e 91 e1e Wal11Vt Dmall1 e oiwmV17 i D U umwv lilUlDWl ll l l Wl alIDB f waiw alWID l E 'il Jl1 J W'Ul OOWW iliJ PL 86-36 50 USC 3605 Non - Responsive _ lOP 181 T EO 3 3b 3 PL 86-36 50 USC 3605 Approved for Release by NS Z on 11-29 2021 MDR Case # 110903 6 I PL 86- 36 50 use 3605 iliSIU l'I' ANOT#IER LAST WORD ON I A T B F33 have been closely following the discussions on IATS and have frequently promised myself that given time I would make some comments of my own The recent article The Last Word on IATS CRYPTOLOG April 1977 provided the necessary impetus But first a few qualifiers and qualifications are in order My experience has a User Desi ner and Maintainer of --AG-22 IATS follow-on process ng an app 1cations systems The following coRPents emanate from an admittedly parochial background • but I feel they have a general relevance to •the IATS discussions • Most of the previous IATS discussions seem o be looking at the problem from the wrong end The concern of the writers seems to rest ith the obligations binding'tne intercept operators USSIO 101 the proliferation of untimely processing systems and the expanding volume of unwieldly data bases lacking analytic integrity I feel it 111ay prove worthwhile to approach the system from the other end -- to seek the opinion of the desk-analyst user who depends on the end product of the daily processing system to make his days productive The basic question to be answered is How does the analyst view the output of the system ond what does he do with it llo J Does the Anaiyst View the Output of the System • My observation is that even though a tremendous a ount of time and effort has been ap'plied to the definition of copying instructions September 77 CRYPTOLOG EO 3 3b 3 PL 86-3 6 50 USC 3605 euon li• Page 1 llttl I 5 Jflt SCllfPIF S lt271 IP lf111 PL 86-36 50 USC 3605 ·· · ' - i stallaticm of IATS equip¥O collllllUJli ati'o• EO 3 3b 3 PL 86-36 50 USC 3605 a sllfdlar ' but a print f complete int ercept links and cbmputer ha 3wat ••an4 the writi g of• Thus thi anal st is foraed to examine 1arge • mtd manually refine it by software t ie analyst vies • oggin selected data ih o some type o SOI • quick deliyry system • • It is his•o•p-1 11 1•0•1 • t-•at 'e-n•o•w-g·e·t_ s • tumznart•fol'lllat A rela d problem wit data fusy lly SPECOL directed-at the a daily pJ nk copy nonul1 1ti hin 24 hours £ • intercept -or sooner that ise4 be oourie store 6 - - --•-- -1s the t they usually lre poin b a o e or at be t two of the elements fol'Warded 3 weeks after in terc • Th f ri ce that would appear in art I Jlalyst summar The paid for his instant deJ iver t •sy 5tem _rs the extt'act can roduce a i · eliminatiQ of the analyst--ann°'ateci leetri• cally forwarded technical vehic S1iUM • • TECSUM EL AIR and the re ul taril • n ned d a ut 1 t e ana yst wer to ma e t e numerous base In a very real sens this iewpqint is seg'arete extract reque f'ts necessary to list all • correct in that we have elimina J k the field • tlte elements of a sUIIIIIUlry the resultant paper • site STR anal st and meved- th b rd n of • f ooa would quickly drive him back to the SxS exuining • seco -echelpn Gard What is needed i s a daily interactively • analysis an reporting to th• NSA desk ilJ dafed SOI Su1111Dary prc»gra111 that will •analyst Today's analys1t 1 P longelll' ap reciates • • the number of analyst• man -hours e mipate q by • • provide summarize target activity informa•our overworked11 intePcept operators and sor • • • tion at the analyst's discretion phisticated computer softldlre No • tter that • • be compiled on th basis of varying time ystem now ea1tly handlts amount ptervals daily weekly SOI period L Jnot even cootemplated 1 1nde monthly • tions does the dirty work iike • be idlmediately accessible preferably by • vouf and conven ently updated _____ an allows t e nalyst to nftlchiqe• • contain•an except'ion report section that s arc t e entire intercept record for ique • • highligllts unuspal occurrences for anaoit of chatter He sees o y th4t h lytic rKolutio p and begin at the beginning with be permanen Iy st ored as a refined de• • • and What Does He Do With It • • tailed anr lyst -verified data base Again from observations I would say tH t ere are four s ptions·that form t e basis the analyst examines the system output and then• of a uccessful 5 UIIIIII ¼' ' system They are manually logs pertinent data in prescribed for- • • 'Ebe summary will e primarily applicable mats The result of this logging process ha's •to stereotyped rasher than high-interest been variously described as the Sx8 cards in th r unusual target ' upper right-hand drawer drawing the circl • The effort requited o the analyst to make or historical notes but the generic tenn l • the system wor must be of a lesser delike best is SO Sul DDary The SOI Summar i S grh than the ffort t quired to keep an anal tically refined data base • • han logs I d base• r ct 0 o 0 t q as • - • •• • • •• Full text of the•target activity intercept must be readi 11 available io the analyst a re referably by VDU • will be co rected only as ar as ' lff ll'll is concerned preferably via P ing the anilystcorrected Sul llllary • The key to a s ccessful SOI Sunnary syt i b e its ability to learn Today's a Jwould be plaY d against analyst-supp 1e pa