Plato – The Birth of an Automated Research Tool

In the early 80’s, I was in Mensa and was trying to find some stimulating discussions of the outer limits of science.  I was an R&D manager for the Navy and was working for NRL in some very interesting but highly classified research.  I was careful to avoid any talk about my work but I really wanted to explore areas that I could talk about.  This was one of several attempts to do that.  I sent the message below to a young professor at Lawrence Livermore National Labs, who was running a Mensa discussion forum on ARPANET, in the hopes of getting something started.  He was working with artificial intelligence in math and robotic systems at the time.   

Remember, this was written in 1984.  The Apple Mac was one year old.  TCP/IP has just been introduced on ARPANET. Windows 1.0 was introduced in 1884 but I did not begin using it until version 3.1 came out.  The fastest processor was an Intel 286.  Most all software ran in DOS.  This message was originally sent via UUCP but I saved it as ascii text onto tapes and then later translated it to disks with the idea of someday writing a book, but I never did.   Enjoy….. 

Dennis,

This is my first contact with one of the Mensa discussion forums.    I found a few guys that were willing to talk to me but it seems I ticked off a lot of others by my lack of due respect for puzzles and my references to the “wing nuts and space cadets” that inhabit and comment on most of the Mensa forums.   🙂   I eventually formed my own forum, web site and discussion groups and a bunch of us proceeded to talk our way into a lot of business together. 

=====================================================================  September 9, 1984 

Hi.  I’m new to this board but I have an interest in the subjects you discuss.  I’d like to open a dialog with some of you about your ideas and what you are interested in and have analyzed or studied that may be interesting.  I’m no Mensa guru but I do like a mental challenge and the application of science but more importantly, I think there is a difference between knowledge and wisdom.  I seek the latter.   

Who am I: I guess what I need to do first is try to tell you who I am and perhaps try to establish a little credibility so that you won’t think I’m really am a National Enquirer writer or some wing nut with wild ideas.  Then I’ll present some simple but somewhat radical ideas to start with and see how it goes.  If there is any interest in going further, I’d love to get into some really heavy stuff about life, existence and the future.  I am particularly interested in discussing cosmology and the human animal, but that is for later.   

I’ve been developing a methodology for predicting events and narrowly defined aspects of certain subjects based on social and technical inputs from a vast information correlation program I use……But that should wait until I find out if anyone is even interested in this stuff. 

I have been working near the Washington DC area for a number of years.   I am a researcher that deals in analysis and logic.  I enjoy a mental challenge similar to what I perceive that many Mensa types like but I don’t enjoy the meaningless math puzzles or secret decoder ring stuff.  I prefer to ask or pursue the real mysteries of life and nature.   

I have a few technical degrees and have traveled and been schooled all over the world.  That was mostly a product of my parents being in the military and my early jobs.  I became interested in computers as soon as they came out.  I helped build an ALTAIR at the University of New Mexico’s engineering fair in 1971-2.  That was where the first “microcomputer” was created.  The Altair came a few months later.  It introduced me to computers but I immediately switched over to the software aspects of computers rather than become a hardware hacker.  I got an EE degree first, so I understand the hardware, I just think its secondary to getting the job done.  Then I got a CS degree and began to see the possibilities.  I did 40 credit hours of studies in computer simulations and loved it.  I was using math so much in my CS degree that I discovered that for one more semester, I could also get an BS in Applied Math – which I did.  Then I discovered that with just one more semester, I could get a degree in Physics so I did that too.  By then my parents were out of money and I had to get a job.  Ever since then I have been messing with computers. I was particularly fascinated by the speed of computers.  I won an award one time for being the only student that solved a particular math problem using an algorithm that would fit into 2K of RAM.  I did it simply by adding one to a variable and checking to see if that solved the equation ‑ if it didn’t I added one more.  It worked.  While working on one of the first OCR studies, I was captivated by the fact that the computer could find any text, no matter how much it had to search, in seconds that might take a person years to find.    That has been a driving force every since.   

 What is my Resource Tool? I liked software but I wanted to get to the things that I could see that a computer could do ‑ not spend my time writing code.  I became good at modifying and interfacing existing software to do what I wanted.  I found that this was much easier than writing my own code.  I got the original WordStar to talk to Visicalc and dBASE on an old CP/M Kaypro so that I could get automatic documents that self‑updated themselves.  That was fun but I wanted to apply the efforts more to real world applications.   

The programming was slow because I tend to think in pictures and I wanted the programming to think in pictures also.  I found a program that would reverse engineer a source code listing into a flow chart of the program.  It was crude but it worked.    I figured it would be even better if you could go the other way ‑ input a flowchart and get a compiler to write the code.     I bought a flow chart program and a Fortran compiler and made them talk to each other so that I could use the graphics of the flow chart program to create a chart of my program flow and then feed it into the compiler to get object code.   I have improved on it for the last several years so that I can input natural language variables and verbs and it interprets for me.  If it doesn’t understand some variable relationship and can’t figure it out by seeing it in context, it stops and asks me.  I now can spend most of my time Using a program instead of writing it.   

