OCR of the Document

View the Document >>

National Security Agency, Cryptolog Vol. 11, No. 11-12, November 1984

mGBPiJuurnmGBPiJl1 5lBl J EUU GBPiJ $Wffi U0 i V lB $0 i lB $ lB 00 $ D D $ f 00 D i 11 DW D lBrn UI Jl11 J1 NOY-DEC 1984 EO 1 4 c P L 86-36 I TRENDS IN HF COMMUNICATIONS U oooooooooooooooo SWITCHBOARD PAST PRESENT AND FUTURE U 1 I oooooo oo 5 OUT OF MY DEPTH U ooooooooooooo oooooooooooooooooooooooooooooooo oo 'oooooooo 7 I Looooooooooooooooooooo ooo oooooooooo 8 C EXPLORING A DOS DISKETTE U ' 12 HUMAN FACTORS U oooooooooooooooooooooooooooooo _ 22 CAN YOU TOP THIS U E Le 19h Sawye r ooooooo 24 PERSONAL COMPUTING IN A GROUP U oooooooooooooo 25 FACTION LINE U oooooooooooooooooooooooooooooooo Cal Q Lator ooooooooooooooo 35 NSA-CROSTIC NO 59 U oooooooooooooooooooooooooo D H W oooooooooooooooooooooo 36 I I 'FIllS BOCl 1MENT JONTAINS JOQI'1WOHB MATI'1HIA-b Ghi'tSSIFIEB BY flSA eSSIt IH 2 BEGJ l1 tSS IFY eN er il inft iut AgCiICY'S Bc ezmiiiatiori Req iled Declassified and Approved for Release by NSA on 10-16-2012 pursuant to E o 13526 MDR Case # 54 net aCID 4009933 EO 1 4 c F L 86-36 Published by PI Techniques and Standards VOL XI No 11-12 NOVEMBER-DECEMBER 1984 Editorial PUBLISHER BOARD OF EDITORS r - Editor Product ion o 1 963-3045S 963-3369s l Collection o o 11 963-3961s Computer Security 1 859-6 044 crYPtolingui i 963-1103s Data Systems o o 1 963-4953s Information Science o 1 1 963-5711s Puzzles David H williamsH 963 1103s Special Research Vera R Filby 96 -7ll9s Traffic Analysis Robert J HanYOk 968-3888S For subscriptions send name and organizat on to 1 Ip14j P L 86-36 To submit articles or letters by mail to PI Cryptolog via PLATFORM mail send to cryptolg at barlc 5 bar-one-c-zero-five note no '0' in 'log' Contents of Cryptolog should not be reproduced or further disseminated outside the National Security Agency without the permission of the Pub lisher Inquiries regard ing reproduct ion and disseminat ion should be directed to the Editor FeR eFPle fAn eSE fJlfnY aCID 4009933 SseRl3'i' speRS TRENDS IN HF COMMUNICATIONS U P L by Nov 84 CRYPTOLOG BE8MJ'i' Bf6til 86-36 1'-----_ _ Page 1 EO 1 4 c P L 86-36 aCID EO 1 4 c P L 86-36 BEiBRB'f BFSlfiE 4009933 Nov 84 CRYPTOLOG BEieR ' BPe Page 2 aCID 4009933 EO 1 4 c F L 86-36 SE8RGBPl'f' SpeIHl Nov 84 CRYPTOLOG Page SESRFiT SP8IlE 3 aCID 4009933 L SBSRB'f SP61ffi 4 c P 86-36 EO 1 4 c P L 86-36 Nov 84 CRYPTOLOG ISI SRI3'F Sf6f Page 4 aCID 4009933 P L Nov 84 P L 86-36 EO 1 4 c CRYPTOLOG S gSRE'l' 86-36 Page 5 112'dfB 8 rIA SSIlHf'f SIWlIfgObS QN bJPY seSRet 4009933 to lot R tPT'DL Of We lorgo t have su sc rif lo ' tnove4 Wi I n tell 86-36 86-36 1 4 c Nov 84 CRYPTOLOG Page 6 IIltIlBl f l YIlt S8 IHI'F eUMRlBt 5 SIlt' 4009933 a 1 4 c P 86-36 Q o OUT OF MY DEPTH U Nov 84 CRYPTOLOG 88IlFIBIHI'f'IAt Page 7 o ero 4009933 EO 1 4 c P L 86-36 '-- 1 'P L 86 36 byl'----_ _IS74 Nov 84 CRYPTOLOG Page 8 EO 1 4 c P L 86-36 E0 CID 4009933 Nov 84 Page 9 go 1 4 OCID 4009933 93 34 93L 86 36 Nov 84 Page 10 Eo - OCID 4009933 86_36 Nov 84 Page 11 4009933 EXPLORING bylL --_lp13 INTRODUCTION LOOKING AT THE DISKETTE Many data systems professionals will go through an entire career using a personal computer and handling small diskettes without really having an idea of how data is organized on them There is nothing magic about it and in fact it isn I t really diff icul t to understand where and how information is being stored and what it looks like on the disk This paper will attempt to explain how a 5 1 4 inch floppy disk is organized under PCDOS 1 1 written by Microsoft for the IBM-PC We will be using an IBM-PC and two doublesided double-density disk drives Most information was obtained from the Peter Norton book Inside the IBM-PC one of the most useful books on il7 omputers I have found and the IBM DOS 1 1 manual--very useful both as to informat ion and as a cure for insomnia The diskette data displays were obtained through the use of the Peter Norton util ity DISKLOOK Microsoft went to great lengths to have the software overhead take up as little room on the diskette as possible This obviously makes good sense A small computer isn't much good to you if the operating system takes up the entire disk As a consequence the directory and allocation table look rather cryptic but once explained they are easily understood Nov 84 Formatting When a user removes a blank diskette from the box for the first time he has two choices He can use it as a frizbee or he can turn it into a reliable storage medium for use in his personal computer We will address the second possibility with the full knowledge that we do not exclude the first possibility at any step of the process When you run the FORMAT program as you already know the diskette under DOS 1 1 is divided into 40 tracks per side 8 sectors per track with a boot record directory file allocation table and optionally an entire operating system Ok let's format it Exploration We'll now take our formatted diskette and place it in drive B --the one on the right In the left drive we'll place the Norton Utilities and type in DL for DISKLOOK This assembly language program will allow the user to look at the floppy diskette in the right drive sector by sector After DL is typed a self explanatory list of options comes up on the screen We choose one that allows us to select a particular sector side and track We'll start by looking at the 'Boot Record' which is on the first sector of the disk CRYPTOLOG Page 12 FeR eFFISIAb ese euJ iY P L 86-3 aCID 4009933 FeR eFFI6IkD eSE 8Htl Side O Track 0 Sector 1 1at ector of Boot record E8279003 01140000 00000000 00000000 00000000 00000000 00000000 00000000 OOOOOOCO 19FA8CC8 BED2BCOO 7CF88860 008ED88E C033D28B C2CD1372 72002E83 3E037C08 74062EC6 06647002 8B00002E 5180092A C184008B F05633D2 33C08AC5 2EF63664 F48BC684 02COl372 205E592E 2936057C 741F8BC6 700308FE C5BI0151 8GBP08002GBP 3B36057C 7C052E88 COEAOOOO 60008E67 70GBP80200 GBPBFE32FF 2EAC247F OEBB0700 COI05EEB EFC3E933 FFBBOOOO B9040088 1GBP72338C C88E08BF 0000B908 00268000 20268040 F4BFOOOO BE8B7DB9 OBOOFCF3 A6750FBF 20008E97 F3A67502 IFC3BEIB 7DE8A2FF B400CD16 IFF9C30D 2D537973 74656020 64697368 206F7220 64697368 6F72000A 5265706C 61636520 616E6420 73747269 6E792068 65792077 68656GBP20 72656164 790DOAOO OA446973 6B20426F 6F742066 61696C75 72650DOA 726F736F 66742C49 6E632069 62606269 6F202063 62GD646F 732020 3 6FG03005 CGOG772F FF837EFC 7EF73B75 05C60676 2FFF89EC SDCA0400 00000000 00000000 00000000 00000000 00000000 00000000 0000000000000000000000000000000000000000 00000000 00000000 00000000 ' ooooooooooooooooooooooo 8E083302 ooooooooooooooooooooooo 3 oo 69E88500 ooooo I oo o ooooo 3 ooooo r1 oooo 8BOE037C r ooo I t oooo dl oooooooo I 708AE88A Q oooooooo V3 3 ooooo 6dl oooo 2EF72665 oooooooo r- Y 6 ltooooo e 36051CE8 1 oooooo Qoooo 6 II ooo 6 1 o 740B56B4 ooooo o oo gl ooooo 2ooo S t V oo 0102C013 ooooooo A o o o o 3 ooooooooooooo 202047E2 oo r3 oooooooooo o M Goo 7DB90BOO ooooooo l oooooo uooooo l oooo OA4E6F 6E ooo u ooooo l oooooooooooo Non 20657272 -Syate diak or disk err 6B652061 or o Replece end atrike a 01000200 ny key'when ready oooooooo 00406963 oo Oisk Boot fei1ure ooo Kic 6FG03069 rosoft Inc 1b b10 coaOi 00750B80 bados co O ooo w oo oo u ooo 00000000 u ooo v ooo l oooooooo o o 00000000 oooooooooooooooooooooooooo 00000060 oooooooooooooooooooooooooo F1gure 1 - The Boot Record This record serves one purpose It instructs the operating system in read only memory ROM as it is known to its friends to go to the location on the diskette containing the rest of the operating system and read it into memory If the system is there it is read in if not control returns to the boot record and an error message is printed out So as the prompt comes up we type in 0 for the side 0 for the track and 1 for the sectpr Zero for the side Real people start counting at 1 computer people start at zero OK why didn't we start at track zero There is no track zero Tracks on the PC were numbered by a computer person who skipped class the day they learned how to count Tracks start at 1 everything else at O many ASCII characters and in fact we don't Look down now at the 13th row and below There we do see a translation and we probably recognize it We are of course looking at the error messages that we get when we have put a non-system diskette in drive A and attempted to boot up To keep the boot record as small as possible Microsoft wrote the program in machine language but the messages to the user had to be in ASCII characters The Directory The next two sectors contain what is known as the File Allocation Table FAT We will skip the FAT for a moment and proceed to sector 4 still on side O As soon as this information is entered the screen display comes up with 6 columns of hexadecimal numbers on the left and a few readable words and other symbols on the right This is the sec tor we requested and is a display of the Boot Record Sectors 4 through 11 have the directory This area contains all the information necessary for the operating system to retrieve files requested by the user to sort them by time and date if required and to perform other operations as needed Looking at the boot record Figure 1 you can see rows of hexadecimal numbers broken into six columns' with an ASCII translat ion next to each row Since the boot record is in machine language 'we would not expect