rameters historicJl target knowledge and AdMittedly this is not a comprehensive list mau-machine corrected current SOI data Thus but it does contain the fundamental data reas•the analyst's current knowledge of the target quired by an analyst to maintain continuity of gro s and he interacts in a feedback loop with an activity and to build a normal activity the s01 Summmary syste the system's knowledge data base from which inferences of abnormal oft e e d it i better able to reactivity can be drawn I believe that for the fine - 1'1 ' ' 1 11'1 -- - l ' l'I As a result the majority of target activity currently maintained longer per10 the reater the target knowledge the less ______ is outfor continuity purposes not intelligence purposes a computer-generated summary can be putted and the few decisions mus e made to produced from the lATS input that will allow refine the data by the analysts more freedom for the analyst to concentrate on One of the more significant aspects of a analytic rather than logging tasks well-designed SOI Sumary system would be its The basic problem with the current daily out- ability to highlight unusual activity in an exput of the follow-on processes is that it is not ception report There are two levels of analyst •• •• • • • is September 77 CRYPTOLOG Page 2 81881T JlitiRI 5 Z i Jliil I lit S C Si EO 3 3b 3 PL 86-36 50 use 3605 iHMJIHilT larget n • ----------------------1 knowledge concerning a availabl t• one time to the system j- current and historical _ The historical target k wledge resides in TEXTA Th t th st or in B Group to a gr i ter extent in BBAR us for at leaS e ereotyped activity the analyS t 's daily job would become one of reCurrent target knowledgi° would reside in the SOI b viewing SOI Summaries and exception reports and Summary that is being Jl'at daily by the making decisions designed to refined the SOI analyst-computer relat1o itsh1p Summary data base bY reso 1ving f The summary k dsyconfl 1c t s settem can draw on both l evtt s O iarget nowle ge ting parameters equating and identifying data ovide for the anal when_ To counter the feel · O f · f 1--- lexceeds normal pa meters For example• • • ing getting out O touch the anal 'c t has provide t• XTA BBAR with the • with a taritet it would probably be ad antag ous i l d h _ lo ' the analysts on a regular cyclic to s or1ca 110 e et retb to the process of reading all _ 1and keeping the hand logs his • • • w o u l als erve as a check and balance on the • • • • r • •• summary syS t m and the parameters and profiles • it relies on•lo-r exception reports • •• If the hypoth is of computer summarization NR0 nded above is •al plied to our current pro- • r gg cess1ng s rstem and dat4 base structure the 1stor1ca now e e on report another level exception reports would also indi- diagram below would be ll 'Rtneralized represen- • cate hilh-interest items 1____________ tation of the to d svstem • I • ' as1s l r• 0 1 IATS • ' '- PL 8 6-36 50 USC 3 605 1 rI _ CURRENT SOI StMtARIES • • I realize that the addition of summart ed data bases will add to Mr Phillips' coacern for the growing data storage problems•• I can only suggest that examinin the storage requirements for _________ in light of having a reduced re 1ne sW11111ar1zed data base may reveal a way to a net saving in on-line storage l would further suggest that the analyst will be using the summarized data bases for the majority of his SPECOL extracts with attendant saving in search time and output volumes The implementation of an analyst-VDU environment and distributed data bases Project RETINA in B Group will allow the analyst to be_come an integral and interactive part of the system and this should increase the integrity of the entire process If Mr Phillips' senior technical people believe that a sy tem • of this t pe is worth pursuing then any discussions to be held 1D1Jst start with the desk analyst and attempt to define his requirements From these requirements will come a design for the back end of the system and from there a determination of what foreseeable future hardware software techniques including interactive analyst involvement can produce Then and only then can a determination of the absolute minimum requ rements for formatting gisting or r flagging be established for the intercept operator In short I'm advocating less operator and analyst involvement at both ends of the system and greater use of the potentialities of the computer software coupled with the coming advances in access hardware September 77 CRYYTOLOG Page 3 ililfiRI• IIPiPP 7 IT OOJIIPF Q11917 p 7 Fill r EO 3 3b 3 PL 8 6- 36 5 0 USC 3605 --·----- ---- IJIOP 91J RIJT Wlllllt I PL 86-36 50 USC 3605 _ _ _ _ _ _ _ Pl6 OF B can handle 1t But if a machine-translation system can be devised so that nonlinguists can determine the intelligence value of the trans lated traffic for possible further evaluation by linguists the problem will at least be alleviated • is••-------I 1 The st line • • • I •9•1 1 the sect nd line • n e t n ird line'is a machine translation • wh1cn a P i ttedly is st ll a little rough but • conveys-the idea of HAPPY BRifflDAY I mean • BIRTifDA ' More on the translation later This j a tic le uses the word 111 arble to refer to errol'61 • which a killed linguist can easily correct