P L 86-36 l Bl JWUUij GBPDlBl UJ Bij tlD ilV 1 5 D ilI 5 1 '1 OO 5UJlD 5 f OOUJ il l1UJWlD UJGBPDUUlDUJGBPD11 Plf P L 86-36 AN OVERVIEW OF PROJECtl 86-36 f 0 o 0 0 0 0 CADRE REVEALS WHAT DATA CONCEALS 0 0 0 oS 1 r 1 V o 7 SIGINTPUBLICATIONMANUAL INPREPARATION V R Filby oo i o o o o 8 CENTRAL COMPUTER COMPLEX IN 1970-19805 A FEW THOUGHTS ON THE N S A LINGUIST Cecil Phillips o o o i o o o o o S Anon oo 15 'j BOOKBREAKERS FORUM o o o MORE ABOUT THE N S A SIGINT SUMMARy LETTERS TO THE EDITOR I ' 16 o o 17 19 TillS BO JYMBNT JONTl INS EUROBBWORB MA'fHRIAJ g e ll BIIlNS2 SIIEURS8 NSAj eSSM 1 2' -IOP--5ECRH- tlxempt frem fiBS 139 116 1 Sate 1 DecI '11r lItil iIlB It till QRPIate Declassified and Approved for Release by NSA on '10-'1 '1- 20'1 2 pursuant to E O '135 26 vl DR Case # 54778 DOCID 4019647 TOP SECRET Published Monthly by PI Techniques and Standards for the Personnel of Operations MARCH 1977 VOL IV No 3 WILLIAM LUTWINIAK PUBLISHER BOARD OF EDITORS Editor in Chief Collection '1 Arthur J Salemme 56425 1I955S Cryptanalysis J 80255 oooo HHHHH oo HHH Language EmeryW Tetrault 52365 Machine Support 1 133215 Mathematics Reed Dawson 39575 Special Research o Vera R Filby 71195 Traffic Analysis Frederic O Mason Jr 41425 Production Manager Harry Goff 49985 For individual subscriptions send name and organizational designator to CRYPTOLOG PI TOP SECRET P L oo 86-36 DOC I D 4 01 9 64 7 S CR T L 86-36 EO 1 4 c F L 86-36 86-36 March 77 CRYPTOLOG Page 1 SI3CRI3'f EO 1 4 c P L 86-36 EJCJC2I1D EEC 1 4 C 86 36 March 77 Page 2 7 3 30 1 4 - L 86 36 DOCID 4019647 March 77 Page 3 DOCID 4019647 fO 1 4 c oLo 86-36 0 SECRET i March 77 CRYPTOLOG Page 4 seCRE'f 0 130cm 4019647 P22 L 8 6 36 - March 77 Page 5 DOCID L -- 4019647 SEEURRE'f 1 ---l March 77 CRYPTOLOG Page 6 SEEURRET EO 1 4 c P L 86-36 F L DOCID 86-36 4019647 EO 1 4 c P 1 86-36 TOP SECRE'f' UMBRA IR51 March 77 CRYPTOLOG Page 7 TOP SBCRB'f' UMBRA P L 86-36 EO 1 4 c DL 86-36 1 4 e DOCID 4019647 TOP SEEURRBCf UMBRlt 'fep SE6RH HII8Il1't S $ REVISED TECHNICAL SIGINT PUBLICATION MANUAL IN PREPARATION V R Filby E12 t is well known that not all analysts everywhere get to see the formal instructions and other documents that affect them in their work For those who may not find out about it through normal channels here is news of a document all analysts and reporters should take the trouble to get hold of and study The revised and updated Technical SIGINT Publication Manual U to replace the present TSPM dated 19 August 1969 is in final draft form and should be ready for distribution in mid1977 It will be promulgated by USSID 200 6Q IFlQ NTIPb 1bt' IQb VIA 6QIIHIT 61bt' INHe Q lbJPY 55 s March 77 CRYPTOLOG Page 8 TOP 8BCRBT UMBRl 55 EO 1 4 c P L 86-36 DOCID 4019647 Reprinted in entirety from C- INERS VoL 3 No 10 Issue No 31 Final 1976 SC rhe most dramatic change in SIGINT computer processing in the next few years is almost certain to come from the much wider development of man-machine interactive processes This does not necessarily preclude growth in regular batch processing or growth in fully automatic processes but it certainly means a strong shift of emphasis in the development of computer processing systems Some of this emphasis comes from a desire to explore the newfound capability of interactive computing but hopefully most of it comes from a realization that machines alone are not nearly as effective as man and machine working closely together These views are presumably not unique to the SIGINT process since comparable growth in the use and sale of interactive computer terminals is being experienced in all kinds of pursuits Interactive computing in the SIGlNT case can mean a variety of things It can mean the ability of an analyst to adjust the collection process while it is going on It can mean the ability to scan some output in real time and adjust it as necessary Or in an analytic process it may give a person the ability to correlate and cross-check new data against the mass of historical information one has collected before or the ability to prepare reports gists and similar inputs to other processes It is clear that there are a number of ingredients to a good interactive processing system One of the most critical of these ingredients is a knowledgeable and well-trained cadre of users This group is already growing but more specific action will be necessary in the future to insure that we use the full potential of the systems which are built t the same time the central computing service must do everything it can to make user interfaces simpler rather than assume that every user has some skill and knowledge as a programmer In addition to making computer systems easier to use the central computing service must provide a number of capabilities and tools for the users in much the same sense in which utilities are provided to homes and businesses In a practical computer sense we see this utility as o an adequate set of computer terminals to serve all users who need access to the system and simplified user procedures o on-line mass data storage systems and data management software which will minimize the complexity of the user interface o a computer network connecting all SIGINT locations which will enable us to put the required computing power into a connected system and provide user access to any part of the system from any terminal and o individual computers and devices and appropriate software which will provide the best computational capability available Some of the details of current plans and broad concepts of usage for parts of the system are described below This detail will be added to and modified as necessary to reflect requireuuuu ments developed underluuuuu The overalll plan which is basically the plan for the central computing facility and externally connected computer nodes is due to be completed in draft early in 1977 Computer Terminals and Terminal Access Systems The largest group of computer terminals for SIGINT operations will consist of alpha-numeric I March 77 CRYPTOLOG Page 9 SECRE'f ILtNBhB VIA CO IHJ't' ellAhNf3LS 6TJLi P L 86-36 DOClD 4019647 SEEURREtF displays and associated keyboard for analyst interaction with the system By 1980 or sooner computer terminal density should have reached the point where there is at least a terminal for every three or four production personnel This should mean that almost every analyst does part of his work at a computer terminal A smaller set of terminals will have full graphics capability in addition to alpha-numeric capability that is they will be able