to see Looking at the directory Figure 2 we do see some sort of information in the first couple of lines and we'll get to that shortly but for the most part we see rows of F6's and every so often an E5 When we formatted the Nov 84 CRYPTOLOG Page 13 FSR SPFI6IkD ElSE StitH --------------------------------- - aCID 4009933 FgR 8FF1Slfm SB Lr disk the program took all 40 tracks 0 - 39 and put a pattern of H110110 F6 in hex on them It then ent to the directory area and each 32 bytes put an E5 The E5 signifies the beginning of a file record Now a sector will hold 512 bytes of data and there are 8 sectors allotted for the directory Eight sectors will have space available for 4096 bytes 8 x 512 Since there are 32 bytes per record a DOS 1 1 diskette can hold 128 files 4096 divided by 32 With this information we'll try to understand the directory Starting at the first line of Figure 2 we can see some readable characters The first ten bytes contain the hex equivalent of the ASCII characters for PROJECT BAK We can see this translation on the right So bytes 0 through 10 show that the first entry in our directory is PROJECT BAK The first few lines of the directory do appear to be telling us something Microsoft organized the entries in the following manner Byte 11 has a 00 as it is supposed to because it is not a hidden file or a system file The user has the opportunity to make a BYTE 0-7 File name an E5 at offset 0 indicates that the the file has been erased 8-10 Filename extension 11 File attribute 02 hidden file 04 file 00 for all other files 12-21 Reserved this implies that in a future version of this operating system some or all of these bytes will be used to further describe the file 22-23 Time the file was created or last updated Time is mapped according to the fo 110wing rules I I 23 I 22 hour I minutes I seconds I 15 14 13 12 11 10 09 08 07 06 05 04 03 02 01 00 hours minutes seconds 24-25 system binary number of hours 0-24 binary number of minutes 0-59 binary number of two-second increments Date the file was created or last updated Mapping is done as follows 25 I 24 year I month I day 15 14 13 12 11 10 09 08 07 06 05 04 03 02 01 00 I year 0 - 119 1980 - 2099 month 1 - 12 day 1 - 31 26-27 Cluster number of the first cluster in the file 28-31 FiLe size in bytes The first word contains the low order part of the size Both words are stored with the least significant byte first Nov 84 CRYPTOLOG Page 14 F8ll 8FFl'el'Ah aSE 8 fhY I - 4009933 FeR Side O Track O Sector 4 1st sector of Directory 50524 4A 45435420 42414BOO GBP4080200 00360000 324A554C 00000000 0000ABA4 E3081000 F6F6 6F6 F6 6 6F6 6F6F6F6 ESF6F6F6 F6 6 6F6 F6r6r6 6 F6F6 6 6 F6 6F6 6 E5 6F6F6 F6f6 6 6 f6r6 6 6 6 6F6F6 F6 6 6F6 6F6F6F6 F6 6 6 6 ESf6F6 6 F6F6F6F6 F6 6F6F6 6F6F6F6 6F6F6F6 GBP5FGF6F6 6 6 6 6 6F6 6F6 6 6 6F6 F6F6 6 6 6 6 6F6 F6F6r6 6 ESF6F6 6 6 6 6F6 6 6F6 6 6F6 6 6 F6 6F6 6 E5 6F6F6 6F6 6F6 6 6 6F6 6 6 6 6 F6F6F6 6 F6F6 6F6 F6F6F6F6 GBP5F6F6F6 F6F6 6F6 F6 6F6F6 F6 6 6 6 F6 6 6F6 E5 6 6 6 F6F6 6F6 F6F6F6F6 F6F6F6F6 6F6F6 6 F6F6 6 6 F6F6F6F6 GBP5F6F6F6 F6F6F6F6 6 6F6F6 F6F6F6F6 F6F6F6F6 eFFISlh 00000000 59202020 80100000 F6F6F6 6 6 6F6 6 F6 6 6F6 F6F6 6F6 6F6r6 6 F6r6r6 6 F6F6F6F6 6F6 6 6 6F6F6 6 6F6 6F6 F6FGFGFG 6 6 6 6 F6 6F6 6 F6F6 6F6 6 6F6 6 6F6F6 6 6F6 6F6 F6F6F6F6 WS6 eN Y 00000000 0000914D PROJECT BAK oooooooooooo K 42414BOO 00000000 oooooo Goo 2JULY BAK oooooo GBPSFGF6F6 FGF6F6FG 6F6 6 6 6 6r6F6 6F6F6 6 F6 6 6F6 F6F6 6F6 F6 6F6F6 GBP5F6F6 6 6 6F6 6 F6 6r6 6 6F6 6 6 F6r6 6F6r6 6 6r6 F6F6F6F6 F6 6 6F6 oo oo oo o GBP5F6F6F6 F6 6F6 6 F6F6F6F6 F6F6F6 6 F6F6F6 6 6F6 6 6 F6FGFGF6 F6FGF6F6 GBP5F6 6 6 F6F6F6 6 F6 6F6 6 F6 6F6F6 F6F6 6 6 F6F6F6F6 F6 6 6 6 6 6F6F6 ESF6F6 6 6 6F6 6 6F6 6F6 F6F6 6F6 F6 6F6F6 F6F6F6F6 Figure 2 - The Directory fLle hidden so that in a directory search the part icular file in quest ion will not appear Bytes 12 through 21 have OD's as they are supposed to Bytes 22 and 23 have the time as you can plainly see You say you don't plainly see that It really is a little confusing so lets take a look space needed and operating system which equals 9 AM if it were 9 PM we would get 21 2100 hours The next 6 bits represent the minutes Using the same logic as before we get 44 The last bits give the number of 2 second increments 10001 would give us 17 indicating 34 seconds Well that's how the time is represented-somewhat confusing but it does minimize the Nov 84 is important to an Bytes 24 and 25 which contain an E408 represent the date and again we'll examine this closel y We first switch them 08E4 For the file PROJECT the time is written as 9l4D This can be represented as 1001 0001 0100 1101 Referring to our file description on page 14 the first thing we do to translate this is to reverse bytes 22 and 23 So we would get 0100 1101 1001 0001 Now the leftmost five bits represent the time This means we have that 0000 1000 1110 0100 Now the first seven bytes represent years past 1980 In our case 0000100 represents 4 ind icat ing that the file was made in 1984 The next four digits 0111 represent the month 7 July The last digits 00100 represent the day 4 in our case So now we know that PROJECT BAK was placed on the disk on July 4 1980 at 9 44 in the morning Bytes 26 and 27 give the starting cluster We will look at that further when we examine the FAT but suffice it to say right now that this file starts at cluster 2 Bytes 28 - 31 contain the file size with the first word containing the lower order part of the size and both words stored with the least significant byte first This is CRYPTOLOG Page 15 FeR eFFI8IAfJ HBFJ aUK F6R 8FFI8IIdJ H6E1 QUbY OCID 4009933 Side O Track O Sector 2 1st sectorol file Allocation Table rrrrFF03 11200113 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00 oo00 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 F6F6F6F6 40000560 4001FFOF 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 F6F6F6F6 00078000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 F6F6F6F6 09A00008 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 F F6F6F6 COOOODEO OOOFFOFF o 1 o 00000000 00000000 1 o o 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 0000000000000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 F F F6f6 F6F6F6F6 F6F6F6F6 ' F6F6F6F6 F6F6F6F6 Figure 3 - The File Allocation Table basically the same idea that Microsoft used when storing the tilDe and date so we won' t go through it again but the file size is 13 024 There without much stress we have been able to understand how file information is stored in the directory by DOS 1 1 The File Allocation Table In addition to sides tracks and sectors the diskette is also divided into clusters with two sectors per cluster Since a sector can hold 512 bytes of data a cluster can hold 1024 The function of the File Allocation Table is to keep track of which clusters are holding which files' data If data is to be put onto a diskette it is allotted room in 1024 byte chunks at a time If a file only consists of 1 byte it is still allotted 1024 This may seem wasteful but every cluster r - quires a corresponding entry in the FAT If we decrease the size of the clusters we increase the size of the FAT and therefore increase the amount of time that must be spent looking in the FAT for the next cluster The designers of this operating system evidently felt that 2 sectors per cluster was just about right Nov 84 Since there are a total of 640 sectors per disket te there are 320 clusters How many bytes do we need to represent a cluster Clearly 1 is not enough it could only represent numbers from 0 to 255 Two would be fine up to 64k could be represented but this takes up too much room Microsoft settled on 1 5 - 12 bits So when we talk about offsets in the file allocation table we are talking about groups of three hex characters a byte and a half Since a nibble is half a byte how about a 'chomp' for a byte and a half Let's first look at offsets 0 and 1 in the FAT in Figure 3 Remember when we talk about an offset in the FAT we are talking about I chomp--a byte and a half We see FFF for offset 0 and then another FFF for offset 1 The first FFF tells the operating system that we have a double sided diskette If it were single sided there would be an FFE The FFF at offset I is a filler reserved for future use by IBM or Microsoft As ment ioned above the firs t cluster of our first file is #2 which corresponds to track 0 sectors 3 and 4 of side 1 These relative correspondences can be found in the IBM DOS 1 1 manual Appendix C After the operating system has read that cluster where does it go next Does our file cont inue at CRYPTOLOG FQ QFFI I6Z H8E Page 16 8 Y DOCID 4009933 Side 0 Track 0 Sector 4 lat aector of Directory E5524F4A 45435420 42414BOO 00000000 00000000 0000914D oo ROJECT BAK oooooooooooo K 40aO OO OO OOOO 24AS54C 00000000 F6F6F6F6 E5F6FGFG F6FGFGF6 F6F6F6F6 FGF6FGFG E5FGF6FG F6F6F6F6 F6F6F6F6 0000ABA4 F6F6F6F6 FGFGFGFG F6FGFGF6 F6F6F6F6 F6F6FGFG FGFGF6FG F6F6F6F6 FGF6F6F6 GBP3081000 F6F6F6F6 FGF6FGFG ESFGFGFG F6F6F6F6 F6F6FGF6 F6FGF6F6 GBPSF6F6F6 F6F6F6FG 59202020 42414800 00000000 oooooo6oo2JULY 80100000 F6F6F6F6 FGFGFGFG FGF6FGFG F6F6F6F6 FGFGFGF6 F6F6F6FG F6F6F6F6 F6F6FGF6 E5F6F6F6 F6F6F6F6 FGF6FGF6 FGFGFGFG ESF6F6F6 FGFGFGFG F6F6F6F6 F6F6F6F6 ESF6F6F6 F6F6F6F6 F6F6F6F6 F6FGF6FG FGFGFGF6 F6F6F6F6 FGFGFGF6 F6F6F6F6 F6F6F6F6 F6F6F6FG BAKoooo