We are n t referring td texts made totally unrecognizible as by sta ic - ---•1t This article describes the results to date late June 1977 Any machine-translation system requires a dictionary to give target-language the into language equivalents or translations of the source-laniruaae the fro111 lanouaoe entries _____ • The volume of traffic is expected to grow beyond the point where available linguists September 77 CRYPTOLOG Page 5 T9P 81JOREf t l'tlDRA EO 3 3b 3 PL 86-36 50 USC 3605 euonuw 1JJJIB111Q T8P - September 77 ----- ----- CRYPTOLOG Page 6 ' ' nne111 UIHBU -- - EO 3 3b 3 PL 86-36 50 USC 3605 •oP 81Hfllli• iHBRL • ________________________ _ And there is no reason why A Group or for that matter B Group and G Group need content •• 1 ···• 1 - _ _ automatic translation I I Both computation and l1ngu1st1cs have Ude great advances in the past two decades and machine translation is an idea not of the past but of the present ---------------------- land future September 77 CRYPTOLOG • Page •op 8118• 11• WAIBllii 7 PL 86-36 50 USC 36051 Iasks 81JfJ Ef JPOR'1 • s THERE A DOCTOR IN THE HOUSE ff different regions that he's a generalist not a specialist you can get some answers You can try to use the hand machine records which can be pretty cryptic if you have no standardi ation and don't read Sanskrit You can set up a task force to reanalyze what's already been done but not documented You can swallow if you're not too concerned about your problem of wanting to throw up the old well-worn phrases you hear from many management-level people including unfortunately even section chiefs such as I just didn't have the people You can For some ime now I have been concerned with check _through old technical messages You the lack of iechnical reporting and documentacan always go to the data base I didn't see tion and if•the examples I use appear to have any reason to waste time writing that up al•a slightly oriental flavorr it's because I wear though it was worth a product We can• t af •a red star oh my hat and carry a little red ford that luxury or It's too ambitious a program As those who know me can attest I book get highly emotional and extremely vocal when We have All heard ad nausewn that Product hear that kind of rationalization from people • is our bread' and butter and without it who should know better we'd be out f business Now I wholeheartedly • agree with 1lhis However which comes first Unfortunately you don't find analytic conclusions ideas and opinions in data bases In • the chicken -or the egg most cases you don't even find the correct In order to write the product someone has data People tend to forget that data bases are • to perform 10me analysis In many cases it's frequently uncorrected and are good only if the • easy for the problem-wise analysts to recoganalyst puts the information in as it happens • nize unusual activity and to issue appropriate not when lightning strikes in the form of an product bu the job shouldn't stop there irate supervisor there are a few good ones • although all too often that's just where it left • docs stop • If we actually insisted that analysts take • When it in c •nevitahlv back come the the time to write up an activity as soon as it's nuec • '- • I over we could save untold and I think very costly 111an-hours now spent in reinventing the wheel We don't act anymore But boy do we ever J eact Other answers include old cliches like How would I know -- I didn't work this region then or I think I remember we saw that in the year one but no one wrote it up or How do you expect me to find it -- we threw all those old records away ave you heard the latest rumor T re is a terrible miasma seeping through the corridors of NSA inf ltrating where it can all the elements of PROD creeping under doors insidibusly attackiOg all who cross its path Do you are You should because it's probably affect irtg you even a you road this Let me say I don't hink it's ju t a rumor It goes under the ianie TechniC l Documentation Syndrome and its constant companion is Apathy lhe answers to these questions should be found in the technical documentation of the problem but are they The answer to that question usually is that no report was ever written If you're lucky and if the analyst hasn't been transferred or has become so cross-trained on When the documentation is not available how much time and energy is wasted in trying to recover the information If it becomes necessary to recheck analysis for any reason where do you go to get the answer It would be logical to look in a technical report but remember the key word here is logical September 77 • CRYPTOLOG Page 9 iUtllli• OPOlili EO 3 3b 3 PL 86-36 50 USC 3605 •• r • • • • • • • • • • How swe t it is on ttilte rare occasions when you can g° to your favo rl CREF person and request a co y of a TSR orfWbrking Aid which is maintained in the libra or posterity and incidentally_ •• for usei yQb remember to include Tl213 on yuur di stributifn li J t that is To illunrate the po • ther take • • • • • • • • ••• • i· If and when 1110ment go right down the tube you take another look because it wasn't au ays documented it's time to find the ole wheel again I've hardly scratched the surface and there's much left unsaid -- a lot of which involves the section-level