WlDVVWlD 1 5lBWWVV lD 5WlB L'Ul V lHl3Ul lHl3 lH OOl3 L1fill3 f OO LI l1WlIlfil f rn Urnlbrn GBPr1 Em D E 1J0 0 86-36 r-_ _ IAND A LITTLE BIT OF LUCK D H W o o 1 iNT' cHAiN' 'GANG ' 'j' KNOW YOUR GEOGRAPHy ' o o o o 6 _ DATA STANDARDS WITHOUT TEARS Mark T Pattie WELL MAYBE A SNIFFLE OR TWO LITERATURE SEARCH ON-LINE oooo ooo 1 f IN DEFENSE OF THE INDEFENSIBLE NOTES ON THE RUSSIAN PQE Emery Tetrau1t LETTERS TO THE EDITOR 15 LET'S NOT FORGET OUR CRYPTOLOGIC MISSION o J Gurin o NSA-CROSTIC NO 22 o o D H W P S 1 t 7 9 10 13 21 16 18 20 'fIllS BOCtJMEN'f CON'fltINS COBElNORB Mlt'fERbU 6LASSIFIEB BY NSA SSSM 123 2 1 Pee gg RE'IIEW eN 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 4009823 TOP SECRET Published Month y by PI Techniques and Standards for the Personnel of Operations VOL VI No 2 FEBRUARY 1979 PUBLISHER WILLIAM LUTWINIAK BOARD OF EDITORS Editor-in-Chief David H Williams 39575 Collection 1 Cryptanalysis 1 IC8S55s IC4902S P L -1 - t Information Science 1 7 1 Language Machine support I S98i 1 Crypt 01 i ngu i st ic 5 1 57115 ICSI 61 5 1 1 50845 -----l Mathematics 1 Special Research ' 1 85185 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 TOP SECR T 86-36 DOCID 4009823 SECRET SPOKe FROf' THE By DAVID H EDITOR WILLIAMS EO 1 4 c P L 86-36 L --_landaLiltleBifofluck One of my first actions after becoming the have never appeared and I'm sure there are new editor of CRYPTOLOG was to sit down and numerous examples both current and ancient read through the back numbers One of my prin- of both of them here and there around the cipal discoveries was finding out that virtual- _A g e_n_c y_ ly all of the articles could be assigned by general type to one or another of a rather small number of categories By far the most frequent was the This is Such-and-such articie a straightforward description of something an organization an analytic technique or skill a machine program a new piece of equipment or whatever This article is the lifeblood of the publication and every issue contains at least two examples of the type A special sub-category of this article is the Here's HOW We HandZe or HandZed Suchand-such piece covering one organization's approach to an operational problem To my mind one of the best of these was The Iron Thumb byl I March 1978 on the subject of MOREIOWN B Group's machine assist to the Chinese language translator Another frequent theme is found in the cry of Anguish item which describes some major injustice frequently relating to promoticns especially in regard to their inequitable distribution to persons of one cose usually linguists vis-a-vis another usually engineers Articles of this sort are often followed by a flurry of rebuttals and counter-rebuttals in subsequent issues Interestingly one subdivision of Cry of Anguish the comment on the performance appraisal or the performance appraisal system never receives any response in defense of that system P L 86-36 One of my favorite articles is Oops We This article which understandably appears relatively infrequently details some organization's adventures and misadventures in going astray in the handling of some SIGINT phenomenon and how they eventually c9ped with Goofed I There are however two types of potentially worthwhile articles which to my recollection 0 February 79 CRYPTOLOG Page 1 SECRE'I' SPOKE 0 EO 1 4 c P L 86-36 ---1 DOCIO EQ 1 4 c P L 86-36 4009823 SECRE SPOKB I So if XDU have any Ipi 1ing up in your desk dp6p a note to CRYPTOLOG Maybe we can help you find the other half of your puzzle 6 if you have any examples of breakthroughs made without the slightest intervention of any scientific know-how and technical expertise send them along Incidentally if you have something you think you might want to contribute in either category or on any other subject for that matter but you feel a slight lack of selfconfidence in your ability to put it down on paper call or drop in at the CRYPTOLOG office We're always glad to provide whatever editorial assistance might be needed The CRYPTOLOG office is located in PI Room 2N039 the telephone is 3957s CLA NEWS RUSSIAN INSTITUTE The Crypto-Linguistic Association has approved the formation of a new special interest group -- the Russian Institute The new group's purpose as defined in the gUldellnes is to provide a forum at which students of the Russian Soviet problem can discuss and exchange information and insights gained from research in the field The guidelines call for the Institute to present Soviet films and lectures by visiting scholars as well as to encourage its own members to take an active role in leading discussions and seminars on various topics The Russian Institute's guidelines define the Russian Soviet problem in a very broad sense the Institute hopes to include presentations on a wide variety of both work-related and more general topics I Membership in the Russian Institute is open to all CLA members in good standing If you are interested in becoming a CLA member or in renewing a membership that has lapsed dues for 1979 $3 00 can be sent to ' iJI Treasurer mP16 m A# L 86-36 2N039 4032s A memo and survey form concerning the Institute have been distributed among Russian linguists If you did not receive a copy and are interested in further information please contact A65 m 2NOOl 4367s U February 79 CRYPTOLOG Page 2 SECRBT SPOK L 86-36 DOCI-D 4009823 P L 86-36 'fOil SECItE'f UrtlBItA THE COMINT CHAIN IIGANG G51 EO 1 4 c P L 86-36 P got a glimpse of how far-ranging NSA's activities are at NSA Day 1975 right after I came on board I got the immediate feeling that there was a lot about the Agency I would never see ostensibly because I had no need to know It also became clear quite early that it would often be hard to find someone with sufficient knowledge to answer my questions even when I did have a need to know I Being inquisitive that's the word that applies to me -- nosey applies to other guys and being interested in learning what the real career prospects are in the Agency I began to ask questions about how things work in other offices I asked my intern friends to compare and contrast what they had seen I asked my classmates at the training school what their jobs were like I also happened to hear conversations involving the occasional visitors to my work space I really wasn't trying to eavesdrop I even went so far as to listen to what was being taught in my professionalization courses The upshot of all my investigation was the decision to write a tunnel vision overview of COMINT operations mainly for a friend in the Agency for whom the word product has little to do with the publication of intelligence That term tunnel vision is a disclaimer so that if anything inaccurate is said it can be blamed not on my personal inadequacy but on my limited perspective one of my original aphorisms states that what you see depends on where you stand With that warning you the reader are now ready for a worm's-eye view of the chain of COMINT Production 1 4 c 86-36 us who have veteran's preferences might have to exercise them to get jobs in HEW or DOE or maybe even forget the preferences and scramble for jobs in productive enterprises Because we must wait for something to happen the best the Agency can do is to try to plan to be in the right place at the right time with the right equipment Fast-breaking events are in an entirely different league Because of the limitations of money politics and technology NSA simply cannot have a finger in every possible pie For that reason private newspaper reporters are occasionally much better equipped for gathering information in remote places than we are or for that matter better adapted than anyone inside the intelligence community The Agency subscribes to three wire services to take advantage of the mobil i ty of thei r reporter O 1 4 c