oooooooooooooooooooooooooo oooooooooooooooooooooooooo oooooooooooooooooooooooooo oooooooooooooooooooooooooo ooo o o o o o o o o o o o o o o o o o o o o o o oooooooooooooooooooooooooo oooooooooooooooooooooooooo oooooooooooooooooooooooooo oooooooooooooooooooooooooo Figure 4 - Directory With Firat File Erased the next cluster or is it located at some entirely different location on the disk Or for that matt er is there a next cluster This is where the FAT comes in We read the first cluster which happened to be cluster #2 We are now interested in the next cluster of our data You take the cluster that you have just read and if it is an odd number you subtract 1 in this case you don't do anything because 2 is not odd You then go into the FAT 2 positions and write down the next two chomps that you find You should be writing down 034 000 last c luster of our file cluster As you can see the FAT is quite important It is so important that there are two copies of it on the diskette one in sector 3 and one in 4 If one cannot be read for whatever reason hopefully the other one can be Both the FAT and the directory are read into memory when the system is brought up and are kept there In general if we have 6 hex digits hI h2 h3 h4 h5 h6 we unscramble them to form two words o h4 hI h2 0 h5 h6 h3 OK let I s try that with the 034 000 In this case the hI 0 h2 3 h3 4 and h4 h5 and h6 equal D's Rearranging our numbers we get 0003 0004 Now we were interested in cluster 1t2 Since it is even we use the first byte 0003 if it were odd we would have used the second byte 0004 Looking at the FAT again it does appear that the clusters were assigned in order Suppose we were interested in the 15th cluster We count in 14 remember with an odd number--15--we subtract 1 groups of 3 hex digits at a time We come to OFFOFF The digits are unscrambled according to the above formula and we come up with OOOF OFFF Since 15 is odd we are interested in the second of the two words OFFF is clearly too large a number to represent a cluster and yes as you have probably deduced it indicates an end of file The 15th cluster of PROJECT BAK is the Nov 84 There is no 16th CRYPTOLOG Page 17 FeR eFF181 c eBB 81lhY aCID 4009933 Side 0 Track 0 Sector 2 1st sector of File Allocation Table FFFFFFoo 00000000 00000000 00000000 00000000 00000000 11200113 4001FFOF 00000000 00000000 00000000 00000000 1 00000000 0 oo000 00000000 00000000 0000000000000000 00000000 00000000 00000000 0000000000000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 Figure 5 - FAT With Firat File Ereaed file could be found There are in fact programs that could make very educated guesses as to what the FAT probably looked like and recreate it for you Removing Files Suppose that for reasons of privacy we decide that we don't want anyone to see the file PROJECT BAK so we erase it Looking at the revised directory Figure 4 there is an E5 in place of the first character of the file name but it doesn't appear that anything else has changed If we look at the FAT Figure 5 it appears that the sectors that were being used for PROJECT BAK are now available They of course are the ones starting at offset 2 chomps and continuing for the entire line Looking through the diskette which we will do shortly you would find that PROJECT BAK is still there and in fact has not been touched So as you can seen when a file is erased the FAT is changed to indicate that the area taken up by that file is now available The first character of the file name is changed to an E5 But that's it The file is still there If you create a new file chances are that you will write over the old one but if you erase the file and then remove the diskette from your computer you have not done much of anything to assure privacy If someone decides to go searching through the diSk the FAT could not be used but the name in the directory could probably be recreated and the Nov 84 Editing Files Suppose we decide to edit the first of the two files on our diskette and in fact make it longer What happens I'll take a fil e known to be longer give it the same name as the first file now on the diskette and we will see what happens Looking at the new directory Figure 6 it would appear that the date and time have changed and the file size is bigger The starting cluster did not change Now lets look at the FAT Figure 7 The starting cluster is the same as before but look at offset #15 Remember we used to have OFFOFF which translated to an end of file If we look at the entry there now by using our calculations we see that this new longer file takes up the same spaces that the older shorter one did and then routes itself around the second file 2JULY BAK This is an interesting and important feature of DOS 1 1 A less sophisticated operating system such as the UCSD p-System would always look for a cont iguous area I arge enough to ho Id the ent ire file If such an area were not found the user would get an 'out of space' error when in fact there may very well be more than enough room DOS however would not do this CRYPTOLOG Page 18 FeR 8PPI8I r15 BSB 8lfbY CIO paR epPieiftL 4009933 Side 0 Track 0 Sector 4 1at ector of Directory 50524F4A 45435420 42414BOO 84080200 808GBP0000 324A554C 00000000 0000ABA4 GBP3081000 F6F6F6F6 F6F6F6F6 F6F6F6F6 E5F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 E5F6F6F6 F6F6f6F6 F6F6F6f6 F6F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 E5F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 E5F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 00000000 59202020 80100000 F6F6F6FS F6F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 Figure 6 - 3 OMLf 00000000 42414800 E5F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 ESF6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 E5F6F6F6 The Ne Side 0 Track 0 Sector 2 lat aector of File Allocation Table FFFFFF03 4000 60 00078000 09AooOOB 000000GBP0 11200113 4001fF6F 01178001 19A0011B C0011DEO 21200223 40022560 02278002 29A002FF OFOooOOO 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00002791 PROJECT BAK ooooooooooo o oo BAK oooooo 00000000 ooooooooo 2JULY F6F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 It' o F6F6F6F6 F6F6F6F6 F6F6F6F6 F6F6F6F6 It o o o o o o It It It It It It It o o o It It It It o o It It CRYPTOLOG It It It It o o It o o o It It It o It It It It It It It It It It It It It It It It It It It It It It I t It It It It It It It It o It It Directory I ooOFSool ooo oooo ' oooooooooooooo Poo 011F0002 It o lit 0 oo 00000000 I tl ' 00000000 00000000 00000000 00000000 00000000 It 00000000 00000000 00000000 00000000 It It It It It It It It It It o o o It o It o It It It It It It It o o o o o It It o o o It o o o o o o It It o o o o o o It It o o It It It It It It It It It It It It It It It o o It It It o o o o o It o It o o It o It It It It It It o o o o It o o It It o o o o o o It o o It It o o o o It It o o o It It Page 19 It It It o o It It It It It o Figure 7 The New FAT Nov 84 ooooooooooooooooooooooooo It It It It It It oo It It It It It o It o It It It o It o o o It o o It It o o o It It It o It It It It It It It o o o It It o o o o It It It It It It It o o o It It o It o It It o o o o o o o o It o It It It It It It It o o It It It o o o It o It o o It 4009933 simply places the file a cluster at a time onto the disk The clusters do not have to be cant iguous Of course there must be enough clusters available It An ASCII File We've looked at an assembly language program the Boot record the FAT and the directory It would be instructive to look at a couple of ASCII files The first one that we will look at will be exact ly that--a pure ASCII file and the second one will be an 'almost I ASCII file--a file make by a wordprocessing program Figure 8 has a few phrases from The Midnight Ride of Paul Revere This is quite readable without anyone's help Notice that in the middle of the second line there is a OD and then a OA This is an ASCII 13 which is a carriage return and then an ASCII 10 which is a line feed The file in Figure 9 was created by one of the more popular word processing programs Wordstar So that we may better understand what we are looking for we will first take a look at the file as it would appear in print disc usslo11 of the essihes of l OlIparing the architecture of iero-proc essors rither than si ply relying on i beftl tIIark or tNO Ilhen ittelpting to deterline the lIel'its of tllO or more different devices III discussing thl 80S6 it should also be undtrStoocl thit lie are discussing the 80B8 Any referenee in this report to the 80S6 also ilplies the 8088 111m there are IIi fferences Looking at the translation on the right side of Figure 9 it seems almost readable The first thing that appears to be slightly wrong however is that the last ASCII character of each word is not quite right What's wrong with it Lets look at the first line Looking at the string 69 73 63 7S 73 73 69 6F EE we know that the word should be 'discussion' but where is the 'n' Lets look at the EE If we removed would have bit we which is a decimal 110 or an ASCII 'n' Since only seven bits are required to represent the standard ASCII charac ters it appears then that what Wordstar did is use the most significant bit of the last character in a word to indicate end of word In Wordstar and most other wordprocessors the typist does not push the carriage