 CLICK! Necessity if the Mother of Innovation

The first real application of this program was when I became a player in the stock market and discovered it was easy to improve my investment decisions if I could get my hands on the right information.  The information was available, there was just no way to find it, link it and give it structure and purpose using the speed of the computer.  That was the start of my effort to create a better information search and retrieval system.   

 The Hardware + Software

In short, I created some special searching software that helps me find anything about anything and then it automatically identifies links, relationships and implications for me.  I know that sounds like a bunch of pie in the sky but it really isn’t all that hard to do.  There are, in fact several programs on the market now that do the same thing, only I did it first on a Radio Shack TRS‑80 in 1979.  Then again on an Apple II+ in 1983 and again in 1987 on a Mac and most recently on an MS‑DOS machine (from PC to XT to 286 and now a 386).   

My method has evolved over the years and now uses some fuzzy logic and thesaurus lookup techniques along with a smart indexing and interfacing to my CD‑ROM and hard disk databases.  I built it over several years in modular form as I added new hardware or new processing capabilities.  The flowchart compiler helped me move the code from one machine to another since the source code (the flow chart itself) remained essentially the same, only the compiler code changed.   I now have a mini‑LAN of four computers and it will pass tasks to other computers, in the form of macros, so I can get parallel searches going on several different information resources at the same time.    That also lets me proceed with the analysis while some slow peripheral, like the tape deck, is searching.   

 De Facto Credibility

This search software will also interface into script files for on‑line searches like IQUEST, DIALOG and about 270 others including several military and government databases and gateways (NTIS, DTIC, FEDLINK, etc.  ) that I have access to as a function of my job.  For the CompuServe Information System (CIS), the command structure that programs like TAPCIS uses makes it easy to initiate an on‑line search.  The slowest part of it is waiting for the on‑line responses from the dial‑up service that I am using but at work I can use some really fast lines on ARPANET. 

I also have access to a few foreign databases that are the equivalent of our UPI, AP and CIS’s IQUEST.  The European (based in Germany) databases have lots of technical data and the Japanese databases have collated worldwide news reports from about 30 nations.  I use some lines from Cable & Wireless that I am able to bill to my job.  The translation services allow me to issue a search in English and read the response in English but the data searched is in one of several languages.   I can get into a lot of this stuff for free but there is also a lot that costs money.  That’s one of the reasons I got permission and started using all these resources at work.    

 Plato is Born

Still, the on‑line search costs are why I tried to build up my own research capabilities.  I use a page‑feeder scanner and OCR software to read in books and other texts to add to the info databases that I can search.  There is a used bookstore near me that sells books dirt-cheap or will take trades for other stuff (non‑books).  This makes it possible for me to buy a book, rip it apart and feed it into the page‑feed scanner.  Then I can throw the book away.  Since I never, ever let anyone else use the database and never quote directly out of the texts, its not a copyright violation.   

400 CD‑ROMs, 90 (14 inch) laser disks, 250 or so tapes and perhaps 5000 disks of compressed (zipped) text files gives me immediate access to my own database of about 500 gigabytes of text or about 500 million pages.  Some of this has line pictures but most of it is just pure text because the OCR software does not translate the images – just the text.  That is a loss but if I think the image is important, I scan it and save it on a disk.  Add to this on‑line access to about 3500 databases, including some I can get to at work, containing perhaps 50,000 times as much as I have, and you get some idea of how powerful my search capability can be.  I call my search program, “Plato”.   

 Concept Searches: With Plato, I am able to input a search “concept” instead of a search “syntax”.  It will automatically cross‑reference and expand the search into related subjects based on parameters I set.  It took a long time to learn hour to phrase my search syntax but I usually get back just the data I want.  Plato saves the search paths, resources and references in a bibliography format if I need to refer to the source.   

When you think about it, it is all pretty simple and commonly used techniques used in lots of commercially available software.  The search of compressed (zipped) text data is done real well by Lotus Magellan.   Lots of search software is available but I found a mix of GOPHER and FOLIO VIEWS with some added fuzzy logic and thesaurus lookup techniques that I enhanced after seeing some spell checkers that looked up words phonetically and with missing letters.  The interfacing was simply a matter of finding hooks in other programs or putting front‑ends on them to get them to talk to each other.  If all else fails, I just use a script file and/or keyboard macro in a BAT or BIN file to simulate the manual typing in and reading out of text.  That always works.  Linking Information Resources: 

There are lots of programs that can search one database or a selected set of data sources.  All I did was add a few extra features (script and macro files) to make it move from one reference to another, to quantify the validity of the data and added some interfacing software that I wrote to make other programs, that already do parts of this, work together.   Using some of the research techniques and capabilities that Plato allows, I have been able to identify some very interesting linkages and cross‑references to concepts that may be of interest to people in this forum.  I have also been able to fairly easily dismiss some of the quackery and screwballs that sometimes frequent these idea exchanges.   