to plot and to produce line drawings through the use of vectors A still smaller number of remote batch terminals will continue to be used to initiate batch processes and to receive printed output It is probable that as the use of interactive terminals grows remote job entry terminals will become more basically remote job output terminals or simply remote printers Softcopy displays that is writing output to a computer file for subsequent examination on a terminal screen will probably replace some print but it is not too likely that all printing will be replaced In many cases line printers will be associated with display terminals so that output can be recorded Equally local storage cassettes may be used as safeguards against loss of data during entry and as a way of insuring the ability to input data at all times With present technology and probably within the foreseeable future the main bulk of the interactive terminals will be attached to minicomputers as terminal concentrators and through these will be attached to the network and will have potential access to all of the system resources attached to the network The actual access will be limited by absolute security constraints and by need-to-know procedures These will be effected by both hardware and software means Hardware controls are certain to be used to limit some terminals and may be used to identify users through badge readers or the like Software controls will consists of authorization tables and passwords which will control a user's access both to files and to processes The total extent of the work to be performed at the terminal level or at the terminal concentrator is not yet determined However it seems most likely that with present terminals efforts will be made to limit them to universal functions These will amount to data entry and data editing functions dealing with lines or screens full of data and probably to the generation of standard protocols for network entry and accesS to a specific system If the terminal intell igence actually storage and processing power grows a great deal as most manufacturers predict then this power may be used to accomplish procedure and data language translations or to provide the capability to do other functions common to a group or class of users At the next or terminal concentrator level more advanced functions will be per- formed than those at the terminal level With controls and with standard principles applied so tware in the concentrator may perform machine e ltlng and automatic formatting and provide hIe-browsing capabili ty Where the primary function is data entry the concentrator will also serve as a fail soft capability to allow data entry to continue until the main system becomes operational again Work not done or not possible in the terminals or the concentrators will be done in the main systems or hosts Communication between concentrators and the hosts will be via the computer network described below To a considerable extent hosts will be specialized processors as they are now If network protocols and individual host protocols and language can be simplified for the user by translation and creation in the terminals and concentrators it may result in even more task specialization by the hosts Alternative views are that networking of host computers may result in automatic load leveling of tasks by automatic distribution of them throughout the network In the next few years the latter seems practically attainable only for mUltiple copies of the same computer within the network In general the network will contain multiple copies of hosts for reliability and backup and for achieving the desired processing power All of the description above pertains to user-to-computer communications through terminals concentrators and computer hosts natural parts of a man-machine process A quite separate kind of communication capability will exist within the network to support person-to-person communications With this capability analyst-to-analyst discussions such as are now handled by OPSCOMMs would be straightforward A mail box ' variation of this capability would permit messages to be left by one analyst for another when direct contact is ot possible because of working time differences etc By signaling his mailbox an analyst receives all his accumulated messages from other analysts or computer generated messages to advise him of the status of some piece of work Data Storage and Data Management In discussing a processing system for SIGINT operations we are certainly talking about a system serving several thousand people -- probably more than 10 000 when the operational aspects of field sites are considered Even if terminal density is as low as one terminal for each ten members of the operational community it will clearly be a very large system with many terminal systems and many hosts In all probability terminal density will be much higher particularly so if MAROON SHIELD concepts extend to all sites To support such a large community of users from interactive terminals it will be extremely important to have a large amount if not all of the recent data on-line While it March 77 CRYPTOLOG Page 10 SECRET Ih NBbE VIA 88 IIPJ'f EURIIANNBbS 8fihY DOCIO 4019647 SECRE'F is not out of the question to have a single large central store for all data recent experience would seem to indicate that distributed data collections are much more likely Some systems will have much more storage than others because of the historical nature of the kinds of problems handled and because the primary impetus in developing and marketing very large mass store systems is toward IBM systems Thus it is likely that we will have unevenly distributed data with more than half of the data in a single large collection Examples of this uneven nature of data distribution are quite pronounced now and will be greatly exaggerated when the IBM storage system called OAK is added to the present IBM 370 168 complex With its 169 billion byte storage and virtual disk concepts it is bound to lead to a large imbalance in storage of data throughout the general purpose system In spite of the imbalance there will be a continuing need for large local collections of data on other hosts in the central complex in order to satisfy specific requirements for rapid response and other special requirements In addition there will be requirements for data correlation-fusing-updating and other functions inter-host For this latter case and for inter-host inter-process operations a great deal of data management is