return except at the end of a paragraph As we can see from the above quote after the word 'comparing' Figure 9 line 2 there should be a carriage' return and 1ine feed At the beginning of the 5th column in the second row we see an E7 As before we strip away the left most I giving us a hex 67 which is an ASCII 103 the 'g' that we expected Next to that is a hex 20 a space and then an 8D Pulling our usual operation with the leftmas t I we would get 00 13 a carr iage return Next to that is a OA 10 a line feed then another line feed and carriage return So it appears that carriage return line feed combinations which are inserted by the wordprocessor have a 1 in the leftmost position of the carriage return byte We can verify this by looking at the end of line 10 There we have a carriage return inserted by the user--end of paragraph Notice that this time we have OD OA OD OA whereas before we had 8D OA 8D OA This is how Wordstar knows when it has gotten to the end of a paragraph If you have ever wondered why you couldn't use a standard wordprocessor to write Pascal programs you are probably beginning to see the problem Since each byte is eight bits and ASCII character representation only requires the least significant seven bits the wordprocessor has the remaining 128 characters that it can put to very good use The compiler cannot always recognize these characters In the case of Wordstar there actually is a mode which will allow you to write programs but to allow this Wordstar has to sacrifice some of its versatility CRYPTOLOG FBR BFF IB IAt significant 110 1110 EE 1110 1110 Nov 84 the most as E Page 20 EURlift ' aCID 4009933 Side 1 Track 5 Sector 5 lat sector of TESTREC 4C697374 656E206D 79206368 6F752073 68616C6C 20686561 646E6967 68742072 69646520 72652GBP00 OA4FGGBP20 74686520 GC20696E 20373S2C OOOA6861 73206E6F 7720616C 69766500 72732074 68617420 66616D65 79656172 2EOOOAOO OA486520 66726965 6E642C20 69662074 61726368 000A6279 206C616E 60207468 6520746F 776GBP2074 2061206C 616E7465 726E2061 626S6C66 72792061 7263680D 68204368 75726368 20546F77 696C6472 656E2061 6E642079 Liaten ay children and y 720DOA6F 6F662050 31387468 72646C79 OA77686F 6F757320 73616964 68652042 64206F72 6F6GBP6967 6C6F6674 OA6F6620 65722061 66207468 61756C20 206F6620 20612060 20726560 64611920 20746F20 72697469 20736561 6874000A 20696E20 74686520 73206120 65206069 52657665 41707269 616GBP2069 65606265 616E6420 68697320 73682060 2066726F 68616E67 ou ehall hear ooof the ai dnight ride of Paul Revel reooo On the 18th of Aprii 1 in 75 oo hardly a an 11 a now alive oo who reaeabel ra that faaeoua day and o year ooooo He aaid to his friend if the British a arch oo by land or aea fro oo the town tonight oo hang 74686520 o a lantern'aloft 1n the 4E6F7274 belfry arch oo of the Mort 7369676E h Church Tower as a sian Figure 8 - An ASCII File Side 1 Track 0 Sector 5 3rd sector of PROJECT BAK GBP9736375 7373696F EE206FE6 207468E5 746965F3 206FE620 636F6D70 6172696E E5206172 63686974 65637475 72E5206F 70726F63 6573736F 72F32072 61746865 GBP96D706C F920726S 6C79696GBP E7206FEE 656E6368 606172GBP8 206FF220 7477GBPF20 656D7074 696GBPGBP720 74EF2064 65746572 20606572 6974F320 6FE62074 77GBPF206F 72GBP52064 69666665 72656EF4 20646576 OA202020 202049EE 20646973 63757373 38303836 AC2069F4 2073686F 756CGBP420 756E6465 7273746F 6FE42074 6861F420 OA8DOA64 69736375 7373696GBP E7207468 20416EF9 20726566 6572656GBP 63E52069 206GBP6563 GBP7208DOA GBP6206069 F2207468 20E1208D 776865EGBP 60696GBPGBP5 F2208DOA 69636573 696EGBP720 616C73EF 77E52061 E5203830 65737369 SDOA7468 63726F2D 61EGBP2073 OA8DOA62 20617474 207468E5 8DOA6D6F 2EOOOAOO 7468GBP520 2062GBP520 72E52080 3838AE20 GBPGBP207468 69F32072 lacussio o tho necesal tie o co parin ooooo th oo architectur o oo1cro- processor rathe tha s iapl relyln o oooooo b ench ar o two whe att eaptin t deterain th oo aerit o tw o oooooo0 r differen devices ooooo I d1acuasin tho o 8086 i ahoul ala b o understoo tha w ar ooo oooo dlecuasln tho 8088 oo o An ref renc 1 thi r Figure 9 - A Typical Sector from a Wordproces8or BIBLIOGRAPHY CORCLUSION There you have it--a walk through a DOS 1 1 diskette There was nothing very hard or compI icated about what was seen or done and hopefully the reader has ained s me inform -tion about an object that s very mportant n their professional life--a 5 1 4 inch diskette Nov 84 Microsoft Disk Operating System IBM Corp Boca Raton FL 33432 1982 Norton Peter Inside the IBM-PC Robert J Brady Co Bowie MD 20715 1983 The Norton Utilities Peter Norton Santa Monica CA 90403 1983 CRYPTOLOG Page 21 FQll QFFHHA J YSE em Y aCID 4009933 HUN liN P L byl any people tend to think that human factors specialists produce nothing but reports papers and advice usually too late and largely unheeded A recent article in the Human Factors Society monthly Bulletin vividly counteracts this limited impression M Buried within human factors R D work are lots of products that have value that will solve real problems and that can and will command the respect of the 'hard' disciplines o For too long we have too often been known as the discipline that produces nothing more than long boring cumbersome technical reports And you know about technical reports--they usually end up on the shelf never to be read A survey made by the author William B Askren Air Force Human Resources Laboratory Wright-Patterson Air Force Base revealed a wide range of pract ical products within various projects carried out by his organization and others He lists them as follows in order from concrete to less tangible prototypes mockups drawings handbooks training curricula training manuals users' guides algorithms models task analyses classification schemes software methodology questionnaires analysis formats requirements regulations standards spec ificat ions costing data human capacity information evaluation criteria proposals recommendations plans alternatives definitions descriptions examples illustrations hypotheses concepts relationships FQ lp13 quirements in such a way that specific produc t s are ident ified Secondly each produc t should be made a separate deliverable to the customer If a technical report is to be prepared as part of the effort it should be thought of as an adjunct to the products I heartily agree and I believe a large part of the problem arises in the total ignorance on the part of most people about the meaning of human factors and what its practit ioners can do At our Agency in part icular most people if they have heard of human factors at all seem to think of it as concerned with the subjective aspects of terminals furniture etc Human factors is everything that is left over when you take away hardware software and cost Human factors these people think is restricted to vague emot ional things 1 ike whether a chair is comfortable a screen is readable or a console is attractively styled We really need to do a strenuous job of education to demonstrate to people that human factors can provide concrete useful products that make a difference to the success of a project In fact at our Agency a great many things that should be done with the advice and aid of human factors specialists are being done by engineers and technicians without the faintest regard for any but technical considerations Many things that are defined as hardware and software are really basic human factors issues and problems For example what about control of the cursor on a VDT Our managers define this as software or hardware because it is implemented in a board or program But isn't the behavior of the cursor a central issue for the USER of a system I think cursor behavior is a HUMAN FACTORS issue not just software or hardware What do you think Askren continues I would propose that it is incumbent upon those of us engaged in human factors R D projects or application efforts to help our customers define their research reNov 84 86-36 CRYPTOLOG Page 22 QHISIA15 l SI i 81115Y ern 4009933 SOME RECENT ARTICLES AND PUBLICATIONS Readability of Text Scrolled on Visual Display Terminals as a Function of Window Size R L Duchnicky and P A Kolers Human Factors Journal Vol 25 No 6 December 1983 pp 683-692 ABSTRACT Readability of text scrolled on visual display terminals was studied as a function of three different line lengths two different character densities and five different window heights either I 2 3 4 or 20 lines All three variables significantly affected reading rate but to markedly different extents Lines of full and two-thirds screen width were read on average 25% faster than lines of one-third screen width Text appearing at a density of 80 characters per line was read 30% faster than text in a format of 40 characters per 1 ine Text appearing in windows four lines high was read as efficiently as text in 20-line windows and text in one- or two-line windows was read only 9% more slowly than text in 20-line windows Comprehension of the passages did not vary as a function of window size indicating that subjects maintained a constant level of comprehension by varying their reading rate Impl icat ions of the resul ts for mixing text and graphics and for limited-capacity electronic displays are discussed Ergonomics in Technical Communication Peter Martin IEEE Transactions on Professional Communication Vol PC 27 No 2 June 1984 pp 62-64 ABSTRACT The technical communicator