Right in the first link then we depend d lJ L 86-36 factors largely beyond our control in order to get needed information into the hands of oUr decision makers Worthwhile News The first link in the chain is a newsworthy event Actually the newsworthy event is so important to end product that a better metaphor might be an anchor or a hook or some other object on the business end of a chain In reality until something important outside the intelligence system occurs the system has no function no reason for being If all threats to national security military and economic evaporated once and for all those of February 79 CRYP' 'OLOG Page 3 TOP SECRET UMBRA DOCID 4009823 'fOP SECRE'f UMBRA ReadabZe Enc yptio These first four links are all outside of NSA and the intelligence community The rest of the chain is directly controlled by the communIty Easy InteY'cept Modern Transmission After a message is encrypted the next step is to transmit it In this fourth link of the chain there are still some factors Qver which the Agency has no control The way a message prob la 11 SjQD Qfr ti lCRY15TOrn the AttaooyT January 1978 L - February 79 CRYPTOLOG Page 'fOP SECRET UMBRA ----I 4 DOCID 4009823 TOP SECRET UMBRh Rapid becY' ption EO 1 4 c P L 86-36 formed by lingu1sts 1$ called dispos1t1on1ng or To perform that task the scanner must not only keep a breast of a fluid set of requirements imppsedby those who oversee the Agency but also stay on top of what has already been published and what the consumers of our product think s hot any given time _ ----------------1 scanning Efficient Staff o Recognition and Language Processing One of the most important functions per- I _---------------------' February 79 CRYPTOLOG Page 5 TOP SECRET UMBRh Inte ested Custome s Casting Off Well there YOU have it folks Everything you never needed to know about production Obviously in such a short space many things had to be glossed over or skipped altogether Anyone whose career field was slighted will just have to expand my horizons through an article in this magazine KNOW YOUR GEOGRAPHY R G Tony Melzer D5 as a body of salt rank more or less landHow many of these seas can you locate Adriatic ---Aegean ---Arabian ---Baltic ---Barents ---Bering ---Black ---Caspian ---Coral Japan ---Mediterranean ---North ---Norwegian --Okhotsk --Red ---South China ---White ---Yellow Solution on page 21 U February 79 CRYPTOLOG Pase 6 'fOP SECRET UMBIU' DOClD UNCLASSIFIED L 86-36 DATA STANDARDS WITHOUT TEARS 11 A635 11 E GBPr-lIIo- sT o oM TJ strayed into the world of data standards quite by accident Probably few do otherwise As data base design manager for a fairly large project I was concerned with the creation of a Data Element Dictionary Directory OED D system mainly to document file designs and to provide some general-purpose software such as data editing The establishment of data standards was expected to be a simple and natural product of the Dictionary This initial naivete was soon shattered when it appeared that the length of time required to coordinate standards with the central authorities are likely to exceed the life span of the project Even worse there appeared to be valid reasons why complete standardization could not be enforced within the project itself What happened was what will always happen Standards took second place to operational necessity The project was completed satisfactorily as was the OED D As far as standards were concerned we did our best turned a blind eye to some abuses and 'hoped for better things in the future I Since then I have talked at length with data standards experts studied how things are done and reflected on the experience of the OED D My conclusion is that data standards can be achieved but that our present approach is wrong We overcomplicate the issues and lose sight of the real requirements we worry too much about irrelevancies we view standards as something which not only can be but must be imposed in an inflexible hard-handed manner Let's clear our minds and start again Consider these two statements 7902- - o Everyone agrees that data standards are a good thing in theory o No one agrees that data standards should be enforced on his project at the expense of operational necessity The crux of the problem is that these two truths appear to be mutually exclusive so that when faced with the question to be or not to be standard we feel constrained to reply either yes or no A better way would be to answer Both We need both to have standards and not to have them Standards without tears means that we want to have our cake and eat it too It goes without saying that this cannot be achieved without some degree of magic On the practical level the magic machine already exists for rendering coarse materials into fine standard gold It is without doubt the Data Dictionary Directory On a more nebulous level a much stronger alchemy is needed to bend our present attitudes onto simpler more flexible more realistic paths Firstly let us have a look at the mechanics of using the DED D to achieve standardization without tears From the plethora of terms in present use I have chosen two data element and data field Forget subfields data chains data use identifiers use modifiers etc At this point they are red herrings In this simple system the terms are defined thus o data elements are things o data fields are homes for things CRYPTOLOG Page UNCLASSIFIED 7 DOCID 4009823 UNCLASSIFIED In other words a data field is a receptacle for holding a data element and a data element is mostly what you want it to be but essentially refers to the content of the data field Do not worry too much about the exact definitions at the moment Concentrate on the idea of two domains because I do not believe this to be arbitrary It is a natural separation and although apparently simple is fraught with consequences philosophical political and practical The immediate practical consequence is that the structure and functions of the DED D are greatly simplified and the associated terminology is made clear The Dictionary section of the DED D describes data elements The Directory describes data fields and how ney are built into records and files or how they are used in software No confusion exists because the Dictionary and the Directory deal with two inherently different concepts A data element name cannot be confused with a data field name because the one is the name of the thin and the other is the name of the box in which the thing is put Similarly a data element abbreviation is the agreed standard abbreviation or code by which a data element is known while a field name abbreviation is just the name of a container in which data can be housed Some important attributes belonging to these two domains are DATA FIELD Data field name Data field name abbreviation Field length actual Position in record Thread from Directory to Dictionary DATA ELEMENT Data element name Data element abbreviation Data length max min Definition Values DATA ITEMS Configuration s The philosophical consequences of the subdivision into two domains relate to the fact that standards are applied initially at the Dictionary level only The Directory confines itself to describing actual use The former describes the ideal world the latter describes the actual world The thread from Directory to Dictionary listed above links the real with the ideal Before the hardliners point out that this is not standardization but the avoidance of standardization it is necessary to stress that this is a starting point only The DED D is not just a vehicle in which standards are enshrined for the multi tude to gape at It is the mechanism for developing standards The success of standardization can be measured directly by the degree to which the data field descriptions in the Directory match with the data