The Use of Computers in Researching the Assassination of President Kennedy

                                W. Anthony Marsh

                  The Third Decade Conference - June 29, 1991

    There are several ways that computers have been used in investigating the

assassination of President Kennedy and even more potential uses. I will outline

some of the ways computers have been used, including some of the ways I have

used computers, and how computers could be used more in the future.

    1. Statistical analysis.

      Although statistical analysis can be done by hand or with a pocket

calculator (which is really just a handheld computer anyway), using a computer

has certain advantages which give the researcher more productivity. More

calculations can be done more quickly and data analyzed in more ways more

easily. Even statistical packages which used to be available only on a mainframe

are increasingly becoming available for microcomputers. Even a beginner who is

not familiar with statistical theory could duplicate some of the statistical

tests used by BB&N in their studies of the DPD tape, given the right program and

valid data to input.

    2. Physical analysis

      Studies which would have been difficult back in the 60's may be easier to

do now with increased computer power. Spectral analysis or neutron activation

tests could be done with smaller sample sizes. Blood testing which might have

been ruled out due to the small size of the sample might now or soon be

possible. This might clear up questions such as what bullet fragments hit which

persons and objects.

    3. Fingerprint ID

      Computerization of fingerprint ID files might make it easier to find

matches with unmatched prints. Computerized searches through fingerprint

databases might be used to prove identities of suspects that was not possible

before. Eventually, computerized matching might even allow one to prove that a

fingerprint was planted or when it was left.




                                       1





    4. DNA fingerprinting

      Although some courts have not accepted DNA typing as conclusive proof,

some experts believe the tests to be 99.9% reliable. By matching DNA material in

a sample to a person, it is possible to prove things that could not be proven by

an ordinary blood test. And DNA material may survive intact much longer than a

blood sample, so even after 27 years it may be possible to learn something more

about the JFK assassination by DNA testing. For example, it might be possible to

prove whose blood covered which parts of the limousine. Was it only JFK's blood

and possibly brain matter on the windshield, or was some of the blood Governor

Connally's? Or is the blood on the windshield now neither men's, proving that

the FBI or Secret Service fabricated the windshield evidence? If someone had

saved the cigarette butts from behind the Grassy Knoll, perhaps a DNA test would

indicate the general racial or national characteristics of the shooter. The

testing might become sophisticated enough and the database large enough to show

that the individual was of German extraction with a rare medical condition and

chain-smoked Camel cigarettes.

    5. Digitized photos

      Computer enhancement of photographs has been used to try to glean more

detail and clean up photographs studied by the HSCA. Sometimes enhancement can

disprove a theory by showing that what looks like a gunman is merely an optical

illusion due to the fuzziness of the photo. Or enhancement might bring out

details such as the shadow of a rifle. Digitizing photographs also allows a

researcher to compare photographs more easily, plot movements, and extrapolate

hidden objects.

    6. Digitized audio

      Because computers are inherently digital, it makes it easier to study

audio sources when you convert the typical analog signal into digital. Even the

least expensive home computers now can digitize sounds and store the information

in files. This makes it easier to manipulate and analyze the data without



                                       2





expensive equipment. Plus there is the advantage of being able to print out

hardcopy for analysis. One project that I have just begun is digitizing the DPD

tape with the Commodore 64. I hope to find patterns which suggest a previously

unidentified shot. And with the right programs, any researcher might be able to

duplicate the work done originally by BB&N or Weiss and Aschkenasy. The program

which I am currently using store information as a .RAW file which is compatible

with many other computers. Other programmers and I are working on programs which

will display and analyze the .RAW files. It might be possible to prove or

disprove that certain messages on both channels are cross-talk of the same

message, or that the bell sound is indeed a bell or that it is electrical

interference, or that the sound right before the siren on the cycle with the

open microphone is turned on is actually the arcing sound of the switch being

turned on to start the siren. That might prove that the open microphone was on a

motorcycle in the motorcade.