must be a 'human engineer' because a communication like a computer system can be a technical success but a practical failure if its design discourages its use Four parallels between technical communication and user-friendly systems are discussed I Its overall structure should be apparent to the user 2 it should be congenial without being chatty or too personal 3 its nomenclature and syntax must be consistent throughout all functions and 4 its logic must not trap users in loops but should lead them straight to their goals Not Seeing is Relieving Invisible Writing With Computers Stephen Marcua and Sheridan Blau IEEE Transactions on Professional Communication Vol PC 27 No 2 June 1984 pp 69-72 ABSTRACT Word processors with adjustable video screens are used to test students' writing facility with and without visual feedback Lack of visual feedback seems to have a saluNov 84 FQR tary effect in forcing concentration onto emerging thoughts rather than words and sentences It also prevents interrupting the train of thought to edit and correct every few sentences 'Local editing' is regarded as counter-productive because it deals with individual words and lines instead of reconceptualizat ion and reorganizat ion For some students however the lack of feedback per se was disturbing and thought disrupting NOTE This is a really intriguing paper with some provocative concepts Especially interesting are the subjects' quoted comments on the inabil ity to see what they had written and the effects on their thinking MED A Study of Revising and Editing at the Terminal Jean A Lutz IEEE Transactions on Professional Communication Vol PC 27 No 2 June 1984 pp 73-77 ABSTRACT This paper presents the results of a comparative study of experienced writers edit ing and revis ing at a word processor and with pen and paper Seven writers performed four writing tasks each of which provided several kinds of data Among these are statistical analyses and graphs of a chronological record of changes made in both modes and summaries of interviews with the subjects after each writing task Results suggest that the word processor directly alters a writer's composing style NOTE this is another highly interesting paper It brings up a point that I have never seen expressed before but which seems to agree with my own subject ive experience over many years of composing revising my own writing and editing others' writing both on paper and at a terminal It appears likely that the terminal provides a somewhat smaller more linear and more restricted psychological window into a document than a book or sheets of paper It functions more like a tape and less like a random access memory This is no problem for most writers Jith their own compositions since they have a relatively clear concept of the structure at a high level When editing others' papers however or using reference material s the smaller mental window can be a hindrance Even in my own writing I often prefer to print out a draft and make final revisions on paper to get a feel for the document as a whole If the terminal does indeed prov ide a smaller more linear psychological window for the user in contrast to printed books or pages there are important implications for automation of some of our analytic tasks I recommend this article MED CRYPTOLOG Page QFFIQI' ijSK QNbJPY 23 aCID 4009933 F8R 8FFl8EAb eBEl 8tfbY 1 4 c L 86-36 CAN YOU TOP THIS CUI o from E Leigh Sawyer B 1961 End of a lengthy TDY to sites in Europel IReady to go home Pan Am flight out of Orly All on board Capacity load of passengers freight mail and fuel 707 trundles out to end of runway Swings around for takeoff Final checks Can see all sorts of funny control devices in wing sliding back and forth going up and down and so on Observe that wing is remarkable hunk of engineering All set now Jets on full power Brakes released See man near runway with hands over ears Must be dickens of a noise out there Pretty close to rotation now Plane lift s off St range thing No familiar clunk noise as soon as wheels leave runway Stranger thing Angle of climb barely perceptible Terrifying thing Plane airborne 10 to 15 seconds and still only 20 feet off the ground More terrifying thing Plane is yawing badly left and right First reaction is rapid calculation of ongoing insurance and special flight insurance to know how much wife will get Total acceptance This is where I buy the farm For the next half hour pilot earns pay for next five years Wins fight to claw plane up to 2 000 feet Intercom comes on Pilot's voice Cool confident The father image Daddy 's taking us for a ride today Lost engine on takeoff beyond rotation Airplane didn't want to fly Sorry for inconvenience Will circfe Orly for next hour dumping fuel Landed safely Passengers wined and dined during engine change Duty free 1 iquor store open Most of passengers were floated on board three hours later Seemed very carefree I passed out before takeoff Woke up over middle of At lant ic Pilot comes to mind Pushing 60 Gray hair Probably logged over 1 000 000 miles Thank you pilot God bless you Nov 84 CRYPTOLOG Wayne Have you ever considered the endless number of unusual experiences both humorous and serious that have befallen TDYers allover the world They would make for a very entertaining book How about a Cryptolog feature called something like Can You Top This If people responded to it it could ultimately make for a very intriguing collection of anec- dotal tidbits Attached is a case in point E Leigh Sawyer Page 24 FQR QFFISE 'm fBEl 8RbY 4009933 FeR eFFfefAh HB eMhY P L 86-36 PERSONAL COMPUTING IN AGROUP IUJ byl IA63 U This paper is a slightly expanded version of a paper given at the CISI Spring Conference in Hay 1984 Introduc t ion Relevancy of the topic The subject matter of this paper--iSIParticularly relevant for at least three related reasons First personal computers are relatively new at NSA In fact they arrived on a large scale less than a year ago beginning around November 1983 Second they are numerous It is estimated that' there are over 2 000 personal computers at NSA at present Third because they are new and because there are a lot of them questions are being raised regarding their usefulness and their capability' of being applied to every day agency problems as well as their effect on the organizations that have acquired them Currency of the data Since they are so new at NSA it is not illogical' to assume that any study concerning personal computers that is undertaken at this time will find the data in a state of flux It is therefore worthwhile to note that all the data that was gathered for this study is current as of April 1984 and is subject to change Definition by features of 'personal computer' In order to properry-discuss any subject it is useful to define it in sufficient terms so that any misunderstanding resulting from its mention will be minimized Accordingly for the purposes of this paper we define 'personal computer' to be a single user micro computer programmable by the user and Nov 84 with a price under $10 000 The three essential features in this definition as discussed below effectively define the set of 'personal computers' to include systems such as IBM PC IBM PC XT TRS 80 and Apple II to mention just a few On the other hand examples of systems not members of the set include the Hewlett Packard 9845 Wicat Systems l50WS and the Xerox 860 The first does not meet the set criteria since its cost is in the $30 000-and-above range The second is a multiuser microcomputer system with a base price in excess of $10 000 and thus it fails to satisfy two out of the three set criteria The last system does not qualify as a member of the set since it is a dedicated word processor and not programmable by the user Single user The 'single user' feature means that the resources of the computer are not shared and are for the exclusive use of a single user under normal circumstances Al though there is at least one operat ing system for microcomputers out on the market that is advertised as a multiuser system Venix a UNIX look-alike the vast majority of personal computers are not equipped with the required hardware or software to handle more than one user at a time CRYPTOLOG Page 25 FeR epPfeH h HBB Emu ern 4009933 Average cost under $10 000 Although the 'price under $10 000' feature seems arbitrary at first glance there is some supporting evidence for such a choice In the April 1984 issue of PC WORLD 46 personal computers were compared in terms of cost and hardware features They ranged in price from a low of $1995 to a high of $14 995 I t is notable however that only one system out of the 46 reviewed cost more that $7300--the GRiD Tempest a briefcase portable with 256K of RAM 384K of bubble memory and a 300 1200 bps modem to list just a few features The average cost for all 46 personal computers reviewed was $3723--well under $10 000 Programmable The 'programmable by the user I feature was included as part of the definition to exclude turnkey systems dedicated to a specific function--such as standalone word processors Normally these kinds of microcomputers do not have the software necessary to allow the average user to write and execute programs on them Collection of the data An informal phone survey was use to determine the organizations within A Group that have