element description in the Dictionary There is no other practical way to tackle the problem A standards authority cannot confine itself to documenting agreed standards because much of the time these do not exist Instead the details of actual use should be recorded and analyzed so that new standards can be evolved Standards will be much more acceptable to users if they evolve in this way from actual use Politically the subdivision into two domains is exceedingly useful The Dictionary may be maintained and controlled by the central authority Separate Directories can be built in each user area according to user specifications with a large degree of freedom saving only that the link from Directory to Dictionary must be present in the form of some standard number or designator so that varying data field uses may be related to a single standard set of data element conventions When a user builds a new file he checks first to see if the fields he wishes to use have corresponding data elements defined already in the Dictionary Say for example that he finds the data element name DATE OF BIRTH with the data element abbreviation DATEBRTH There is a strong possibility that he will adopt the identical name and abbreviation for his data field thus conforming to an existing standard But maybe his system hardware or software demands an abbreviation not more than three characters in length He may then say I accept the standard data element abbreviation DATEBRTH but I must use DOB for my data field name abbreviation Two other situations must be catered for the case of the user who rejects a standard for reasons of personal preference and the case of a file or a software system which exists before the standard is set up where the effort required to change it is unacceptable In the first case friendly persuasion is called for In the second there is no question but that the user wins His is an overwhelming case against the indiscriminate enforcement of standards The DED D system must be democratic not authoritarian It is the only reasonable way to proceed And most people are not rebels They will accept a standard if they are able to Thus the DED D becomes a powerful force towards encouraging standardi ation whilst being able to accommodate all deviations The terms I have been using to express the idea of two domains data element and data field are not sacrosanct The concept matters more than the terminology There are other words which I rejected earlier as red herrings e g data chain data use identifier I did not mean to imply that they do not exist merely that they are of secondary importance If my data element is observed under a microscope it does indeed resolve itself into a number of different types DATE becomes a data chain made up of YEAR MONTH and DAY DATE OF BIRTH is seen to be a 7902-- CRYPTOLOG Page 8 UNCLASSIFIED DOClD 4009823 UNCLASSIFIED composite structure made up of the basic noun DATE plus a usaRe modifier The mistake is to imagine that this causes any proplem Data Dictionary designers are sometimes bewitched by ideas of levels and hierarchies The user needs definitions of YEAR MONTH DAY DATE DATE OF BIRTH even DATE OF BIRTH OF FATHER without worrying about type Keep it simple Just enter them all into the Dictionary as plain data elements There are many easy methods of expressing the various relationships within the Dictionary It also appears on close examination that the data element is not really the thing itself but the name and description of a set of things Thus DAY OF THE WEEK is a set name The actual members of the set are values such as MONDAY or TUESDAY These may he calle i data items or anything else but again this poses no problem They are simply entered into the Dictionary subordinate to their data element Some clarification is needed however to determine the exact relationship between data item and data configuration To sum up standards cannot be created in a vacuum They must be developed from current usage and this development cannot be carried out without automated assistance of the type offered by a Data Element Dictionary Directory system which documents the real world as well as the standards world There are two ways of tackling standardization the easy way and the impossible way We've been trying the latter for some time now It's time for a change m WELL MAYBE A SNIFFLE OR TWO Ma rk T Patt ie P13 '0 NSA DATA STANDARDS CENTER J Since the topic 0 fl lartiClEii faLLs within the area of respons biLity of the NSA Data Standards Center we sent a prepubLication copy of the articLe to Mr Pattie the chief of NDSC and asked him if he wanted to comment on it He did ADED7D b 1 d t' i 36 6- useful and valid as the material plac eq in it The saw Garbage in garbage out applies here too That is why I take exception to he referens t6the alchemy of ren ering coahe material into gold which Mr sp aks of in his article The DED D -an provide valuable information about already cxisting stanqards and it can al so help us pU ll together el ements that ought to be regarded as potential standards Simply placingl'Daterial in the file omehow I have I am will not automaticallY make it gold though spending mor time writing to CRYPTOLOG for it may onl be u seful within Sing l than I should but Data Standards pro ect and If I understand Mr _1 perWithout Tears demands a response Mr haps only within apiece of a project Some --' iises words very prettily and if I did one would have to work with such material to not already know better I might think he had make it acceptable to another Agency element made a worthwhile point However he gives himself away right in the first paragraph with Mr 'j tin his castigation of experts his there appeared to be valid reasons and authorities has really set up straw men why complete standardization could not be enfor his pot shots The NSA Data Standards forced within the project itself StanCenter has always held that while we would dards took second place to operational necesprefer to see Project Managers use existing sity If we change necessity to convenience data standards we recognize that individual I think we might be closer to the truth projects may work without them Whatever terms they corne up with though are not stanusually when someone begins a project he dards and should not be regarded as such We looks around and tries to make use of already existing material standards so he won't have would have to test them out in other uses before we could determine what the Agency status to reinvent the wheel He then builds upon should be And that as I pointed out in the what is useful to him and attempts to develop previous paragraph is where the DED D can be new material as necessity demands Within very helpful It is a tool for us to use It his own project though I would think he is not the end-all would be able to enforce standards Otherwise he must end with a Data E ement DictionA more detailed comment on Data Standards ary Directory that will be anything but stanWithout Tears is being prepared for next dard and reliable month's issue of CRYPTOLOG U S thefe efi hat I 1 7902-- CRYPTOLOG Page 9 UNCLASSIFIED DOClD 4009823 UNCLASSIFIED LITERATURE SEARCH ON-LINE I P L rl1213 A re you looking for information that may have been in open-source literature or documentat10n An item from an unclassified source Or an unclassified item published by o government source or by a commercial organization working under government contract If so DIALOG may be your answer p blished DIALOG is a comprehensive collection of on-line data bases maintained by Lockheed Information Systems Palo Alto California Subject areas include science technology engineering social SCiences business and economics Sources of information include research reports technical reports journal articles government