needed We feel that a number of things will be necessary in order to deal with distributed data First some kind of data management plan will be desirable so that data will not be distributed helter-skelter over several storage devices Such a plan should address necessary and unnecessary redundancy It should also address the software tools which are to be provided to enable users data base administrators and others -- as well as programs and computer systems -- to have convenient access to the data Some work has been done in NSA toward developing a data management plan DAMAP but clearly a great deal more work remains to be done especially if anyone is to make regular use of distributed data bases To a great extent we will not have very much experience with trying to use distributed data bases until the computer network linking computer hosts is operational As far as data base management systems DBMS are concerned these are key tools for handling data Hand in hand with DBMS usage there must be a growing awareness in user organizations of the need for individuals who know the data base and control its development -- generally referred to as data base administrators The data base administrator and his management have the potential to make data base operations useful and productive In their efforts to do so it is imperative that they achieve standardization of data elements standardization of data codes and standardization of files in the sense of requiring basic data elements in there Data dictionaries and data directories will be key tools for data base administrators and others who face the job of producing and managing the data bases In achieving the SIGINT operations process1ng system data base management techniques W ll be guided by a unified systems discipline w1th a goal of a single data base management system Standardization will play a much larger role than in the past Until adequate data standards are developed existing data management will be continued but its future growth will be limited The goal is a coherent whole for managing data supported by a comprehensive set of data management tools Experience with networks such as COINS points to the critical need for accelerated work in NSA's data standards Emphasis should be on NSA or SIGINT data standards rather than large efforts at meeting non-SIGINT standards T es may be im tant in some 'case but many of them have little bearing on SIGINT problems Further SIGINT data standards work should concentrate on the most frequently used data elements and types so that the benefits are maximum Standardizing data elements which are rarely used or used in narrow areas of SIGINT is useful but it does not contribute as much to the areas of broad use A final note on data standardization implementation -- a phase more difficult than developing the standard This must be vigorously but rationally pursued with NSA management taking active interest in the arguments and counterarguments Good standards are usable standards Computer Networking Another key ingredient of a large interactive system such as that required by SIGINT operations is a way of linking computer systems The scope of SIGINT processing exceeded a single large computer many years ago In the interim specific computer-to-computer linking and loose coupling of several large computers through shared storage have helped with the problem projectl NSA' S computer p L 86-36 network1ng plan 1S the generalized solution to linking as many systems as necessary into an integrated processing complex I rai smce 1Jm - S b nn nmn- mpn L niques for linking computers and terminals with standard connection protocols It does not of itself guarantee standard processes standard data bases or any other set of standards except for the basic network protocols NSA's computer networking is based on technology developed for the Department of Defense Advanced Research Projects Agency ARPA It is a packet switched network in which every node is connected to at least two other nodes This gives additional reliability to inter-host connections As mentioned above the use of these March 77 CRYPTOLOG Page II SBEURRE'f' II zNBbB VIA 88 lnJ 8Ih HJ8LS l3NL'l 86-36 DOCID 4019647 SECRET connections is not controlled except by the net- will reduce the demand to have everything work protocols network control programs in the together but from a practical point of view this will take some time perhaps 2 to 3 or hosts and the provisions for inter-process more years communication As the network comes into use some additional controls will corne into use Security will be one of these Job process and message accounting will add another dimension of control The details of an overall control or management system are yet to be developed but will be formed from the experience of using the network and from growing user requirements for process control or process management The requirement for process management by users is developing rapidly in parallel with the ability of processes and processors to be linked together Effective use of the network will require extensive user knowledge of computer processes and computer facilities as well as knowledge of the data mentioned earlier A network information center is being developed to assist the users in understanding all the capabilities and complexities of the network 11deV 10pment6fthe computer network will take a number of years Its efective use is a key factor in the development of the Agency's potential to use computers The Equipment Complexes Equipment complexes or sometimes computing complexes is a term which has come into increasing use as we couple two or more pieces of quipment or link equipments through some cOlllllloJi peripheral device The use of equlpment mplexes for planning and management has been accelerated by the need for operator and maintenance economies and by the practicalities of consistent operation of dentical and related systems The principaJexample of such an equipment comp ex jl the main IBM 370 complex now called By developing the large IBM systems so that they share disk storage a significant improvement was obtained in the flexibility of the four systems At the same time it was possible to reduce the staff necessary to operate the machines In other cases we can also reduce maintenance costs by collocation of machines I Immediate plans for the large IBM 370 systems are to add a fifth IBM 370 168 to the present four systems and to create one multiprocessor system It is planned that the multiprocessor will serve