 And Then What?

I am a serious and scientific researcher and I am not interested in some of the nuts and liars that grab scientific or technical words at random and make up their own versions of reality.  On the other hand, I consider the majority of science to be somewhat boring.  I may not KNOW everything but I don’t need to if I can find out what I need to know in only a few minutes on the computer.  Besides, even if the answer to any question is right there on the screen, I still have to read it and after awhile, that mounts up to a lot of reading.   

It’s like having a dictionary.  Anytime you wanted to know what a word means you’d look it up, but most people wouldn’t sit around all day looking up words just for fun.  Now imagine the same thing with a very good set of encyclopedias.    There would be a lot more information but after awhile, just knowing that you can find it would be enough.  Now imagine a set of encyclopedias that contains 87 billion, 500 million pages of text!  That’s how big my dictionary is.  Ok so its not really that big but we are talking about the size of hundreds of libraries. The one advantage that I think I have over many people is that I believe that the answer to most of our questions are out there somewhere.  Many people don’t even think to ask if they really believe that the answer is not available.  Let me give you an example.  I worked as a part-time consultant to government contractors for a while and I often dealt with clients that were preparing a proposal for a contract.  When I tell them that I can get detailed information about what their other competitors are doing, most think I can’t or it would be done by illegal means.  I can and it’s legal.  I can, in many cases get not only what the competitors are going to bid but their cost structures and their past performance.  I can even get the salaries of the people doing the bidding.  After awhile, my clients start asking me to get information that it would never occur to them to ask for before I came on the scene.   

 Monotony: Getting back to that incredible large dictionary, it might be fun to look up stuff for a while but pretty soon you would stop looking up random subjects and try to find some real challenges.  I got to that point about 4 years ago, shortly after I finished the prototype for my first PC based search software.  I have expanded its capabilities as new databases became available.  The addition of the scanner to read in hardcopy text was a big improvement.  I was able to select books in topic areas I wanted or to fill in gaps in coverage.    The scanner(s) has been going, on average, at about 2‑4 hours a day for the last several years.   

 The Hawking Incident

As I added new data, it was fun for a few days to search for some incredibly minuscule detail.  Or to try out a fuzzy search and chase down some concept.  I particularly liked writing to Steven Hawking and telling him I thought I had determined the size of the universe.  He was very polite when he said, “I know!”.   

That incident was one of many where I began following a trail of information that made me believe I had “uncovered” some new idea or concept that “I” had not heard before only to find out that upon deeper research, it has already been discovered.  With all this information, it is a very humbling thought to realize that someone out there knows at least some part of all of it.   I guess there is something to be said for being able to consolidate and cross‑reference all of this information and focus it down for a single person.  It has the net effect of allowing me to ask questions that lead me into areas that I would never have known to follow into.   

It is very useful to integrate across scientific study areas.  For instance, medical people seem to know very little about electronics or physics and vise versa.  The result is that scientists in each field limit their view of the world by only seeing it from their own field of study.  Only in the last few years has there begun to be a cross mixing.  Things like a tiny pill made of SMD (surface‑mount devices) that a patient swallows.  The pill has a sensor array and a transmitter that sends data to a receiver outside the body.    The term non‑invasive gets redefined.   It seemed like ages before they began to introduce virtual reality to medical systems and robotics and yet it seemed to me to be a perfectly natural mix.   I felt that as soon as a movie like TRON was made, that it would be only a matter of time before robotics, animation and computer graphics would be combined into a 3‑D viewer but it seems that it is just now catching on.   

But What Has this all got to do with you?

Now it is at this point that I must chose a topic to discuss with people of this forum.  I enjoy almost any intellectual discussion from religion to cosmology to the human potential but I prefer a topic that is perhaps a little further out than most of these and that mixes a lot of hard-core science and math with some logic and speculation.   

I am very curious about the fringes of science.  The areas where conventional science is afraid or unwilling to conduct real research but that has an unusual following of “believers”.    _____________________________________________________________ So, Dennis, what do you think? 

___________________________________________________ 2007 Update: 

In the late 1990’s, I updated Plato with a modern windows GUI interface and OOPs OCX files and modules.  I expanded into a Dbase DBMS engine and SQL interfaces.  I was able to multiplex multiple modems using some ISP software so I could use multiple lines of input.  Later, I extended this to multiple computers on a TCP/IP network using broadband.  It still relied on macros and keyboard simulators to interface with other commercial and proprietary software but its parallel operations equated to massive procession power. I have continued to make use of a lot of web sites and online services that I can access as a result of my government work and that gives me a huge increase over simple web searches.  I also have improved the bi-directional translation capability so I can tap into databases created in other countries. 

I have also since expanded its ability to search for themes, concepts and related ideas while improving its ability to quantify the relevance of those findings.  It still takes hours to resolve most of my searches but I let it work overnight and sometimes over the weekend to find and resolve my searches.   The end result is a very useful tool that I find helpful but, as noted above, it is not perfect and still falls far short of the human mind. 

Leave a Reply