reports and publications periodicals pamphlets reviews monographs conference proceedings and books All data bases are regularly updated to include the most recent information Additional data bases are being added continuously All hese data bases can be accessed on-line Tl2 Information Services has several terminals connected to the DIALOG system Many of the indexes covered by the DIALOG data bases are subscribed to in hard copy by NSA These indexes are available for use in the NSA library DIALOG greatly reduces th tim necessary to perform most searches as compared with unual techniques for searching literature The accuracy and exhaustiveness of searches performed through DIALOG allow the researcher to retrieve a greater nUlllber of Number of InforName of mation citations data base period update producer covered period ABI INFOR '1 Data Courier Inc louisville Ky COMPENDEX Engineering Index Inc New York August 1971 to present 65 000 monthly 1969 to IN5PEC present Institution of Electrical Engineers London England The following table lists a few of the data bases cur- I rently included in the DIALOG system o Abstracts of s1gn1ficant artlcles of 1nterest to MAt AGEMENT lind ADMINISTRATION from approximately 400 publications in business and related fields Representative publications are Harvard Business N Y I tainable through a comparable manual research effort Because of the savings in research ti and the increased thoroughness of the search results DIALOG is an extremely cost-effective research tool A word about the cost of DIALOG With DIALOG the custoaer government agency co ercial or2anization etc pays only for the services that are used Search charges are based on the time the terminal is connected to the computer prorated to the nearest thousandth of an hour There is also a charRe for each citation printed and mailed to the custoaer Rates vary depending on the data base used The expense of the DIALOG system has been included in the Tl2 budget Although Tl2 analysts who operate the terainals aust be cost-conscious and stay within this budget they make every effort to provide complete search service for all operat ional probleJlS Contents of data base Revieru January 1970 to present - I J documents WIth a higher degree of relevance than those ob- Review For'tune Dun's Review $65 10c Sloan Management and Journal of Marketing Research Machine-readable version of the Enginee ng Index 600 000 monthly monthly annual which provides the ENGINEERING and INFORMATION communities with abstracted information from the world's significant engineering and technological literatur The EI data base provides worldwide coverage of approximately 3500 journals publications of engineering societies and organizations conference proceedings selected government reports and books 11 045 000 The SC1ence Abstracts family of abstract journals monthly indexes and title bulletins has been published since 1898 Today it forms the largest Englishlanguage data base in the fields of PHYSICS E LECTROTECHNOLOGY COMPUTERS and CONTROL Foreign-language source material is also included but abstracted and indexed in English INSPEC includes the Physics Abstracts Electrical and Electronic Abstracts and Computer and Control Abstracts The principal subject areas are indicated by the major headings of the unified classification approach developed for the data base e g Atomic and Molecular Physics Computer Programming and Applications Computer Systems and Equipment and Elementary Particle Physics Journal papers conference proceedings technical February 79 CRYPTOLOG Page 10 UNCLASSIFIED Cost to NSA on- line conneat hour fuU record printed offline $65 10c $45 10c 86-36 DOCIO 4009823 UNCLASSIFIED reports books patents and university theses are abstracted and indexed for inclusion in the INSPEC aata base Approximately 2000 journals are scanned of which 200 are abstracted completely -MA-N-A-G-EM-E-NT-- S-e-p-t-e-m-b-e-r 2-5-0-0-0---- -C1-1r-r-e-n-t-i-n-f ' o-r-m-a-t-i-o-n-on- -a vanety 0 t BUS I NES S a-n-1d - ----CONTENTS 1974 to onthlY At AGEMENT related topics to aid individuals in $55 10 Managepresent business consulting firms educational institutions ment Congovernment agencies or bureaus and libraries in decision-making and forecasting Articles from ap tents Skokie proximately 200 U S and foreign journals proceed Ill ings and transactions are fully indexed and abstracted to provide up-to-date information in the areas of accounting decision sciences finance industrial relations managerial economics market ing operations research public administration etc Government-sponsored RESEARCH DEVELOPMENT 6lu OOO NTIS 1964 to biweekly and ENGINEERING plus analyses prepared by federal ' 5 10c National present agencies their contractors or grantees It is Technical the means through which unclassified publicly Informaavailable unlimited-distribution reports are tion Sermade available for sale from such agencies as vice U S NASA DOC AEC HEW HUD DOT Department of Department Commerce and some 240 other organizations of ComState and local government agencies are now bemerce ginning to contribute their reports to the file SpringThe data base includes material from both the field Va hard and soft sciences including substantial material on techno ogical applications business procedures and regulatory matters Many topics of immediate broad interest are include such as environmental pollution and control energy conversion technology transfer behaviorall societal problems urban and regional planning Information ill all fields of SOCIAL SCIENCE PAIS INTER- uUetin 45 000 $60 l5c quarterly including political science banking pUblic adNATIONAL 976 to ministration international relations economics Public present law public policy social welfare sociology Affairs In- fForeigneducation and social anthropology 25 000 citaformation language tions taken from more than 1200 journals and more Service 'ndex than 8000 books pamphlets government publicaInc New 1972 to tions agency reports and other documents are York N Y present added yearly 'yrs WI'S lPredlcasts TerIJ lnal System maIntaIns a number varies according DATABASES of BUSINESS-related files which are available on $90 20c to file PrediDIALOG The EIS Plants data base provides immedicasts ate answers to a broad range of questions concernInc ing the U S industrial economy The F S Funk Cleve nd Scott Index cover both domestic and interland national company product and industry informaOhio tion They contain information on corporate acquisitions and mergers new products technological developments and sociopolitical factors The Federal Inctex covers such federal actions as proposed rules regulatIons bill introductions speeches hearings roll calls reports vetoes court decisions executive orders and contract awards The Washington Post and federal documents such as the Congressional Reaord Federal Register Presidential documents and Commerae Business Daily are indexed on a regular basis The Pl ediaasts Mar'kBt Abstroaats data base provides information on acquisitions capacities end uses environment foreign trade market data new products production regulations and technology ' Significant information is ahsLracted from February 79 CRYPTOLOG Page UNCLASSIFIED 11 1' DOClD 4009823 UNCLASSIFIED thOusands of newspapers business magazines government reports traJe journals bank letters and sp ctal reports throughout the world SCISEARCH Institute for Scientific Information Philadelphia Penna January 1974 to present 1 820 000 monthly SOCIAL SCISEARCH Institute for Scientific Information Philadelphia Penna 1972 to present 490 000 monthly SPIN American Institute of Physics New York 1975 to present N Y SSIE CURRENT RESEARCH Smithsonian Science Information Exchange Washington D C last two years e g 1977 to present