needs which demand high reliability or availability This will include immediate front-end processing of field data arriving from I I and support of critical terminals The whole 370 168 complex along with the OAK mass storage system of 169 Ihj J j gn bytes of gr rs 1 rwi c i The present IBM 370 158 serving primarily administrative functions will be upgraded t an IBM 370 168 and formed into the larger complex by loose coupling through shared storagea nd by linkage to the computernetworL The pres ent IBM370j158 used for Information Stora ge and Retrieval IS R will serve principally as a developmental system and will beused in background mode and in nondevetopmental times for large batch processes Future large IBM systems or IBM-based systems such as AMDAHL which may have to eadded as supply capacity will be coupled into the main complex The linking of lIOre systems as multiprocessors wi11 be dependent on the success of tne MP system planned for late 1976 In order to meet DOD's goal by bringing time-critical and batch processes together the IBM 370 complex will have to be made t 6 respond much more rapidlY so that bulk data can be vailable as selected data such as r orts In fat i likely that and the I BM m -' ''rT''''l' '' ' -- t '' T'T- ' ' 'ri er to accomplish this The Time CPi #ea l Comp lex This complex is also expected to continue to grow rapidly The succes s of the TIDE sys tem ove r the last 6 or 7 years hasproducedt le present overload state of the pairiof UNIVAC 4945 and has led to the development of PREFACE plans I As planned the UNIVAC 1110s'ln tWill be used to relieve r t o f processing functions The Central Data Processing complexl as distinguished communications functions While converting processes from the U 494 to This complex is the historical successor of more than 40 years of automatic data processing the U 1110 a data base management system ADP which began with IBM punched card machines DBMS will be introduced to provide for more orderly handling of 1ata anQ ito simplify future in the 19305 As noted earlier it is the programming As the development proprocessing area where the most data is stored ceeds it will be necessary to address the and where the input and output at least of broad question of alpha-numeric terminalS for traditional data records greatly exceeds any other set of NSA computers Because of storage NSOC and to begin development of plans for a It capacity available in this complex and because communications su to replace of increasing requirements for data correlation is estimated that r Jill continue to operate this complex continues to grow at a considerable until at least January 1980 and possibly a couple of years after that pace It is possible that computer networking I I c J March 77 CRYPTOLOG Page 12 8ECKiT IIAN8h13 VI G8l IlN'f' 6IIANNEbS 8NbY IP L 86-36 DOCID 4019647 P L 86-36 8 CK T - -__ Iwhichi S no sI tedtoserve b th full graphics nee s and so e alpha num ic needs will be diverted to fullgrapl1ics oIlly usage as Ibecomes opel'ational Acare ful study of all fUll graphicsneedswin e undertaken with a goa1 of developing a set of requirements for full graphics While these requirements will ultimately drive the full graphics solution it is likely that I will fill most of these needs over the next 2 to 4 years The High Compute Complex This compleX of comput rs is now almost entirely synonymous with cryptanalysis To some extent this is because the cryptanalysts have carefully guarded it and to some extent it is because thel let of CDC 6000s preceded the first CDC 6600 for cryptanalysis 10 al computer and more complex questions could be relayedto h MC T i s jde is al ready projected inl applications where the broaunc lona sImIlarIty to Morse ollection is almost identical except for tuhiIlg The presence of high bit rate signals at operating facilities may dictate the need for a special network pr data bus paralleling the r study in R as a part ofl I and other projects It is most likely thatt Iwould be used for control information LllU LlldL Lhe parallel high-capacity linkage would be reserved for only signals This would limit the number of required nodes etails of these connections to the network and the interaction of operating aCl lIes with the Main Computer Complex are As far as he cryptanalytic side of the prob- emerging as the projects develop It is lem is concerned the high-compute-complex plans reasonable to expect that requirements gathered are described under aC Group concept called forr Jwill provide more understanding of HYPERCAN -- for HIgh PERformance CryptANalysis what is needed Under th1 s concept four CDC 600s using the 86-36 P L NSA IDA Operating System illbe linked by a 50 megabit data transmission bus In addition EO 4 c a new and more advanced super computer JWill be sought to add to the capacity 01 me conr lex We hope to add this new capacity in about 2 years and thereby double the straight-compute capacity Interactive terminals will not be developed at the output for the new system Instead access to the new system will be primarily through the CDC 7600s and the 50 megabit network Thi is expected to optimize the use of available resources and to minimize the impact on users by retaining the present NSA IDASYS procedures and languages I l 11 '''' I Operating Facilities I l J The term operating facility is being used as a catch-all for the NSA end of remoted collection operations Presumably everyone of these operations needs some interconnection to the main computer complex To date most of the interconnection has been describe as a connection to the computer network --L __ with little more specifics than that Almost without exception design of these complexes has ignored the necessary and desirable interaction with existing data collections on the central computers In viewing overall operations processing in the next few years it seems certain that more complete plans will have to be developed using existing resources as well as the new equipment One mode of operation which may emerge is to limit operating facility computers to problems requiring extremely quick response and to use the main computing complex MCC for basic support In this mode the MCC would update basic files in the operating facilities by downloading from the central analytic files In the operating facility immediate response problems would be dealt with by the March 77 CRYPTOLOG Page 13 SBCK T IIJd J8 bB VIA t eOr lnJ 8I1ANNBbS SNbY DOCID 4019647 III P L 86-36 EO 1 4 c SECRE'f --------------- Mar h 77 o CRYPTOLOG o Pago 