    7. CD

      It appears that CD players are quickly replacing records and tapes. It

would be great if someone could arrange to put the DPD tape onto CD. Or

conferences such as this one. If you remember the record that "Gallery" magazine

sent out with the article about the JFK assassination, the same type of things

could be done on a CD. A CD version would be more durable, not subject to

defects and damage, and you could cue important portions for study with more

advanced CD players which can remember and repeat tracks. There are lots of old

recordings which should be preserved before they deteriorate beyond recognition.

CDs may soon be writable for a reasonable cost so that research audio can be

archived for future reference.











                                       3





    8. CD-ROM

      A CD which holds non-audio/visual data can archive great quantities of

data and make it available to a researcher almost instantaneously. One CD-ROM

can hold the equivalent of an encylcopedia. Instead of searching through stacks

of papers in a library, one could insert one CD and search through all the

papers in a matter of seconds. Cross-indexing could allow you to find an obscure

reference that you might have missed with a manual search. All occurances of a

keyword could be located and printed out so that you could quickly see a pattern

that you might have missed with a lengthy hand search through documents.

    9. Optical Character Recognition

      New hardware allows the computer to scan a printed page and convert the

information into a form that the computer can remember. Either on a device which

looks like a copy machine or with a handheld scanner, you can pass a light over

a printed page and have the computer convert the image into a text file. This

could be useful for entering old documents into a database, rather than retyping

them.

    10. High Resolution printing

      A new generation of printers which can print many dots instead of just

lettes like a typewriter gives researchers more flexibility in handling printing

out documents. Cheaper laser printers and more powerful software give the

average and unsophisticated user desktop publishing capabilities, such as a wide

choice of fonts and styles, graphics good enough to include photographs, and

easy reformatting of text. Another project that I'd like to see developed is

reconstruction of documents. Some old documents are in such bad shape that they

can not be read. Copies of copies and deterioration can make documents unusable

to researchers in their current forms. If they could be carefully converted and

cleaned up, one could print out versions which would look almost exactly like

the originals. One way that I've used such a reconstruction is to recreate the

original text from a 1963 Department of State telegram. The original draft had



                                       4





been X'ed out in many places, but I was able to see most of the text underneath

and typed it into my word processor. Scanning would make quicker work of such a

task and perhaps OCR software could be taught to see the original text under the

X's. I had only a 9-pin dot-matrix printer to reprint this telegram, but a laser

printer might be able to reprint it exactly so that one would think it is the

original telegram draft.

    11. FAX

       Fascimile transmission allows one to send a document in almost its

original form anywhere immediately. This could be a great help to a researcher

who needs to find a missing piece to the puzzle quickly. A document could even

be sent to someone who doesn't have a FAX by arrangement with a local store that

offers such a convenience. Not only is there no waiting for the mail or

wondering if it will survive the mail, one does not take any chance of loosing

the original. Local libraries are now offering fax. This makes it possible to

get copies of articles from anywhere across the country. If you are not able to

find a copy of a rare book or obscure magazine, your local library might be able

to locate it and have it FAX'ed for you. That is how I was able to get better

quality copies of some magazine and newspaper articles from a library which had

the original material instead of microfiche.

    12. Databases

      Storage of research material in a computerized database offers options not

available to paper storage. Hand searching can be time consuming and is

inherently error prone, as fatigue creeps in and one misses otherwise obvious

details. Having the data in a computer form allows one to use the computer power

to access all the data quickly. Searches for all references to any word can find

references to documents which might have been missed before or dismissed as

irrelevant. The Assassination Archives and Research center has ongoing database

projects available to researchers, such as Mary Ferrell's index, Daniel

Brandt's, and Jeff Meek's. If all the available documents were put into a



                                       5





computer database, one might be able to find documents which the researcher

might never have thought to look for. The use of Boolean operators would make it

possible to find unusual references or narrow one's search through a maze of

references. Cross-indexing a photo collection such as Sprague's might help you

find a photograph that you didn't know existed, or track an individual's

movements through various photos, such as The Umbrella Man.