Multidisciplinary index to the literature of SCIENCE and TECHNOLOGY prepared by 151 It contains all the records published in Science Citation Index and additional records from the Current Contents series of publications that are not included in the printed version of SCI SCI SEARCH has two important and unique characteristics First journals indexed are carefully selected on the basis of several criteria including citation analysis resulting in the inclusion of 90% of the world's significant scientific and technical literature Second citation indexing is provided thus allowing retrieval of newly published articles through the subject relationships established by the author's reference to prior articles SCISEARCH covers every area of the pure and applied sciences Multidisciplinary data base indexing every significant item from the 1000 most important SOCIAL SCIENCES journals throughout the world and social sciences articles selected from 2200 additional journals in the natural physical and biomedical sciences SOCIAL SCISEARCH includes many important monographs as well The data base covers every area of the social and behavioral sci ences Unique information retrieval technique in addition to more conventional retrieval by words author journal name corporate source etc it is also possible to search by way of the author's cited references 50 000 SPIN Searchable Physics Information is designed monthly to provide the most current indexing and abstracting of a selected set of the world's most significant PHYSICS journals Each month approximately 1400 additional articles from the journals published by the American Institute of Physics including the Soviet translations are added to the data base Author-prepared abstracts enhance the relevancy of retrieval from the data base SPIN covers all major areas of physics as well as mathematical and statistical physics astronomy astrophysics and geophysics 180 000 Reports of governmental and privately funded monthly SCIENTIFIC RESEARCH projects either currently in progress or initiated and completed during the most recent two years SSIE data are collected from the funding organizations at the inception of a research project and provide a source of information on current research long before the first or progress reports appear in the published literature SSIE CURRENT RESEARCH encompasses ali fields of basic and applied research in the life physical social and engineering sciences Details on all the DIALOG data bases and further information on how DIALOG can help you can be obtained from $70 10ct $70 10c $35 10c $90 20c o T122l Room 2C05l 32S8s or o T1233 FANX-III Room 81820 8705s eU o T12l3 Room 2N090 5759s February 79 CRYPTOLOG Page 12 UNCLASSIFIED DOCID 4009823 -UNCLASSIFIED IN DEFENSE OF THE INDEFENSIBLE Notes on the Russian PQE Emery Tetrault P16 A called operational time not to mention subbotnik time as well that the same someone said Enough already And so the search for an alternative method began This alternative turned out to be a two-part test not unlike the Russian I A that we all know and love today Early in 1976 we ran a kind of feasibility study with the help of some people who had just taken Russian PQE 12 About 32 hardy souls took an experimental test battery the results of which were correlated with the PQE Part I which they had just completed The correlation was quite high and more significantly the first six ranks on the experimental battery were filled with five of the six people who had passed the PQE The results were good enough for the Language Career Panel and the Russian PQE Committee and the decision was made to go ahead with the new test forms The last two Russian I A exams have represented somewhat of a departure from the forms and procedures first used in 1976 The Russian Completion Subtest has changed very little although verb pairs are now given in the word lists rather than single verbs but the Translation Completion Subtest is a substantially different animal from what it was originally On both subtests examinees are allowed to use dictionaries this was not the practice at first For all languages except Russian Part I A The Completion Subtest is a text of 220-240 of the PQE is made up of a 500-550 word text of words with 50 words deleted and 70-75 candithe language variety often found on the op-ed dates for these blanks given on a facing page page of a newspaper The test task is to produce a smooth idiomatic translation into Deletions cover a spectrum of difficulty from the trivial to the nearly impossible Items English In reality few if any test papers can be said to describe a kind of bell curve realize this standard but that is a story most of them are medium-tough with a few easy for an other day and another article Up until PQE 13 July 1976 the Russian I A ones and a few hard ones thrown in on the ends was like this too However someone somewhere of the spectrum Items deleted are both syntactic and lexical both bound and free rebegan adding up the person-hours involved in coverable only from situational context Morrunning this kind of PQE for Russian Test preparation was a negligible part of the total phology is tested as a by-product since most but scoring the test was a nightmare Grading content words inserted into blanks have to be inflected to fit both linguistic and situational translations has never been an exact science context A brief sample without word lists and therefore it was done by a committee -is given below five senior linguists in the case of Russian Each grader graded every paper independently o o o Bce Hati8 J Cb _- C __ using the approved grading system and then the committee met to make a preliminary evaluation npo cmeCTB R ABapMM He Papers were labeled Pass Fail and Maybe regraded during a series of follow-up meetings and all the Maybes were regraded during a series fo 1'0 o HO 06cTaHOBKa of follow-up meetings As the test population cos u aJlSCb n e o KOr D S for each successive Russian POE began to reach 120130 this process started taking so much soFebruary 79 CRYPTOLOG Page 13 lthough the current Russian-language population at NSA is made up for the most part of shy self-effacing scholars its members can be uncharacteristically vociferous on the subject of language testing in general and the PQE Professional Qualification Examination for anyone who has just returned from 10 years on the dark side of Mars in particular This is a healthy phenomenon but it should not merely assume the form of a diatribe against the pinheads who devise such tests not because of humanitarian concerns but rather because such a course of action is ultimately self-defeating I am convinced that there are some individuals who have failed the Russian PQE particularly Part I out of test fright or more accurately out of test repulsion They have shared so many horror stories about the PQE with their colleagues that it has become for them an insurmountable barrier a psychological Berlin Wall As a result their performance on the PQE may not be consistent with the actual level of their language skills This is sufficient reason for someone to present at least a factual account of current test practices if not a spirited psycholinguistic defense of the PQE I shall limit my remarks to Part I A the non-voice open-source section of this lengthy and occasionally unpleasant process UNCLASSIFIED DOCID 4009823 BOA Tenb ROTopaR A UNCLASSIFIED MOP030B Ha o aBTo amH e in the kinds of situations WhiCh '''''-lAwS'''' CkT HHCTHTyTy ftrocr npoBo x03ft VI ' TOHHeJlR j06P HHHCHO nJlO b BH ep n AHCTaH HK npH of physics biology medicine no psychology He number of subfields Bo H36emaH e ne d CTOJ1RHOBeHHR co Consequently a within the field of ' IK H L Yll eii ualllHHoi1 OH pesRo H8 f 8 n HA TOpj 03a H aBToMo6Hnb paSBepHyn l onepeR - porM This