14 1I ' JBbE VIA EQPbII JT EIlA JNEb S QN15Y DOCI-D 4019647 CONJi'IBBN'f'IAIJ t 1 L Anon 86-36 both linguistic and factual information Obviam being somewhat arbitrary in using ously there is no such person It should come the term linguist What 1 mean here is an as no surprise then that the N5A linguist at individual who is capable of performing the any level is apt to run into problems that infull range of tasks required at NSA in which volve either language or factual knowledge or knowledge of a foreign lang age is primary both outside his own experience not covered Let's forget all hyphenated job titles in in any of his training courses or anticipated which some degree of language knowledge is by the latest crop of scientific linguists coupled with some other skill You can carry and not to be found in any reference works this hyphen at ion business to t le point where available to him If we agree that it is it becomes difficult to locate anyone who b i doesn't claim to use a foreign language to some extremel im or ant in u wo if extent in his or her specialty It sometimes 1 m -' 'rm I' 'l''''''''''I' '' ''' ''''1 --- l' ''''' ' ' '' 'e n - c a n''''''Y'''o''''u'''''''r'''e''''a''''s''''o Jna bI Y seems that the pretensions of such people are linguist What choices do you in direct proportion to their ignorance of the offer him Must he confine himself strictly to language claimed I he level of his job description and pay and Nonlinguists especially have a curious o Ignore everything else Or should he try to tendency to treat language jobs as though we at cope with everything that flows across his desk the Agency were in complete control of infloweven though many of the problems linguistic and ing language materials Thus a particular otherwise will be beyond his competence language task in a very low grade of course Backing off for a moment let's consider is described as highly stereotyped low-level again the first question raised what is a re uiring no more than a minimum lanQ uaQ e knowlinguist anyway Is he a person who can proledge I EO 1 4 c vide a precise description of a language emP L 86-36 ploying all the scientific terminology currently in vogue Or is he someone who can use the language as an educated native would Is it realistic to assume that the personViho can describe a foreign language can also use it in the manner indicated Probably not Of the two which one does the Agency's work -- the describer or the practitioner J So what is a linguist expected to know -everything It would be a miracle indeed to discover a native of this country who has mastered his own language including all the words in the dictionary and besides that knows all the subjects that can be discussed in English in short a walking encyclop dia of I am reminded here of the never-ending debate over the term bookbreaker Is he or she the p r son who breaks into a code and provide a tentative description of its size and s tI'ucture - - with perhaps a few hypotheti cal recoveries Or is the bookbreaker the bne who actually reconstructs the code book to the point where incoming messages can be decoded and translated upon receipt 1 would go March 77 CRYPTOLOG Page 15 CO FID NTIAI P L 86-36 WtlvBbI'J Ii' eaMIH'I' EUR1IA1H4l3b8 SHb' -bbcID 4019647 EURONFIBBN'I'IAt even further and insist that the reconstructor should upon request be capable of translating any decoded messa e -- and have his translation survive a language check by the most experienced linguist in the area whc may not be friendly A demand like this tends to divide analysts into two groups the describers and the reconstructors These two groups will disagree forever as to who is more deserving of the title bookbreaker but that isn't the point Their great debate is really more over status than anything else In my own mind I expect as much of the code reconstructor the one who meets the standard imposed above as I would of the true NSA linguist as defined at the outset Neither one has the right to assume that the task at hand will be tailored to fit his own limi ta tions or narrow interests To begin with both must be literate in the foreign language involved literate as defined in Webster's -- able to read and write I o o o o o o o o o O IHQElrFl tJ EEB ooo ooooo ooo ooo o oooooo o oooooo oo o ooo o ooooooooo o oooo ooo o o oo oooo o o oo ooo ooo oooooooooooo oo oo oo o ooooo oooo o o o o o o o o o 10M P 16 eo o o o o o e-e o o o o o e o o o o o o o o o o o o o o ' 10 LI NG UIS TS AND ffi YP TA NA LYS 15 INFO PROGRAMMERS AND TEAM CHIEFS m TE eM PR I OR 1 TY FLASH THE BOOK BREAK -lNG -ER TAK E 3R D MEA NI NG PR EV WOR D PLURA L 6 91 6 roR UM 36 MEET 8691 PREV VERB FUTURE 7t 45 9098 STOP 663 6 FOR 4871 INFORMATION 2231 51 48 BEGI N SP ELL 1745 0944 8412 2910 3337 Why make such a big thing about mere literaSimply to plant our feet firmly on the ground in this whole matter We shall never get the linguistic paragon who knows all the answers offhand nor can we afford the risk of depending forever on those who never know enough The real NSA linguist is the person who is literate to begin with -- and who has an infinite capacity for growth To quote Will Rogers We are all dumb on certain subjects and this description fits linguists like a P L 86-36 glove But with the ability to read a foreign language at sight we can overcome our ignoranc'e - by searching for the answers in the very places where a native scholar would look just as we regularly consult standard reference works in English If we still can't find the answers the trouble may be that the other fellow's reference sources are better than ours cy o I 4320 SPELJ-'o 8297 2121 2133 4631 o o o o o o eCOMl'IA TELEPHONE 15 16 5868 14 9374 12 1313 SEC UR E 8297 eCOPll'lA 5148 BEGl N SPELL 89 14 1 1 43 8176 4631 9098 3697 4994 -P eFl LLER o EN D SP ELL It 16 STOP PLEASE RELAY TO lNTER Jl -JN L D 2901 KI NG lOG 671 PAR TY 3397 PREV WOR D PLlfiA L 9098 STOP EDT March 77 CRYPTOLOG Page 16 CONFIBHN'fIAL IL l JQbE VI 8Sf tItJ'T' 81lAtJPJl3bS 8HbJPY DOClD 4019647 CONPIBI3N'fIAIl into the reporting business was the establishment of the NSA SIGINT Command Center successor to the PROD Watch Office and predecessor to our present NSOC At some point during this time it must have seemed logical to produce a daily summary of SIGINT highlights for our customers and for NSA executives And so the NSA SIGINT Daily Summary SDS ---the immediate predecessor of the NSA SIGINT Summary -- was born The SDS was divided into three sections MORE 1 P L 86-36 The year 1965 The pZace the NSA