supervisors doing their primt uty job They should train their people on their problem explain the reasons for the necessity of logging certain information and what it means take the time to answer questions not foist them off with I don't have time now be enthusiastic themselves review to see that things are done correctly • and in a timely manner like right then not next year if ever and document the results I've known cases where something different was • not logged anywhere because a so-called analyst • ili io 111 _ _ _ - --iiiiiiiioiiii _______ thought Well my case never did that so it • lllUSt be a mistake or a garble I don 1 t think • I'll say anything or even write it down 1 ' Who 11 know or care a hundred years from now Yep It 1 s a mistake all right Boy is it a ' • • mistake and believe me friend I care ' ' Let me lii you over t head one- more time Do these same problems exist in other areas ' 'with the Inner and toutho lightly bn two other of the Agency I've heard of this technical things all at- the same tS While puch of syndrome elsewhere Are they being swept what has been said so fat has been a under the desk or the black cloth Or are ' ' rimaril tow• rd the tr c anal st they patronizingly shrugged off with Well you just don't see the big picture • • To me the big picture involves taking that ---- ---- -----------rnitty-gritty technical stuff and documenting it is contihually concerned with so I can provide the customer ultimately with Right If 1 said wrong good sound product reporting put in perspeci r 1r w ililil•liilllii•oliiiilt _1_•_ _ _t_h_r_oiiiliil• h - e•t - il tive based on the subject for the customer before he asks for it not after I want to instill in t he analysts the reasons why technical logging and reporting are not just busy work truetofoundation of our existence -•• If these havetf't been put to a permanent ata but beingtheable say I have the technical facts ••base that can e manipulet or technically to back up my product words -- where I can find documented in _ ome fashi1 n _1 11111ro 1a111 11 _ 1111111 _ __'III 1 em when I need 'em ••sa in in a roduct re o -t We I have spent too many man- woman- •· - a-v 'e r 0 n e s m' 'lli a ' r l ' n ' g s ' y o ul'l 'll' 'n'll ow - - 'lm lu'lll lf'to ' our what ever-hours writing wrap-up seven-year • chagrin -- but ifon•t aslt me to prove it because I etc pl'oduct reports based on redoing or • can't find the documentdtion Think of the flap trying to redo the technical analysis to sup timewecouldhavesavedf prarainydayifwe port what's being said When someone says could have looked up these kinds of messages term study to me I want to go into a corner • in some techni al vehicJie We did staTt this and suck my thumb on but who an find it now What is the answer What can we do about it i- _____ ___ ___________ iwe can insist on solid technical documentation We can and should establish guidelines for technical reporting the same as we have for product - - you wi 11 report when If you've read this far and think all of This t e o i 1w·u 1 1 1 101 1r -a t 1111 i2p_ _ 1111 11g-w j t b---- this sounds lilte a cry in the wilderness it ia other good1es _ ___ HELP I'm tired of being the only nag l -- -' '- - ----- - When we•re P S Before I get off my soapbox I would flush with people you lllllY -- Just may - like to pose one final question lsn 1 t it interfind these data in technical and product esting that part of the professionalization test reports but when the crunch comes and people for Traffic Analysis is writing a TSR are needed elsewhere these two problems unless extremely high-interest areas of the ' 1----- -i-----i-l'---- ------ I ------•---- ' ' I '- ' ' 0 September 77 • CRYPTOLOG • Page 10 I PL 86-36 50 USC 3605 WHITHER THE R A - The follOIJing article represents the personal and professional opinions of the author and should not be intel'pFeted as an officiaZ statement on the paz t of the Panel J T fl his is a short paper ambitiously aimed at resolving a controversi l question To what end does the Special Research Analyst SRA serve in the U S SIGINT system There are more SRA people in the U S SIGINT organization than in any other single SIGINT career field and the question of what constitutes the substance of their work is surrounded by a sometimes vigorous debate An official definition exists but does not seem to have taken hold It somehow appears easier to define what a linguist does or should be expected to do -- or to describe the duties of a traffic analyst cryptanalyst telecolDlm lnicator engineer etc -- and get a general agreement Still whither the SRA Being the military member of the SR and IS lnfomation Science Panel and working at an SR assignment in A7 Office of Operational and Strategic Studies has motivated me to try to come to grips with what it is that SRAs are bent towards The idea is toge at the function by trying to get at the results what is the SRA supposed to produce Regrettably convictions of those who are certain they know what an SRA is differ from others who are like-minded This is distressing in its administrative ramifications What follows is a contribution rather than a fixed proposal -- one which could further confuse the issue In part i t is aimed at the person who aspires to be an SRA or who