sample shows some of the possible kinds of deletions bound prepositions C and Ha partially keyed items npHHa nejl HT BHXO l1e nO l1 etc and the more or less free informational items OnaCHaR Bnepe ru'l Any plausible response other than the keyed answer is tried out on a native speaker Scoring is simply two points per blank with half credit given for a good selection incorrectly inflected asspect choice is considered a selection not an inflection The Translation Completion Subtest has gone through a lot of growing pains as a test form At first it was thought that the underlying translation should be as different structurally from the original Russian as the law allows However the tendency now is to produce a reasonably smooth English text and let the natural differences between Russian and English particularly the differences in informational structure as marked by element ordering do the job without too much outside help Once again we have a brief sample a Russian text followed by a partial translation CerO HR RpHMHHanHcTy Henb3R 6HTb npOCTO Rp MMHanVlCTOM llpH paccneAOBaHHH AeJl OH qaCTO nOna aeT B T8KVle cMTya VlH r e HymHo rny6oRo SHaTb cHSHKy XHMMID DHOnOrMID nCMXOnOrM ooo IT03TOMY B COBpeMeHHOy KpMMMHaJl CTHHe nORBJln CR 60nee Y3RVle npOcfeCCHH The criminologist today t 4 tt4 himselfto criminology Ito 'C4W'AC- 'b - em r he will often Q ' The weighting or point value attached to each blank in the translation depends on the segment of Russian text being processed not on any particular number or combination of English words Thus any response that fits the context given and that accounts for the semantic features of the original Russian is correct and in actuality there are many different correct answers in this form of test just as there would be many different correct translations of any given text When a different set of tests is used for each PQE cycle there is justifiable concern about the reliability and validity of such a testing procedure For this reason all the potential test forms for a particular PQE cycle are first tried out on a small number of certified linguists and the best example of each test type is selected for the PQE In the case of the Russian Completion Subtest a native speaker of Russian is also asked in fill in his or her responses a practice which we initiated with PQE 17 and which has proven to be extremely useful This form of pretesting addresses the issue of cont nt validity Predictive validity and reliability are ensured by having a small group of controls four or five certified linguists take the PQE along with the test population The performance under actual testing conditions of this control group is then used to set the Pass Fail cut These individuals who make up the control groups for each PQE are performing an extremely valuable service and they should be encouraged and if possible rewarded by such bodies as the Language Career Panel and the A Group Language Advisory Committee Maybe a steak dinner wouldn't be too far out of line There is no doubt in anyone's mind that the testing procedures now used for Russian PQE Part I A save time and money The results of PQE 17 for example were in the hands of the Language Career Panel in a matter of a few weeks rather than 3 or 4 months and with little or no time taken away from line functions other than the time used by control groups to take the PQE alon with the test group But obviously cost effectiveness is not enough Part I of a language PQE must above all measure the right level and the right kind of February 79 CRYPTOLOG Page UNCLASSIFIED 14 II DOCID 4009823 UNCLASSIFIED language proficiency I believe it does and by abrupt changes in the Pass rates which I believe that the individuals who have passed fluctuated between 11 percent d 40 percent this form of PQE actually meet the minimum While no particular Pass rate 1S sacrosanct general standards set forth in existing deabrupt shifts in the percentage of people passscriptions of L-3 Minimum Professional Proing a PQE usually mean that test standards have ficiency With the exception of a small changed since the collective proficiency of the number of people who may be spooked by the NSA Russian-language popUlation doesn't change format most people who fail this exam do so all that rapidly because they don't know Russian Any other The present-PQE can and should be improved comparable measure of proficiency would proand it is the duty of the Russian-language duce the same results The errors I see re ommunity if there is such an entity to keep peated 50 or 60 times on a'hundred papers can be traced to certain features of Russian struc- the heat on the Language Career Panel to do so One way to achieve this goal is for supervisors ture and lexicon and not to any particular to encourage their professionalized Russian arrangement of blanks on a test form linguists to participate in the PQE process There is a tendency in discussing the curboth as pretest subjects and as controls In rent method of assessing L-3 general language this fashion most of the bugs would be shaken skills in Russian to forget what this method out before 150 people see them and test validity replaced Translation testing as it was and reliability would also be greatly enhanced practiced by Russian PQE committees is not This is one area in which a relatively small beyond criticism for a discussion of the investment of operational time would yield rationale behind the Agency's first Russianimmediate results language certification examination on 6 and No one is suggesting that current PQE pro9 October 1967 see The Russian-Language cedures are above criticism constructive or Certifying Examination by Arthur J otherwise but such criticism should never be Salemme QRL -- The QuClX'terZy Review for uninformed To simply dismiss the whole process Linguists Vol II No 2 February 1968 It as the product of warped minds or pointy-headed was hard for the PQE committees to judge the theories is to perform an ultimate disservice to relative difficulty of one text over another the cause of professionalism a cause to which and as a reSUlt one PQE could be and most all but the most cynical proclaim their certainly was more or less demanding that the u one preceding it a fact which was demonstrated allegiance at every opportunity LETTER TO THE Yesterday I reviewed the Personnel Summaries of 11 applicants for an overseas position Of all the papers that attend one's career at the Agency I believe that the Personnel Summary is the most important and influential EDITOR To the Editor CRYPTOLOG If CR PTOLOG had never existed except to printl larticle Some Tips on Getting Promoted 1I December 1978 the service to NSA employees would have been worth the entire CRYPTOLOG budget-since its birth I was very pleased to see such an important message receive the publicitryou gave to it The Personnel Summary can be a potent effective document for promotions rotations and overSeas assignments I am convinced that in the last 18 months at least three 'extra' G4 people have been promoted by the NSA boards -- because their Personnel Summaries were carefully written and reflected work well done Within G4 I use the Summaries to help with our rotation program CRYPTOLOG is a great publication Has been for years My personal files contain several selected articles clipped from preVious CRYPTO LOGs Butl larticle will be so useful I ask that you send me 100 copies I will circulate them in G4 I will make them required reading for touring interns I will use them for my class in Promotion Theory I Thank you and able service to the Agency Charlie Itor a valuBosti k G4 u Editor's Comment Charlie's note was just the first of several requesting additional copies ofl I rticle Anyone else wishing to--o b---tain addi tlc nal copies mat do so by