SIGINT COl7'lJnmd Center The occasion the beginning of the NSA SIGINT Swrmary And you are there Sorry WaUer r I I SCJIhe May 1976 issue of CRYPTOLOG contained an article by William Hunt SA DDF which described in some detail the functions and purposes of the NSA SIGINT Summary On the whole it was rather complete and quite interesting But for the period between the inception of the SIGSUM and Bill's association with it -- a period of B or so years -- the article provided virtually no information at all Having with some others devoted a part of those years to the business of conceiving and developing the SIGSUM I am perhaps in as good a position as any to fill in the gaps In the process I will also clarify some statements in the article that are unintentionally misleading made so simply by the fact that Bill had no part in the conception and early development of the report and therefore could not be expected to know how it all began In the early 1960s a number of events occurred which taken together had a far-reaching impact on the role NSA was then playing in the U S intelligence community Where prior to that time the role of the SIGINT establishment had been mainly that of collector and proces sor from the Cuban missile crisis onward we experienced a marked increase in requirements for end-product reporting of SIGINT developments Our customers wanted SIGINT information to be put in its true persp ctive At the same time of course we were reninded to avoid even the appearance of producing finished intelligence for which we were not equipped technically or statutorily Coincident with the entry of NSA o World Highlights mostly G Group items o Red Extract A Group items and o Gold Extract B Group items Items were carried to the Command Center during the morning and early afternoon edited there and delivered to the FLEXROOM in the early evening hours for hard-copy and electrical preparation Later in the evening the electrical version would go out over the wires to a worldwide distribution which is essentially the same as that which the SIGSUM uses today The hard copy most of the time would be delivered to its Washington-area customers early the next morning I say most of the time because 1 can recall one or two times when its delivery was not so early and some of our customers let us know about it The SDS had some shortcomings The quality of the items it printed was to put it charitably inconsistent Items were often submitted in h dwritten form nd had to be deciphered the idea wa6 not pa66 onately emb aeed at Lea6t not 4ight away and typed before they could be edited Most serious was the virtual absence of any input to the Gold Extract The reason for that was that P L 86-36 putout its owndailysufuii'iafja rid theSDS was not only competition for it but since a EO 1 4 c large percentage of their respective audiences were the same there was the problem of redundancy For these reasons chiefly the then deputy chief of the Command Center 1 k sent me on a mission to each of the PRoD Group chiefs to appeal for better support for the 5DS The result of the conversations which then took place was a decision to scrap the 5DS in faygr P L 86-36 of the daily report now called as it wa sthen the NSA SIGINT Summary migh SUpp e Contrary to what one the idea was not passionately em9raced at least not right away i arid some others in the Command Center had reservations about the size of che effort that would be necessary And there was some concern expressed that we would Iget flak from CIA OIA and State on the ground that we infringing on their prerogatives But the SIGSUM was clearly an idea whose time had I March 77 CRYPTOLOG Page 17 U PIQr ii VI' 6QIlHI'I' ElIAtIfIECS eNb ' DOCID 4019647 P L come All the Group chiefs and Chief P05 were strongly in favor of a first-class daily SIGINT summary report in fact had een fQr s o fl e J ime and persuaded the then ADP J that the time to strike was at han It took about 2 months to do the preliminary planning for the SIGSUM The6riginal team that was responsible fO rdeveloping the format and layQut consisteclof two people -- Dave Cossum and me Warner Parsons SOQn joined us as visual-aid cClordinator and I land c J I Iwere the principal artists whQ developed the cover design and page formats I_ 86-36 clarity anddirectness by General Carter in a le Herto DIRDIA General Carroll in early 1966 and e it wqs fUlly developed within a year after it fir t appeared The reader might have alsQ concluded if he has read this far that it is indeed both possible and 7 ' frQm my point of view at least -- desirable to trace the SIGSUM's growing -pains By mutual agreement with the Group Chiefs the SIGSUM was put together every day by an editQrial board on which each Group had repreAt this point it should be made clear that ntati n In addition a CREF representative the SIGSUM had no official status as yet Mr Miss Ruth Schley prOVided cQllateral and L Ilhild only given his approval to explore scientific and technical information support the ways and means of upgrading NSA daily-summary and Warner Parsons served as art editor reporting and to prepare a mockup a printer's EO 1 4 y of our proposedpubli at n He would then decide between two alternative seeK tne ave Director's apprQval Qr shelve the idea In the 1 or-1n- 1e and I was designated course of the exploration some amusing situation developed The Qne that I remember most clearly his alternate The hard decisions which became the basic perating philosophy of the SIGSUM had to do with estimating the number Qf maps We were hammered out by this grQup often in the would have to keep on hand to cover all reasonlate hours of the evenlln One of the Direcably prQjected reporting situations for a given tor's briefers period of time -- 2 months I think Our plan 1L l alsQ satL o-1 n-Q n -- was to place the resulting order with a MI' I lat CIA with whom we had a connection After considerable postulation speculation and P L Among he many lauda o y wi e6 multiplication Warner arrived at a n ber that none of us including him believed Our as ec eived a6te he in 6 i utiC ft 06 tonishment was quickly exceeded however by the uproa ious lau hter the number evoked in 06 the SIGSUM 1 pa ieula ly our connection then Chief Qf the Geographic Branc RE now CS The ec all ho 6 e 6 ent by the Sec eta number was 1 500 000 maps Incredible as it may seem the figure proved to be very near 06 State CINCSAt and CINCPAC the mark group of people and the others who followed them yticularly Warren Keniston deserve an equal share of