talkers Association numbers more than 100 members June 79 CRYPTOLOG Page 20 SI43EURRe'F SPOKe - - - ' -- '---_ _ _' ----_ _--- - _--_ ------- --------- DOCID 4019666 BEeBE SPOKE _________________ UNClASSIFIED C A A NEWS DID YOU KNOW FRANCIS X paRRINO It's always hard to say goodbye to an old friend Frank Porrino came to work at the Agency is 1955 after a tour with AFSS At the time of death he was Chief of Al and President of the CAA Quick caring good-humored professional -these are some of the words that he brings to mind If you ever worked near him you will remember his laugh and the infectious way it spread to those around him I CIA has a collection of over 20 000 motion picture films and videotapes in its Pictorial Services Branch Office of Central Reference This collection dates back to 1940 and includes films and tapes in several languages The collection ranges from foreign newsreels documentaries and feature pictures to videotape recordings of foreign and domestic television programming of intelligence interest The subject content is very broad covering the entire spectrum of political military economic biographic scientific technical cultural and goegraphic area i terests There are tapes and films of world leaders Soviet space flights world-wide political and military events military parades in Moscow and Havana to name but a few FUTURE CAA PROGRAM NOTES 27 June 1400 hours Room 2W087 I These films and tapes some of which may lABS Ockha m's Razor- A L - h 'i-l-o----1 coJitainin 1elligence information not received P sophical Approach to ProbleJIISolving in printed form are available for NSA analysts to borrow or to View at CIA Th Pictorial 7 August 0900 hours Friedman Auditorium Services Branch has the techniealcapability Vera Filby E4l The Impact of the Privacy to play back virtuallyanY ideofoI mat P L 86-36 Act on SIGINT Reporting _ _ F o r additiona linformation calli I I ICNSA liaison at CIA on 983-82l5s if you're at Ft Meade or 9983-82155 if you're at FANX CLUB AND ORGANIZATION OFFICERS-THE PAGES OF CRYPTOLOG ARE OPEN TO YOU FOR ANNOUNCEMENTS OF YOUR GROUP'S ACTIVITIES SEND YOUR SUBMISSIONS TO Pl CRYPTOLOG OR CALL 3957s LKLASSIFIED P L The publications appear to come from a rather eclectic bookshelf The subjects range from contemporary American fiction to religious mysticism to some rather dreary texts on communist political themes If you are a voracious reader with a wide-band approach to the selection of your reading materials maybe you can help Bill identify some of his bits and pieces If you'd like to give it a try and if you have a green badge call him on 4382s June 79 CRYPTOLOG Page 21 P1 MY79 S3 1343 SECRE'f SPOKE 86-36 This document is from the holdings of The National Security Archive Suite 701 Gelman Library The George Washington University 2130 H Street NW Washington D C 20037 Phone 202 994-7000 Fax 202 994-7005 nsarchiv@gwu edu