calling the CRYPTOLOGoffice ext 3957 U February 79 CRYPTOLOG Page UNCLASSIFIED IS P L 86-36 Dacm 4009823- UNCLASSIFIED LET'S NOT FORGET OUR CRYPTOLOGIC MISSIO J Gurin R5 n acquaintance once de cribed to me how he had to stand by l el'plessly w lile his very own company welled its railks with what he terms nonproductive people He had started the firm himself to cash in on an idea he had developed for monitoring the public's reaction to TV fare A psychologist he teamed up with engineers statisticians programmers and others to develop and market the product which did very well It is now widely accepted by the television industry as one of the most reliable indicators of how well any program is doing But as the company grew and prospered he found that he was required to add all sorts of people who had little or no relation to the basic mission of the enterprise He needed a legal staff public relations people personnel counselors etc and admitted to some resentment at having to keep people on the payroll who played no role in improving the product or selling the service I am the one who does the hiring and firing he said and yet I can't control the addition of nonproductive types to my payroll A counted for But we have changed -- and I'm not saying that it is better or worse today just different ' There are many reasons why a large organization takes on people who have no direct involvement with the basic mission of the organization and many of those reasons are convincing enough But there is always the danger that the overhead will grow disproportionately large and that the sound economic balance of the enterprise may be threatened and productivity undermined In a recent article in The Washington Post Charles Peters former bureaucrat and now editor-in-chief of The Washington MonthZy attacked the federal bureaucracy citing some horrendous instances of waste and inefficiency No doubt many of his charges could be refuted by those in possession of the facts but some sounded authentic or at least familiar One of his tactics was to distinguish between what he called essential work and the work performed by those employees with titles like Planning Analyst Schemes Routing Specialist Manager of Creative Services Social Priorities Another illustration of this sticks in my Specialist Suggestions Awards Administrator mind In a magazine article a few years back Fringe Benefit Specialist or Confidential a retired Dean of Men from a small respected Assistant to the Confidential Assistant all private school wrote about his reaction during real titles a visit he made to the school some 20 years after retirement He was saddened to find It would be difficult without detailed inthat although the size and abilities of the formation to defend jobs with such seemingly student body were about the same and the frivolous titles although it is perfectly courses they took hadn't changed much the non- reasonable to assume that some could be justified to the satisfaction of the taxpayer Perteaching staff had increased greatly He was haps we have jobs at NSA that would sound frivocertain he wrote that the quality of the lous to those not privy to the jargon of bureaugraduates and of the instruction was much the cracy or to that of SIGINT or COMSEC activities same as it had been in his time So why all the extra staff Naturally the tuition costs Given the propensity of any institution to reflected their presence but he could find no add positions which may be only distantly justification for them in the output of related to its mission what is needed is some graduates way of fixing and maintaining a balance between those employees engaged in the basic mission Much the saffie thing happens with any instior missions of the organization and the other tution and this Agency is no exception I suppose that at one time in our history as the employees To be more specific it would be advantageous to be able to grade all employees work force came through the gate at the end of a day all but a few had been directly involved by the relevance of their particular jobs to the fundamental mission of their organization with processing intercepted traffic Most Having done this it should then be possible for likely the categories of traffic handler crypt clerk cryptanalyst traffic analyst and management to ascertain trends and to manipulate translator made up the bulk of those coming the balance with some precision through and after one added clerical help and But how do we define those jobs which are a handful of bosses all but a few were acFebruary 79 CRYPTOLOG Page 16 - UNCLASSIFIED DOCID 4009823 j j UNCLASSIFIED be on jobs carrying a 2 Typists and others who assist in this kind of operations should be on 3 jobs The farther the job is from the basic mission the higher the number To repeat the only function of the number is to indicate how close the job is to the basic mis sion not to indicate the value of the operation A security guard performs a vital function but the number attached to his job would Yes yes it is a complex probiem It cerbe high since his function is distant from tainly requires subjective judgments but then the basic mission Warehousemen and medical so do all hiring and staffing actions What personnel perform vitally important functions scientific rules govern tables of organization but their jobs are not basic to the mission Promotions Grade levels for jobs The value of a system of job-grading such as So what I suggest is that each job in NSA be this would lie in providing a display of the rated to indicate how cluse it is to a basic distribution of personnel resources by mission of the Agency This is not to imply mission-relatedness One could even envision that jobs which are not closely related to a deriving a number from the system which would basic mission are not important Far from it represent the average of all the job numbers It is just that the distribution of jobs must be a matter of proper proportion and that pro- multiplied by the number of incumbents It would be useful for example to know that if portion should be visible if it is to be kept under control If for example DIRNSA proposes our average was 6 22 last year it climbed to 6 93 this year Could the increase be explained to absorb a personnel cut with minimum impact away to the Director's satisfaction or should on SIGINT production he should be able to he set as a goal the reduction of the number specify with precision the types of jobs among so as to assure himself that we are not losing which the cuts should be made or in which sight of our real purpose in existing and they should not be made spending the taxpayer's money Suppose we assign the number 1 to those Several of my more cautious friends have jobs which call for direct participation in the suggested that this article hould be credited production of SIGINT product Cryptanalysts to an anonymous author to protect me from postraffic analysts appropriate programmers and others who are directly involved in getting the sibly furious reactions But I protest in advance that it would be reasonable to expect product ready for the consumer should be on that my own job would be given a very high jobs that carry the 1 label First-line number as it should although I consider it a supervisors who are not doing the work themuseful one selves in addition to guiding others should U most closely related to the mission of NSA And isn't the degree of relatedness a subjective judgment rather than a measurable quantity And as Bob Cooper recently returned to DDO from CSEC pointed out to me at lunch the other day the Agency surely has more than one basic mission CLARIFICATIOrl SOLUTION TO NSA-CROSTIC No 21 CRYPTOLOG January 1979 W Averell Harriman Soviet-American Relations Cryptologic Spectrum Summer 