    13. BBS

      An electronic bulletin board which can be accessed from your home computer

via modem could greatly expand researchers' ability to communicate and exchange

information. Instead of sending letters and waiting weeks for replies,

researchers could log on to exchange information very quickly. Someone might

pose a question about some technical point outside that person's expertise and

find it answered by someone who is an expert in that field right away. A public

message area would allow everyone to keep up date to date on the latest research

and exchange information that everyone can benefit from. Private message areas

are quite helpful when someone wants to contact another researcher, but may not

be able to find the person home at the right time or wishes extra privacy that

calling by voice can not offer. Conferencing allows callers to 'talk' to each

other in real time rather than leave messages. Files areas allow you to store

information permanently so that anyone else can retreive it at any time. This

could be text files, documents, maps, graphics, or data files. I know there has

been some interest in setting up a BBS for researchers and I would be willing to

help set up one if someone else can handle financing it. One possible

alternative would be a national BBS, rather than a local board. CompuServe is

available in most cities, although it is expensive. Other systems such as BIX or

GEnie may have lower rates, but may not be as easy to find a local access

number. Another possibility is FidoNet. That would allow one to leave a message

on a local BBS which then sends it around the world until it is picked up by the

other local BBS. Many of the 'echoes' carried by FidoNet are particular to very



                                       6





narrow interests and it may be possible to convince the administrators of the

system to allow a JFK echo, however convincing your local BBS to carry it just

for one person may be impossible. It might be necessary to set up your own

computer to connect into the echo directly, which may become an expensive

long-distance call.

    14. Word Processing

      This may seem too obvious, but there are still people who prefer to use a

typewriter. I don't think I could ever go back to using a typewriter. Using a

good word processor with your computer allows you to do more with your text.

Underlining, boldface, compressed or expanded text, etc. can be used much more

easily if the text has been entered with a word processor. Old letters can be

reprinted to circulate copies, and letters or articles can be exchanged or

collaborated on via disc so that changes can be made and resaved easily.

Everyone makes mistakes and it is often easier to fix them when using a word

processor. Using its search and replace feature, spell-checker, style-checker,

etc. makes it easy to correct errors and update information. One problem to

watch out for though is what format is used to store the file. For example, IBM

typically ends each line with a carriage return and line feed, while Amiga uses

only a line feed so that a carriage return results in a blank line. Commodores

typically use a carriage return and may ignore a line feed, but might print a

document with only carriage returns all on the same line. Some word processors

store their native files in a non-ASCII format and you have to perform some

special operation to convert a file into a normal ASCII text file. With some

programs like Wordstar it may be as simple as turning off the high bit, or

toggling off document mode. With others such as Commodore programs, it may be

necessary to convert the PETSCII or screen poke codes into ASCII with another

utility program.







                                       7





    15. Ultrasonic measurement

      Another pet project that I'd like to see someone accomplish is to design

and build a highly accurate device to measure distances with ultrasound. There

are ultrasonic tape measures on the market for the home carpenter, but they are

not very accurate and their range is too limited. Polaroid used to sell the

ultrasonic unit similar to their Spectra camera which uses it to auto-focus, but

I'm not sure if it could be adapted to obtain the necessary accuracy. Of course,

the application I have in mind is to finally get an accurate measurement of

Dealey Plaza. There are been any number of versions of maps and placement of

objects. The inaccuracy of maps was one of the factors taken into account in the

acoustical studies done for the HSCA. Greater accuracy would allow the studies

to be duplicated with greater precision. And the movement of objects and people

through Dealey Plaza could be plotted more accurately.

    16. A map of Dealey Plaza

      Using a computer drawing program, I have drawn a map of Dealey Plaza.

Although I have hand drawn previous versions, I find it easier to use a computer

now. As I have no artistic talent, I prefer to let the computer draw lines and

curves. One problem, on the other hand, is that some line and curves will not be

drawn smoothly. This case of the 'jaggies' results from the fact that the

computer can only store and send out dots. More sophisticated programs used with

the right printers can send out commands which can draw smoother lines and

curves. Right now, my program can print out 640 dots across the page by 720 dots

down the page. So, my map has a scale of 1 dot ( or pixel) equals one-half foot.