personal computers A questionnaire was then prepared that covered the following topics 1 plications is beyond the scope of this paper Topic arrangement The topics covered in thi per are roughly the same as those covered in the questionnaire Section 2 covers the distribution of personal computers in A Group Sections 3 and 4 deal with off-theshelf and custom software and the way in which they are applied On the other hand the final sections deal more with the human aspect sect ion 5 examines ope rat ional polic ies for personal computers section 6 takes a look at users' problems and concerns while in section 7 user support for personal computers is investigated 2 Distribution of personal computers in A group This section looks at several aspects relating to the distribution of personal computers throughout A Group the distribution of current hardware in relation to the distribution of proposed hardware the ratio of users to personal computers the types of personal computers and the quantity of each the functional break-down of personal computers and the distribution of personal computer usage type and quantity of personal computers 2 off-the-shelf personal Figure 1 computer software used 3 applications on personal computers 4 ope rat ional po licies regard ing personal computers 5 problems and concerns relat ing to personal computers and 60 pes In A Group By Office r - -c u' ' ' o n d p 'O p O d 6 du 'P m ' _ Al 6 personal computer hardware support and software The questionnaire was then applied in personal interviews arranged with each organization in A Group that was determined to have personal computers A total of 34 organizations were interviewed IZZI Totols 92 current 29 pr posed o 221 Current pes Propo ed PC I Although the results of the interviews revealed a fairly large proportion of interesting and innovative applications using personal computers a detailed description of these apNov 84 CRYPTOLOG Page FQR QHISIP r lH Ei eUbJPY 26 I aCID 4009933 FQR QFFISLA 15 18K QUhY Current and proposed quantity of personal computers In Figure I the data Show that A Group currently has a total of 92 personal computers with another 129 systems proposed for the next six-month to one-year period It should be pointed out that some organizations indicated that they would like to acquire a quantity of agency standard terminals AST in the coming year Since however the AST has not yet been officially selected it remains an unknown and for this reason data of this type were not included with the data for proposed acquisitions The distribution by Office is as follows AO currently has two personal computers and would like an increase of two in the next year for a total of four Al has at present 28 and desires 22 additional units for a total of 50 There are currently 14 personal computers in A2 and 28 new units are desired for the coming year for a total of 42 A3 would like to increase its current six unit s by three for a total of nine A4 presently has two personal computers and would like 14 units in the coming year AS with four personal computers at present would like to add 32 new units in the coming year Finally A6 current ly has 36 personal computers and would like 10 additional units in the coming year AD AI A2 A3 A4 AS 3 users 2 personal computers 1 5 1 141 users 28 personal computers 5 1 50 users 14 personal computers 3 6 1 12 users 6 personal computers 2 1 3 users 2 personal computers 1 5 1 7 users 4 personal computers 1 8 1 A6 96 users 36 personal computers 2 7 1 Types of systems and their distribution The data in Figure 3 show that there are only five types of personal compute s in A Group that satisfy the criteria outlined in Section 1 3 above These are IBM PC 74 units IBM PC XT 9 units Radio Shack TRS 80 5 units Apple II 2 units and Tri-information Systems BTS 2000 2 units Typical equipment configurations In Investigating the equipment configurations for each of the five types of personal computer it will be not iced that color monitors are conspicuously absent One of the main reasons for this absence seems to be that color monitors currently are TEMPEST approved only for zones 4 through 5 Another general trend due to the TEMPEST character of the equipment is the almost total lack of hard disk drives There are only seven personal computers in A Group with hard disk drives at present Numerical comparison of users to personal computers The data indicate that hroughout A Group there is an overall user-to-personal computer ratio of 3 39 to 1 with 312 users for 92 personal computers As shown in Figure 2 these data are broken down by Office with the greatest concentration of users and personal computers appearing in AI A2 and A6 The totals are as follows Figure 3 Types Of pes In A Group 92 T lol Uliits APPLE II 2 2 _--r ___ Figure 2 pes In A Group By Office Num rlcal 150 Comparison or Us rs to p s A' 140 130 120 110 GBP01 4 P L 86-36 '00 90 eo 70 60 O A2 AO 30 20 10 0 - A Group totols 312 us-ers 92 pes Users pes Nov 84 CRYPTOLOG Page 27 fOR OfHSPJ IH K QUhY IBM PC BO ''' aCID 4009933 FQ QFFISI QSB QN JPY configurations include 2 floppy disk drives and a printer--usually a dot matrix graphics printer Typical equipment configurations for each of the five types of personal computer are listed below Figure 4 PCs In A Group By Office 50 - -- St Q - I Qn cv o o u m b T ' m ln I cu oo IBM PC Most of the IBM PCs have 2 floppy drives l printer a monochrome monitor and l28K 256K 384K or 5l2K of RAM--depending in part on how they were acquired Those units acquired through the PCIC come with 512K of RAM while personal computers acquired by sole source or competitive bid vary in the amount of memory and type of peripherals according to the organizat ion I s requirements at the time IBM PCs used only as dumb terminals do not in general have printers - A6 t2Z J Dumb terminal IBM PC XT As a general rule the standard configuration for IBM PC XTs is SI2K of RAM 1 floppy drive I-10M hard disk drive 1 printer and a color monitor As indicated elsewhere seven of the nine IBM PC XTs have color monitors Present and future acquisition methods Three principle means of acquiring personal computers have been used in A Group Radio Shack TRS80 Aside from the usual pcrc per configuration of 2 floppy drives one has 3 floppy drives 1 printer and a mono- 1 chrome monitor the TRS 80s have either 48K or 128K of RAM 2 sole source procurement and Apple II While both Apple IIs hav'e 2 floppy drives and a printer one has 48K of RAM and a monochrome monitor and the other has l84K of RAM and both a color and a monochrome monitor BTS 2000 Both BTS 2000s have 64K of RAM 2 floppy drives 1 printer a monochrome monitor and 3 communications ports Functional break- The question posed here was how many personal computers are being used solely as dumb terminals as opposed to stand-alone computers As indicated in Figure 4 there are personal computers in AI A2 and AS that are used only as dumb terminals 86% or 79 of the current personal computers in A Group are being used only in stand-alone mode while 14% or 13--8 in AI 4 in A2 and 1 in AS are being used as dumb terminals Nov 84 FS 3 lo n from another organization 64% of the organizations interviewed indicated they would not change their method of acquiring personal computers in the coming year Use of personal computers daily The data in Figure S indicate that the average time per day that personal computers are used in A Group is 5 77 hours Personal computer usage broken down by Office is most in AI A2 and A4 It should be noted that a contributing factor to the high daily average in Al is the fact that six personal computers are being used as dumb terminals 24 hours per day The Office averages are as follows CRYPTOLOG Page 28 SFFI81M BSI3 SUbY aCID 4009933 Figure 5 4 Applications PC Use In A Group By Office II -r-_ _---- -_ _ c A c V ' OO o c d c 0 ily-'-'PC-'-'U ' - 10 e A Group A r0ge AO AI A2 A3 A4 A5 A6 17 Three general categories of applications are examined in this section applications with off-the-shelf software applications with custom designed and written software and applications which interface off-the-shelf software with custom written software It should be noted however that there are currently' only two examples of the third category in A Group For this reason as well as the fact that one of these appl icat ions uses only off-the-shelf software Figure 7 treats them as commercial software applications The purpose of the data arrangement in Figure 7 is to show how organizations use their personal computers Notice that 50% of the 34 organizat ions interviewed have appl icat ions using off-the-shelf software only while 35% of these organizations have applications using both off-the-shelf software and custom-written software Only 6% have only custom-written applications while 9% have no applications Organizations with no applications are in general using their terminals as dumb terminals only H Oo l 4 5 hours per day 10 6 hours per day 9 3 hours per day 2 hours per day 8 hours per day 2 6 hours per day 3 4 hours per day 3 Off-the-shelf software Figure 6 shows the off-the-shelf software that is currently being used in A Group Organizations were also asked to subject ively rate the software they use on a 1 to 10 scale of poor to excellent o The results are shown in column three Note that the