a n y c r e a l t that is to be given for the SIGSUM The result of all their efforts was a first-class intelligence report that was widely praised as being a worthy addition to the daily reports that were already available to the government's highest-level decision-makers Among the many laudatory wires received I particularly recall those sent by the Secretary of State CINCSAC and CINCPAC We were also told unofficially that it was warrn received at the White House a number of copies with Walt Rostow's marginal notes very much in evidence were mysteriously returned to us for disposition Needless to say we were all both pleased and relieved to have the SIGSUM so highly regarded especially by some who might have considered it competition with their own daily summaries The SIGSUM had received its baptism of fire so to speak and had come through it unscathed I L __ __ r- At this point I must take issue w1th Bill on three points he made in his article and clarify the situation by emphasizing the following true statements e the SIGSUM was not conceived and designed by the Assistant Director for Production e there was indeed a clear-cut requirement for it which was set forth with admirable March 77 CRYPTOLOG Page 18 EO 1 4 c P L 86-36 CON'FIBENCfIAL H llJI li HI' QHnr r IWINH6 8 1f Y 86-36 DOCID 4019647 SECRET Letters to the Editor I ton with _ 'i le le l r' that amazed me _ nn '1 in Y _ I we were one of t e people there T 1S was the story Under normal circumstances I would have exercised my prerogative to ditch C-LINER's Impure Mathematics Final 1976 issue in the open trash receptacle where it belongs It clearly has no SIGINT value no intellectual value and in my judgment not a modicum of socially redeeming value No wonder the author chose not to put his name to it But I can't do that because the thing was published in an official government document under an official SECRET Codeword classification -- circumstances which could hardly be considered normal Hence the burn bag To the Editor CRYPTOLOG I had a scare the other day but it all turned out okay and I thought that you and some of your readers might find amusement and r'ru i n in To the Editor CRYPTOLOG I I feel reasonably confident that C-LINER in taking upon itself this one degree of freedom too a y as offended the moral and profession-al sens1t1V1t1es of other people around the agency as it did mine It is inconceivable thattheaveragemiSsi6ri 6 deri1 edNSAe would P L give assent to the seepage of this type of writing into SIGINT literature Where is the sorely-needed NSA CSS voice of authority to say Yes As a society we have indeed mortgaged our self-respect to the likes of Hefner and Guccione But this is where we draw the line EO 1 4 c P t 86-36 I ri Had we been missing an obvious bet How could such a scheme possibly work We had to find out more about it Hopefully we will not have to wait too long before the Director's voice is heard in this regard As for C-LINER itself all things considered ma be it is just as well that it's now a thing of the past Appalled name withheld at writer's Copy to request Chief C Mr Speierman Director NSA CSS General -Allen UNCLASSIFIED Editor's repZy In the final issue of C-LINERS the second reincarnation cycle of the C Group Machine Processing Information Bulletin its editor David J Williams urged potential authors to send their articles to CRYPTOLOG He also stated If you disagree with any of the materials in this issue you can carryon the fight in CRYPTOLOG I am sure that Arthur will L ----------------------- be happy to carry your rebuttal 1 ' m pleased Without denying that this scheme which has that someone took up Dave's second suggestion proven to be very successful is an example of Wouldn't it have been nice if someone had a simple yet ingenious solution to a problem of taken up his first suggestion that quickly long standing I am relieved If it were possible to train a linguistically naive person to Dave explains the lack of an author's name listen to any language and hear specific words as follows the item was supposed to have a headline Tales from the Past or Golden imbedded in conversations some basic ideas would have to be changed and I am too old to Oldie but the headline got dropped during r page composition The general feeling was that s t ar t ove Jack Gurin R5 the item was in the public domain and m_igh SEERE'f ee J have been written by someone lin the O 1 4 c early 1950s Dave says that he is highlyP L 86-36 I I I EO 1 4 c EO 1 4 d P L 86-36 March 77 CRYPTOLOG Page 19 SECRET II e NQbE HIA QQ lINY SIIANNE158 QNbY 86-36 DOCID 4019647 UNCLASSIFIED amused at the reaction it evoked from you because when the C-LINERS editorial board was considering it for publication they showed it to several professional mathematicians All fel t that it was a cute b professionally sound and c definitely worth printing With a consensus like that how could Dave have failed to print it UNCLASSIFIED d we are grossly overpaying linguists at the lower end of the GG scale while grossly underpaying them at the high end Well sir that is a most extraordinary group of suggestions and a group that I mostly disagree with Let me start at the beginning a I expect thatl Imeant to imply some degree of quality in recommending that the desk linguist be included in the 1 5-18 category But does he really expect that someone in this To the Editor CRYPTOLOG agency can or will come up with language analyst billets in that grade range I don't think so Thel li llterview in the December issue of CRYPTOLOG is outstanding I also think And I think the rationale behind such a decision would be that he isn't likely to be worth that that the interview idea is a brilliant one much money Now that's a hard cold look at it There are undoubtedly more than a few of our people at various levels of the management chain but it's probably closer to the truth than any and also more than a few of the stalwart eXperts C O SC manual that allows 9-18 as the a th Of the who would be amenable to sharing their views ex language analyst Sincel did not indicate that he was discussing multi linguists periences and expertise in this format The I lint r Tiew provided an insight into or multi skilled people I am assuming that he is includirigthe highly qualified singlea story that may never have seen the light of day had it not been for CRYPTOLOG Truly SIGINT foreign-language certified language or voice language analyst in his numbers Ifhe honestly journalism at its finest belieyes that such an analyst is going to