This is enough accuracy for most of my studies, but creates a problem when

having to round out distances. One thing to watch out for when rounding is to

keep objects in nearly the same proportions as the originals. Aspect ratio is

another problem. Depending on the type of computer and monitor screen, a perfect

circle may not appear round on the screen, but may print out round. Or it may

appear round on the screen, but not print out round on paper. It may be

necessary to modify something to get the image to look right.

                                       8





      Eventually I'd like to get my map into a form that others can use with

their computers. At present there is no conversion program to change it into a

form that could be used by another type of computer. CompuServe has a format

called GIF ( Graphics Interchange Format ) which can be displayed by a wide

range of computers. I hope to be able to convert it into a GIF and then upload

it to CompuServe. An alternative format is MacPaint, but that format offers only

576 dots across and would chop off part of the right side of the map. GIF allows

up to 16,000 dots across by 16,000 dots down. It might be easier to convert it

into another format first as it is only a black and white image. RLE ( Run

Length Encoding ) is a black and white format, but only offers 256 dots across

by 192 dots down. I would be glad to convert portions of the map into that

format if anyone wants to paste together the pieces. By combining a series of

pictures of Dealey Plaza it would be possible to create an animation sequence.

Some programs allow you to flip through a set of pictures like a slide show.

Others use page flipping for smooth animation like a movieola. The new GIF

standard 89A supports animation as well as text.

    17. Echo Analysis

      I have written a program which analyzes the DPD tape impulses in much the

same way that Weiss and Aschkenasy did in their study for the HSCA. Using a

computer has several advantages and a few disadvantages. Someone who knows

nothing about the echo studies could use the program by typing in a few

variables and get a list of possible matches in a few seconds which might have

taken them hours to figure out. My program includes some range checking, which I

recommend for any such program, so that impossibly large or small variables are

not typed in. Plus, another useful feature is that it provides default values

for the variables so that anyone who is not familiar with the evidence can start

with a practical set of conditions. One limitation of my program is that it only

examines direct echoes, not compound echoes. In the HSCA studies, some of the

echo paths were found to have been multiple reflections off a couple of objects.



                                       9





Those are easier to find or rule out when doing the study by hand. Trying to do

that on a computer, you run into the problem that if you allow too many multiple

echo paths, you can find a match for almost any impulse simply by combining the

right echo paths. However, some would not be possible in real life. One leg of

the combined path may be hidden by another object or the echo would have become

too soft to be considered a good match. My current program only allows you to

study shots 1,2, and 5. No data for shot 3 is available until I can digitize and

analyze the DPD tape myself. Both BB&N and the FBI have denied my requests for

them to release any DPD tape data, so I may have to recreate the strip charts

myself. Shot 4 is the head shot from the Grassy Knoll which Weiss and Aschkenasy

have already studied. Another program similar to this which I have been working

on would reverse the proceedure by using the echoes from the test shots to look

for objects in the locations they were in at the time of the test shots. Again,

the problem here is lack of good data. I can not be absolutely sure of where the

gun was located and I have only a close approximation of where the microphones

were located. I have estimated other variables such as the temperature. If there

were enough data available about the test shots, they could be used as a type of

sonagram which would pinpoint the locations of major echo-producing objects in

Dealey Plaza. That greater accuracy then could be used to narrow the margin of

error in the Weiss and Aschkenasy study and my program.

    Most of my programs are written in BASIC, but unfortunately that means

Commodore BASIC, which is not directly transferable to other computers. I can

convert them into ASCII LISTings for anyone who wants to try rewriting them for

Microsoft or another BASIC. I've written a version of the DPD echo study in

PROMAL which can also be compiled on the IBM or Apple II if someone has the same

version of PROMAL.

I would be willing to help any other researchers convert my computer programs

into a form they can use. And I would appreciate any suggestions or criticisms

that would improve my computer programs.



                                       10