software with the greatest number of users seven or more is available through the PCIC Looking at the packages exclusive of the operating system MS DOS which have the greatest amount of users it is interest ing to note that 56% of the dBaseII users had specific complaints--which may account for its 6 8 overall rating On the other hand only 33% of Word St ar users and 37 5% of Lotus123 users had complaints about these packages The complaints referred to were in response to question number 10 of the questlonnaire what needs of your organizat ion are not being met by current off-theshelf software Applications with off-the-shelf software Off-the-shelf software is being used in A Group in three principle ways for text processing for management or project tracking systems or for data base applications Text processing The software being used for this type of application is Word Star Volkswriter Easy Writer Edix Wordix and Professional Editor At least four organizations indicated that they use Word Star as a program editor while one organization uses it to produce graphics-like charts Tracking systems GBPE management project The software used 'in this type of application is Lotus123 Multiplan and Supercalc--the spreadsheet packages In general these applications are relatively small Data base applications For the purpose of this y a data base on a personal computer is classified as small less than 50 000 bytes medium 50 000 to 175 000 bytes and Nov 84 CRYPTOLOG Page 29 FQR QFFUlIM T iSI i QUb Y aCID 4009933 FeR eFFI6IAf J BSE 8t1f JY Figure 6 PERSONAL COMPUTER OFF-THE-SHELF SOFTWARE IN A GROUP Average Rating by Users on a 1-10 Scale software name o o o o o o hardware APPLE DOS APPLE PASCAL CONDOR CP M DBASE II DDIT EASY WRITER EDIX WORDIX LATTICE C LOTUS 123 MAXIMANAGER MS DOS MS DOS BASIC MULTIPLAN NORTON UTILITIES OPEN ACCESS PC CRAYON PC FILE 3 PC IX PFS FILE PFS REPORT PL I-86 PROFESSIONAL EDITOR PROFILE 3 PLUS QUICKCODE QUNIX QUNIX C R BASE 4000 SCRIPS IT SUPERCALC SUPERSCRIPSIT VOLKSWRITER WORD STAR rating APPLE APPLE IBM APPLE IBM IBM IBM IBM IBM IBM TRS 80 IBM IBM IBM IBM IBM IBM IBM IBM IBM IBM IBM IBM TRS 80 IBM IBM IBM IBM TRS 80 IBM TRS 80 IBM IBM Nov 84 CRYPTOLOG number of responses 8 8 5 7 8 7 6 8 9 4 5 8 7 3 8 8 8 7 7 8 6 6 9 8 6 8 8 5 1 2 2 1 16 1 2 1 5 8 1 22 7 7 1 1 1 1 2 7 1 7 8 6 10 1 3 2 1 1 9 8 3 7 2 9 2 8 3 6 3 1 7 8 FQil QFFHJlSb ijSI3 2 1 7 9 2 Page 30 8t1f JY 21 aCID 4009933 FeR epPfefAt H3 OMLf Figure 7 APPLICATIONS WITH PERSONAL COMPUTERS IN A GROUP 34 divisions or branches interviewed I N - H 0 USE C U S TOM - W R ITT E N A P P LIe A T ION S none multi one ----------- ----------- ----------- none 3 ----------- ----------- ----------- COMMERCIAL SOFTWARE o 2 one I' 8 II o 2 I I ----------- ----------- ----------- APPLICATIONS I multi I 9 7 I I 3 I II I ----------- -------- -- ----------- TOTAL ORGANIZATIONS 20 Nov 84 CRYPTOLOG 11 Page 31 i'QR QFFIQI 4 15 YSE 8lftJY 3 34 4009933 FeR eFFI8Ikb HSE eMbY large over 175 000 bytes For the data base software currently being used it is a given that a data base is limited to the size of the disk--with the exception of those systems with hard disks this means an upper limit of about 300 000 bytes exclusive of any indexes The software used is mostly dBaseII with a very small number of applications using Condor and R Base 4000 dBaseII applicat ons range from small data bases to large data bases with typical applications being inventory control personnel tracking job profile tracking and clearance tracking For organizations using dBaseII ten have small data base applications two have medium data base applications and four have large data base applications Applications with custom designed and written software The tVaried applications those that are custom designed and written using a programming language such as BASIC C or Pascal Although as Figure 7 shows only 41% of the organizations interviewed have written any custom designed software they are broken down into seven functionally different applications These applications are categorized by Office as follows are AI a text editor written in C and two separate device control software projects written in C and BASIC respectively A2 collection processing and reporting in the field 5 Operational policies and procedures In general it seems that there is an absence of formal operational policies regarding personal computers due to the lack of large enough concentrations of personal computers in anyone organization The two exceptions in A Group A4 and A6 are discussed below The remainder of A Group organizations have in general some informal procedures they follow that deal with security and training for the most part Informal security procedures include periodic back up of files locking up data diskettes at night ensuring that the personal computer is turned off at night while training involves either getting a manual and learning by trial and error using an automated tutorial if available or having an experienced user tutor the new user On the other hand both A4 and A6 have formal written operational policies A4's is st ill in draft form The topics covered in the draft version operational policy of A4 are security standardized files procurement of software supplies management of hardware and assistance A63 with the largest single concentration of personal computers in A Group has two operational policies one dealing with installation and the other with operation The IBM PC Hardware Software Installation Plan covers installation phases installation personnel installation schedule equipment control identification of standard software package software distribution plan identification of IBM PC supplies to be stocked and training schedule material and categories A3 signals analysis software using Pascal on an Apple II A5 graphics written in C A6 text format translation from personal computer format to a format compatible with PWB written in BASIC and collection management software written in C Applications interfacing different software packages Al has a budget planning system that interfaces 10tus123 Word Star and dBaseII A6 has an educational software appI ication that interfaces dBaseII with custom software written in BASIC Nov 84 FOR The IBM PC Hardware Software Policy deals with the following topics CRYPTOLOG nI I hardware management acquisition and distribution of supplies maintenance software evaluat ion software management training and security Page 32 'lJ ISS enr y DOCID 4009933 FeR eFFIe Mb 9Bf 6fft Y 6 Problems and Concerns around The final question on the questionnaire asks the interviewee to list any problems or concerns the organization has in general regarding personal computers Out of a total of 34 organizations interviewed 25 indicated that 'they had such problems and concerns A total of 66 problems concerns were listed for an average of 2 6 per organization The responses fall roughly into six categories acquisition and installation of hardware and software training and support general operation and maintenance of hardware human interface with the hardware and software intermachine communications and security Acquisition and installation of hardware and software Sample responses were no dust covers are provided or are generally available not enough personal computers are available wait ing for the AST impedes ae quis it ion of personal computers which it may render obsolete there are too many bureaucratic levels to go through in the procurement process supplies are hard to get and there many delays there is no mechanism for upgrading the personal computer hardware software evaluations need to be centralized published and distributed on a regular basis to the users General operat ion and maintenance of personal computer hardw Typical problems listed in this category were the graphics printer is too slow personal computer response time slow and there is not enough memory or peripheral storage for some applications is too Human interface with personal computer hardware and softwar ome sample responses in this category were the keyboard is poorly designed and therefore difficult to use accurately images on the monochrome monitors are fuzzy and difficult to read which causes eye strain the available software and documentat ion are too complex for user requirements and difficult to understand there is not enough space to put the personal computers the graphics printer is too loud and disturbs co-workers and several personal computers in the same area give off enough heat to make the room uncomfortably warm Among the problems software manuals are difficult to understand and use personal computer training and support is not centralized thus causing the user to waste time and effort learning by trial and error training at the branch and division level needs to be conducted regarding the usefulness of personal computers and are there is a need for color graphics but the standard configuration does not have a color monitor and Training and support I isted were Intermachine communication This category includes problems expressed concerning networking as well as communications to a host computer Sample problems listed were networking is not being addressed despite a requirement for it thp reliability of file transfer between a host and personal computer is doubtful there is an overall