believes he or she already is one The SRA produces intelligence by anaiysie and by one or more forms of reporti ng of which the end product is but one As in the case of 1 1 September 77 • CRYPTOLOG Page 19 '10tiPIRIHPl'I t L -------------- - -- --- r nrr Ill C SSl IEZS CHbi I - QOtifilRRfl•1 •1 a person suspected of a c-rime the SRA must be shown to have motive means and opportunity Analysis and reporting is the means and the opportunity Reports can take several forms They can be oral written graphic pictorial or combinations of those media The actual form is not as essential as the substance of the report Importantly the SRA produces signals intelligence Roughly speaking about 85 percent of intelligence resources are expended in collection and processing prior to the final production process For the most part the active SRA works within that 15 percent area where final production is performed should not be structured or undertaken to di vide thllt which the target is trying to put together It is not reconstructive if it does that The target is always trying to bring something about and the SRA's product should try to recornJtruct whatever his target is constructing Because the SRA is rebuilding a system it is vital that the SRA have a good working grasp of other aspects of total target reconstruction TA CA SA language and so forth Of no less importance is the SRA's fundamental grasp of collection and information processing The SRA 1 s efforts to re construct target construction will be limited in the extreme unless these other reconstructive aspects are grasped All intelligence is of two broad kinds reconstruction and estimation The former feeds the latter often directly All analyst TA CA SA etc reporters are in the business of reconstruction In connection with the above rather uncomplimentary analogy this is the analyst's motive the analyst is impelled to reconstruct something It is the specific shape of that something which differentiates among the various kinds of analysts This occurs simply because that shape and dimension is translated into different results according to the means and opportunities -- the discipline at hand whether TA CA SA language or SRA What the TA is actually about is the reconstTUction of a t rget communications entity the CA is supposed to reconstruct a crypt system the SA wants to re-create a signal They are taking what they can discover about various aspects of a target and reconstructing it according to its actual construction Explicitly and implicitly each kind of analyst is motivated to reconstruct one or more aspects of a target reality In a sense this is modelbuilding There are few if any cases where reconstruction results in total fidelity The objective is to come as close as possible To do so ultimately requires a broader knowledge than one based on actual SIGINT sources themselves To get close to the most unachievable truth requires viewing the target -- and the analytic effort -- in perspective and in context So-called collateral'' is useful in this respect Many vantage points are useful but whatever these methods it is necessary to penetrate the target to see the world as the target sees it Accordingly what is the SRA motivated to reconstruct What aspect of the total target does the SRA endeavor to model or rebuild My idea from studying my own work and that of my colleagues is that the SRA is about the reconstruction of the target's system By that 1 mean that the SRA is trying to put it together' in something closely akin to the way the target puts it together Man is a constructive animal even in his destruction He is forever making wholes of parts or trying to Analysis Military target reconstruction at the SRA level is a facile example There are numerous other good examples but I am more familiar with this one The SRA working a military target is trying to reconstruct in various product forms a target military event or operation in itself a kind of system having input throughput and output The event is something the target is -constructing and the SRA should be reconstructing A 1110re specific example extant in target reality would be a military exercise The SRA working that kind of problem wants to reconstruct the exercise This cannot be done without prior reconstruction of target communica• tions in their various traffic signal cryptographic language and other aspects to the degree necessary to permit final production by the SRA In thi progression the SRA enjoys no vaunted rol only the unceasing responsibility to reconstruct a live and hopefully lifelike evont Finally I think that in a real sense the SRA is wont to reconstruct target decision-making In the example of military exercise reconstruction the SRA should aim to put into his producta feeling for a target general officer's decision and directions to his forces Remember I said a feeling I do not imply that the SRA is necessarily an artist in the conventional sense What I have in mind is that the SRA is trying to get at target decision-making and the event results of that decision-making By reconstructing events the SRA is trying to cast some light on the nature of decisions conditioning the event Beyond this point analytical intelligence work crosses over into the business of estimates which is