1976 Vol 6 No 3 In April 1945 I came back to brief President Truman on Stalin I found he didn't need very much briefing because he had read the Yalta documents with great care and all the post-Yalta cables He was a great student of history and that was why he was able to deal with these new and complex problems with such skill U The November 1978 issue contained an item stating that the Language Career Panel will award Certificates of Achievement to military linguists who pass the Language Proficiency Test Effective 1 October 1978 any military linguist who receives a passing score on the LPT will receive a certificate those who score five points or more over the passing grade for each of the two sections of the LPT will receive the certificate with honors Military linguists who passed an LPT prior to 1 October 1978 should apply to the LCP for a certificate The composite score requirements given in the November issue can be misleading and should be disregarded _ F o r o further information contact C ptc J M36 ext 8267J U J P L 86-36 February 79 CRYPTOLOG Page 17 UNCLASSIFIED DOCID UNCLASSIFIED 4009823 N$A-4 fi By n H W No 22 The quotation on the next page was taken from the published work of an NSAer The first letter8 of the WORDS spelZ out the author's name and the title of the work WORDS DEFINITIONS A Hill western city 201 139 182 127 225 B Patronizing one word prisoner going downstairs 2 wds C Removing facial hair again u Complete thoroughgoing comp E bridge in physics F Allen or Frome G Holmes' response when asked to identify the citrus plant A 5 wds H Mournful 1 Annoy tease J ---- and needles 2 wds K Thomas or Horace L Plants of the nightshade family M Involving the lending of money at excessive interest N Hometown of the comic strip trolley car O Undergoing evolutionary change P Strong surface current flowing outward from the shore Q Sleepwaikers R Broad sash often worn with formal attire S Logical sponsor for TV series about NSA 2 wds T Play 1961 and movie 1966 by Thomas Bolt about Sir Thomas More 5 wds U Chinese dentist time comp V Fellow roughening of the skin from exposure W Clumsy ungraceful comp February 79 CRYPTOLOG Page 18 UNCLASSIFIED DOCID 4009823 UNCLASSIFIED x Copy of a document made on specially treated paper Y Plant louse z Donizetti opera 4 wds I February 79 CRYPTOLOG Page 19 UNCLASSIFIED j Answer next month U DOCID 4009823 L 86-36 CONFIBI3N'fIAL he two articles on Project SPELL prompt me to add this p s to those tales of the ill-fated P S Mine was a passing involvement -- some few hours I recall -- but an involvement that allegedly contributed to the demise of the project T I gather that ASA did put a great deal of time effort and money into the project for as I heard sometime afterwards the $3 million project had been scrapped Don't askcme to vouch for that figure All I know is that at the time $3 million was a whole lot of money compared to a transcriber's salary I probably just saluted in reaction and went back to transcribing chores Those were the days when a transcriber was paid to transcribe and I think I should not say more What was my brief encounter with SPELLMAN Well I was a transcriber of questionable repute and a typist of some proficiency Oh I might just as well say it -- there probably weren't two or three others at the time who could do as well at the keyboard especially on the IBM Selectric with the Cyrillic keyboard So along with the tin ear which everyone had to some degree or other I had a pretty good reputation as a typist Thence I Nas selected to participate in the SPELLMAN I don't know what criteria were used to grade the test The results of my test were compared with the results of six SPELLMAN persons who had taken the test on their stenotype machines Six persons six parts 36 scores Holly informed me later of the outcome -- I had scored est better in 3S of the 36 cases Amazing I don I t r ' ---------------------l know ut if the recovery rate however Ed 1 4 uant fIed of a tin-ear transcriber es eeiall c L 86-36 WIt a typewr ter was t at go the SPELLMAN concept I don 't know what Holly and others co sern d with the project expected from my taking the test I have a notion the pl ywas to drive another nail in the SPELLMAN coffin To be perfectly fair and objective about it perhaps someone who could not type quite so rapidly should have been chosen to take the test I A J Salemme I Remember SPELLMAN CRYPTOLOG July-August 197C 1 More B S CRYPTOLOG oc t b e 'r lI '9 '7 ' 8 --- P L 86-36 February 79 CRYPTOLOG Page 20 CONFIIJENTIAL iii II1dIBLE nit eel rflff eIIA1414ELS eNLi DOCID 4009823 UNCLASSIFIED choices for the top jobs I thanked him but said I was not interested for I fear I would find it too frustrating to have to explain the facts of life in intelligence gathering to someone who should already have that knowledge I am sure that this organization was set up with the best of intentions but I am also sure that the actuality will in no way resemble the ideal Like Lou I am not optimistic LETTER TO THE Mark T Pattie Jr Chief NDSC EDITOR u IN CASE YOU'VE ALWAYS WONDERED To the Editor CRYPTOLOG In the December issuelr------- 'To Whom It May Concern describes a sad situation but leaves out an even sadder part of it I believe he was present in the auditorium when those who were selecting their supporting casts were giving their pitch to NSA personnel who might be inter ested in working in the new organization When I left the auditorium I had the sinking feeling that I had been listening to a couple patent medicine barkers The top man introduced himself as someone who had no previous position in the Intelligence Community and he said that he looked upon that fact as an asset No predispositions I suppose When he finished his spiel one of his chief assistants took the floor to describe the actual jobs he was trying to fill He said he had had a brush with the intelligence part of his former agency when he first went to work ther but succ d- ing jobs took him far afield He left a number of questions unanswered but I thought that with the proper selection of his personnel the concept might work Several months later the second man returned to NSA to interview those who were interested in hearing more about the jobs In my session with him I learned that he had been using the intervening time to fill the higher positions with people similar to himself with little or no experience in' intelligence work What he was seeking he said was assistants who would provide the proper support to his I J wasta lkingnottoolonlLagowithoneof the young ladies of the Agency who was telling me about a recent encounter Seems she had been in the ladies' room washing her hands when one of the members of the custodial force came in and walked up to the wash basins She looked at all the little paper towel halfmoons lying on the basin rims and on the floor and muttered something about all the thoughtless people who dirty up the restrooms My friend tried to suggest quite gently that maybe the reason all those semicircles keep tearing out of the edges of the paper towels is that the towel dispensers are loaded too tightly at refill time Nonsense said the custodian The reason all those pieces tear out is that people grab for the towels when their hands are wet dhw U Solution to KNOW YOUR GEOGRAPHY ily Tony Melzer DS 10 11 12 13 14 IS 16 17 18 1 North 2 Norwegian -3 Baltic 4 Black 5 Caspian 6 Red 7 Yellow 8 Okhotsk 9 Japan Bering South China Barents White Arabian Aegean Adriatic Mediterranean Coral U CLA - NCS FEATURE FOREIGN FILM FOR FEBRUARY - UIJE7SU JAPANESE SOUNDTRACK WITH ENGLISH SUBTITLES THURSDAY FEBRUARY 8TH AT 0930 FRIEDMAN AUDITORIUM U February 79 CRYPTOLOG Page 21 UNCLASSIFIED Pi-Dec 78 S3-27141 P L 86-36 CID 44009823 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