make I am in the process of collecting information it beyond 12 or 13 without derrion trating a for a couple of articles which I promised you skill beyond turningforeignsounds0l'words some time ago I hope to surprise you one day into English sounds or words using a ski lin and even complete them a single foreign language -- well I justdon' Keep up the outstanding work knowwhatTIleasureheisusing e p L f -------'IG95 b A few years ag IWOUldha y eagr dfhat the upperJevels of agellcYmanageme t wereat Editor's repZy best unrecepti Te to t heidea of Jl Fomoting 1 in Thanks for the kind words We're glad you guists Jnojonger believe llat I have seen like the interview fOrmaL and we hope to use it tooma promotions in the ast couple of years again roon someone else who liked the I I t o the 12-15 range and while lam certain that 1 interview as well a the Kathy Rio kll md there are indi vidu Jcases of inequity I would article on the SR test was _ Jof the strongly disagr that the re are any grounds SR Career Panel He requested 50 extra copies 0 for a classa ction Our activist language the December issue and intends to give them to panel WOllltlrio doubt be offended by the suggescurrent and future SR interns for mandatory tion1 at it did not know what was going pn in reading of both items th la nguage world Perhapsl lis not aware that current hiring plans for new agency _____________________________ _U_N_C_L_A_S_S_I_F_I_E_D_ -Ipersonnel are almost totally devoted to the acquisition of language technicians r-I To the Editor CRYPTOLOG In response tolr-------------- Iarticle Let's Give Linguists a Bigger Pie ce of the Pie CRYPTOLOG December 1976 I say No more pie 'til you eat yOlJrspinach la ticle states or implies the following suggestions I a some sort of quota system should be implemented that would allow the distribution of GG 15-18 personnel in this agency to include x desk linguists b various levels of agency management are unaware of the problem as perceived by 1 1 c an advanced degree in language is prima facie evidence that we can expect superior performance as an operational linguist eel The issue of advanced degrees in language I think has been largely resolved by several developments a general lack of availability brought about by an overall reduction of language majors at many universities a requirement to retrain language majors in real-world applications of the language they studied in the academic world -- and all too frequently this training has been at the very basic level hence expensive and last but not least the average grad structure problem which says that we got to hire more people at lower grades do less promoting above grade 12 and perhaps most important keep the language analyst in that job -- the job on which his paycheck is quite likely to depend Again I realize that it sounds harsh but the times are changing While the agency might like to help the language analyst to move into a different March 77 CRYPTOLOG Page 20 UNCLASSIFIED 86-36 DOCID 4019647 UNCLASSIFIED 1 1 field and very often did that in the past 1 would expect to see a tightening of the belt as more and more interest is shown in the problems we face Ironic isn't it In the years when those language problems were backburnered we always kind of assumed that when the linguist got the spotlight he would just zoom ahead so fast we wouldn't be able to keep up with him Now that he is in that spotlight -- and believe me he is there -- we are going to see I expect more actions similar to the recent personnel decision that dropped the entry level for linguists and others higher standards for language professionalization and tighter controls on the language field in general FYI our language-hire program for the next couple of years will probably fo us on hig l-school graduates at grades 2 3 d While I don't have muchofa problem finding some underpaid linguists in grades 5-9 I personally am unawar of any who think they are underpaid abov those grades But here again we maype discussing two different ideas __ Iseems to be suggesting that as a matter of high agency policy linguists do not get promoted I would like to suggest that he look lower Promotion recommendations come from the divisions and offices and by and large they get their recommendations accepted One final remark the linguist who flees the field had better take with him some skills other than language I don't know many desk analysts TAs or CAs who are supergrades I f there is ever to be a group of supergrades in the language field they wi 11 come from the ranks of mul ti-lingulsts linguist analyst managers or other mul ti-skilled people And that's the truth Dan Buckley B Language Coordinator SOLUTION TO NSA-CROSTIC No 6 January-February 1977 I Puerto Rico in Anglicisms L - NSA Technical Journal Vol XVII No I Wi ter 1972 reprinted in NSATtichnical Journal Special Linguistics Issue III 19761978 i linguistic 'laissez- faire 1 has existed in Puerto Rico for a long time Although schools and newspapers have actively encouraged correct use of Spanish only the most undesirable borrowings from English have disappeared and new ones have appeared UNCLASSIFIED UNCLAS IFIED SPAHAKOPITA 10 sheets phyllo 12 x 15 inches 'I cup bUller melted 2 packages 10 ounces each fresh spinach 1 tablespoon salt 2 eggs 2 cups collage cheese small c'Jrd 1 Cup grated feta cheese 3 tablespoons parsley 2 green onions with tops minced 5 11 and pnppor f you're planning to enter this year's CMI or CLA Essay Contest or the newly established IAI Essay Contest you'd better get a move on The deadline for the CLA Essay is was 4 March 1977 and for the CMI and IAI Essay Contests is 25 March 1977 Complete information about all three contests can be found in the Fall 1976 issue of the NSA Technical Cut phyllo sheels into halves and place 10 pieces in a butlered pan 7 x 1 J inches Brush each sheet of phyllo with melted butler Wash spinach and remove stems Cut leaves into Y -inch lengths Sprinkle with salt and let stand for 15 minutes Beat eggs Add cheeses parsley and onion Squeeze liquid from spinach Fold spinach into egg mixture Season with salt and pepper to taste Spread mixture over phyllo sheets in pan and top with remaining 10 pieces of phyllo brushing each Journal sheet with remaining melted butler Bake in prehealed m c lemte oven lHO'F for 40 minutes Cut into squares and rv ho1 MaLes 6 servings I 'Greek spinach pie UNCLASSIFIED March 77 CRYPTOLOG Page 21 PI-Jan 77-532-25100 UNCLASSIFIED P L 86-36 00% imw 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