something in addition to and beyond reconstruction I hope that these brief reflections are helpful This is my own view which does not necessarily represent the position of any authority in the Agency and I would not want it construed as such I am merely trying to get a handle on the nature of the SRA's work by linking SRA motives with SRA means opportunities and products reconstructions September 77 • CRYPTOLOG Page 20 80ffPIDllff fl 15 - PL 86- 36 50 USC 3605 UNCLASSIFIED THE THAI SEMINAR JIOG r 1 692 new tra1n1ng program in Thai language for preprofessional and postprofessional NSA linguists may provide a useful model for similar training in other languages Although it has been in existence only since September 1976 three courses have already been conducted and planning is underway to set up an Indonesian program along the same lines In addition one Thai seminar participant whose current assignment involves French and who is under pressure to attain certification in French declared This ill exactly the t ype of course we need i n French So there may be a French version soon as well The purpose of the Thai program is to provide opportunities for preprofessional and postprofessional Thai linguists to participate once each year in a continuing education program to grow in the language to increase their knowledge of Thai cultural background to discuss language problems which may be bothering them and to receive periodic professional stimulation through immersion in the Thai language The program is designed to offer intermediate and advance training as a follow-up to the Thai Basic Course Currently eight seminar-type courses -- two intermediate and six advanced -are available However one advanced seminar on current affairs may be taken repeatedly since the study material is always different Seminar sessions are held once a week for 12 weeks at a remote facility with a native instructor Each session lasts 4 hours during which all discussion is in Thai language Required reading assignments on Thai cultural subjects sometimes supplemented by additional reading materials are given one week in advance Each student is assigned responsibility for making an oral report to seminar participants on a portion of the required reading Everyone may discuss the oral reports describe language problems they have encountered and ask questions or contribute experiences related to the subject Instructors are native Thai speakers Their job is to • offer criticism or correction when students misspeak • answer student questions regarding either subject matter or language • moderate the seminar and • stimulate discussion if conversation lags Because the amount of discussion generated by different topics varies considerably other teac hing techniques are used to supplement discussion such as having students read aloud or transcribe from dictation Besides the obvious advantages of this kind of program for maintaining and improving language skills two points seem to me to deserve emphasis First this program significantly improves prospects for a professional career in Thai language For the past 20 years neither advanced training nor periodic professional stimulation has been available to the Agency's Thai linguists With the addition of this program and the equally new Thai segment of the Foreign Language Cassette Series Fl ACS in which cassette recordings of Thai radio and television programs and verbatim transcripts are provided to the linguist for self-study purposes the junior Thai linguists of today can look forward to considerable help toward the professional growth of their language skills Second since for a variety of reasons highly skilled linguists in Third World languages often change jobs to become managers reporters programmers linguists in other languages etc it behooves the Agency to provide the incentive and means for these people to maintain their skills during periods when their primary focus lies elsewhere Again the Thai seminar program and FI ACS are steps forward to correct the previous lack of means I hope that some day the Agency wi 1 l do more than talk about effective incentives for maintaining language skills not currently used on the job Perhaps satisfactory completion of an annual seminar course of the type described above might serve as the basis for an annual monetary award or a QSI What is needed at the very minimum is a policy statement directing supervisors to encourage linguists especially professionalized ' linguists to attend such seminars In our Thai program we have been successful in attracting three professional linguists currently working in jobs unrelated to Thai language and one whose job is only slightly related However some of these people feel their participation has been unenthusiastically or perhaps even grudging Ir accepted by their supervisors This kind of attitude must change There is wide recognition these days that improvements in the language field are very important to the Agency Even though there are numerous facets of the language problem that deserve attention so that what we touch on here is a small part of the whole of one thing we can be confident The Thai seminar program is a step in the right direction ••••••••••••••••••••• September 77 CRYPTOLOG Page 24 10fOLJt81Ufl IB