lack of planning for training and support and not enough to go Nov 84 CRYPTOLOG Page 33 F8R 8PPleIltb 9SFi 8fft Y ern 4009933 I there is a need to have hook up to main frames J there needs to be an efficient quick method for inputting large quantities of information from a host computer and I communicat ions there is a need for Xerox 860 compatibility Security Problems category included listed under According to official A63 memoranda dated 22 February and 1 May A63 is equipped to provide the following support to A Group personal computer users I hardware software consultation I software demonstrations and training I hardware demonstrations and training I off-the-shelf software evaluations and a personal computer users group this there is a need for a removable hard disk capability organizations fear the compartmentalized data there is no password security and I TEMPEST should be done before acquisition of hardware proli ferat ion of The PC users group meets once a month for an hour and is an excellent opportunity for A Group users to exchange ideas and discuss mutual problems In addition to these services A63 currently provides personal computer applications development support to A606 and A2 8 Conclusions 7 Support The questions addressed in this section deal with where A Group personal computer users currently get hardware and software support if any and the existence of a local centralized support within A Group Sources currently cited GBP l Group Figure 8 shows that seven sources for personal computer support were cited These include I individuals in A Group' Al09 A63 B6 the PCIC T2 and T3 Interestingly enough eleven organizations 32% indicated that they receive no support at all A slightly larger number indicated that A63 provides support-although it should be borne in mind that fourteen of the organizations intervi ewed are in A6 In other words it would seem that the role of A63 to provide computer applications support for all of A Group is not yet widely understood--at least in the context of personal computers PCs have been in A Group only a relatively short length of time and thus it should be stressed that the situation regarding PCs in A Group is transitory and will likely change in many ways T ere is naturally some reticence on the part of prospective users who are used to the power of a main frame to recognize the utility of the PCs At the same time inhouse training and software evaluat ion were not initially available to user's on an official basis and are only now starting to develop As a result inexperienced users are experienc ing some frustrat ion and hence many PCs are not being fully utilized Other reasons include the fact that most organizations do not have communications to main frames yet although they have a requirement to work with data stored on the main frame Current PC Hardware Software Support 1o T s ou ' o c t d a Y A G ' Ou P u '3 12 10 7 6 Figure 8 o 3 2 Dual sourceS' Nov 84 CRYPTOLOG Page 34 F6R 6FFfelAb l fSrl 6Ub' cjt d In some cos S _ aCID 4009933 F8R 8FFleHib aSE 81irJ'l ACTION LINE U Dear Faction Line Something must be done before it is too late We are approaching a situation akin to the critical mass in the cafeteria I refer to the to and fro gang You can see them any day wandering about looking for a table knuckles growing white holding their trays as desperation assails hem their eyes flickering endlessly in 360 sweeps and their food so near yet so far growing cold What is needed are bold and innovative measures A few of these are addressed as follows Use larger forks doubling the size of forks would double intake In this regard forks should be provided in two configurations one twice the length and one twice the width in order to conform to varied oral specifications The large fork factor LFF when combined with CG R would materially accelerate consumption time CT In short LFF CG R CT The same approach should be used for spoons and when peas ar e on t he menu knives Thus we can also add both large spoon factor LSF and a large knife factor LKF Therefore even when a person has soup and a hot plate with peas it still obtains that CG R Make people eat faster This may be subdivided into two categories 1 chewing and 2 gulping It is suggested that M7 monitor chewing and gulping norms as they currently exist on the part of those eating in the cafeteria Based upon physiological stress standards as they affect the alimentary tract i t is believed that more rapid chew and gulp rates may be achieved Special courses or seminars to this end could be established in order to enable personnel to meet newly prescribed standards for chewing and gulping In order to raise motivations a new skill field in this area should be established which could lead to certification and even bonuses e g FLAP Food Loading Acceleration Pay Further for those-achieving chew gulp rates CG r faster than prescribed compensatory leave should be allowed This could be achieved by a time card check in check out system at the entrance to the cafeteria or even possibly tied into the new badge system by assigning everyone a GIN GBPastron omic Intake umber Nov 84 LFF LKF LSF CT Use higher tables or lower chairs This lessens the distance the arm must travel between plate and mouth The optimum distance is suggested as two inches as opposed to the more normal distance of 18 inches This would increase by a factor of 9 the rate of food intake 9 x RFI When combining RFI with increased CG R and the realization of LFF LSF LKF CT should be reduced to approximately three minutes a savings of 27 minutes based upon current standards When calculated against a loss of $300 000 000 per annum for the nonproductive cafeteria time of 15 extra minutes a day beyond the prescribed half hour the savings of 27 minutes a day would add $540 000 000 of production rate per annum beyond the normal work year of the federal establishment With perhaps $40 000 000 of this employed to subsidize free antacids the new cost-out of increased production rates would still be substantial Lets move out smartly It is time to clip the wings of this octopus CRYPTOLOG Page 35 4009933 EOR OPE C A fi I l 6lftoY HeJte ' anotheJt puzzle in the cJt lptic vein In you' e no 6 with thM type 06 de Mnition a nd woldd lilz e to give J t a y call the a uthM on 7103J and you I U be e an e xp ta na t o n 06 how theA e c fu I Me put toge theJt N'SA-Crostic No 59 by David H Wilijams _ ' If - -r 16 E 1 P 32 78 95 o 17 H 18 M B H C 34 G o - T - S3 o 35 L 49 T 50 1 A 63 V 64 P 65 Ii J Y 81 J 80 o 111 W 112 U 96 o h II G 67 II 083 I 84 o o 52 Q 53 P G 40 o 116 J 117 C o - _- --133 LI 134 S 54 o _ _--M --70 -----Y 69 68 o Q 55 - - - ' ' t - - t o I o G Y 58 o 8Il'VI8 -1l I - - -- - - o ' o o ' f-- dl 167 181 l 182 K 183 W 194 R 195 B 196 r ' 100 209 P 210 B 211 V 212 J 213 Q - 9u 29 K 30 A N 4S V 41 L 159 I 60 S 61 F _ - 76 I 77 n 4 o _ - LI 75 o _ - ' I i iT Ii -- Y' -_ _ --- I - _- ' - - f ---- 71l Z _ _- I o o 171 r 172 P 173 a o -- - _ _- Page 36 i T 187 B 188 It 189 1 190 P 2ii S 219 L 145 -9- 187 176 sixth ring 4 wds 174 1 _______ 1- E Where the much bejeweled lady wore her 141 I 142 a 't '' ' 21 - f2iS'Ji16Wf-- --'2I7-1 D Ax me not where to bury this 2 wds 125X 126 C ' ISS 1 156 S 157 A 158 X 184 S 185 Y 18h C Call this Greek by phone on Roman 10 o 1I 94 G TORllr109 C 11 G 122 U 123 12411 '''-'t'''fT'-- - CRYPTOLOG J 1191 A Did Dale counterfeit tickets to battleground 2 wds B When visiting with a boor never sing Dorothy's great song 3 wds Nov 84 S 15 i 121 lh8 ' lbY D o U 14 Y 92 0 93 --1--1'-- - 152 a 15 R 154 N -- 74 105 I IDb V 147 J 148 A 149 P 150 W 151 E o G 13 N ---_ - 136 G 137 X 138 r 1 G - - -'--- S 57 ---- S 73 L 43 II 42 _ _ - I 56 I o N 4 _-- P 72 12 o ' ' - -- - r9 '1 71 T All 35 E 132 176 B 177 Il 178 2 179 M 180 P IYI G 192 E 193 C 207 a 2118 F K 39 Q 100 ' 101 X 102 A 99 159 C 160 I' 161 T 162 P 1b3 Y Ib4 F o A 38 lifO o ---- --I-- _ f - - ' - _- 85---A 1132 114 E 115 a 127 1 128 0 129 P 130 E 131 0 175 Y 11 9 1 8 117 --- 1--- _ - f - - - ------ 5 s r 98 143 E 144 Y 145 B 146 P G6 B 37 5 36 o _- --- 1-- - II 82 G 97 ---r _ r - _ _ -_ _ - 1'5 _--f- - Y 48 E 79 ---- C4 I ---'c -' - 'TJPY-c - 1 1-- 47 o 62 v dhw o peR 8PPlelj1d5 tfflB elfbY aCID 4009933 F Ease peas are for the doctor G Try to use Indian blight in indiscernible circumstances -5- 73 43 136 % H Frank camera I The New Colossus The other structure with which Alexander Eiffel was associated 3 wds J He wanted Ned Dee and Sam K Cross off King Tut's middle name L Ran into little Edward and gave him share M Bury between his N These reptiles enjoy tea amid the cardinal points O Literally Greek magma P In a state of stimulation it has a fixed period of expiration 3 wds Q Auto pioneer after whom two cars were named - the other one being the Reo ---sz 54 213 ---gg R That liver I ate made me feel very macho s That mineral I bent seemed to go on and on and on T This Latin is amid late back type of drug U Tossed with rewarding results V Enchanted he marched off W Caesar's white road X This African grain is not efficient Y When I saw rainbow effect on iced screen I cried z Los Angeles Ram's embraceable one a He laughed as he bantered with the 1African queen Nov 84 Pl-Feb 85-53-66829 115 152 142 207 173 CRYPTOLOG Page 37 123 ---g r 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