jwg

For various reasons lately I've thought about my career. So here is a quick summary post.

After graduating from MIT (Math) in 1960, I was a Physics PhD candidate at Columbia. That first year I also had a part-time job teaching 7th and 8th grade science. At the end of the year I decided I didn't want to be a mathematician, a physicist, or a teacher. You had to worry about the draft so I thought I try one more year at grad school while figuring out what to do. My father expressed one of his one-sentence brillant suggestions: "you should take a computer programming course". Another one was when I was wondering about where to apply to college he said: "You are going to MIT".

A few weeks into that course it became clear what I wanted to do. I moved back to Cambridge took a few classes at MIT as a Special Student while contemplating admission and had a part-time computer job. And in the spring I applied for jobs. You could get a draft deferment working for a critical industry which inluded computer companies.

My first interview was at Honeywell - to work on a Fortran Compiler project. I got a verbal offer at the interview and on the way home and stopped at a bank for a loan application so I could buy a new car. A few days later I got the written offer. I accepted it negotiated the start date and then went the dealer to purchase a brand new MG-A.

I wrote the code generator for that compiler and then became project leader for another compiler project. When that was done I joined the Systems Planning department - which was frustrating because they weren't very receptive to new ideas. They sent me and a colleague to an MIT symposium about Multics - a new operating system project that was a joint MIT, Bell Labs, and GE (the hardware vendor) project.

Several months later in early 1967 I (and my colleague) were new employees at GE. The GE office was in Tech Square - a remote office to the division in Phoenix. It was called the Cambridge Information Systems Laboratory (CISL).

I worked at CISL as a programmer, then development manager and eventually CISL manager. Bell labs dropped out early in the process (the Multics people on the project became the people who invented Unix). GE sold its computer business to Honeywell and Multics became a serious product. -multiuser with emphasis on security. It ran on an expensive mainframe (5-10 million dollars, a huge machine room - and not as powerful as my phone). We had very prestigious customers: Ford, General Motors, the Pentagon, NSA , the US Geologic Survey, Several Canadian government agencies (we were strong on dual language support - a Candian requirement), and a bunch of French agencies and universities (Bull, the French computer company had been affiliated with GE and then Honeywell for years).

In 1985 it was decided to "cap" Multics and CISLwas shutdown in 1986. A bunch of my team moved to Honeywell offices in Billerica to work in a new project (with some of the other people there) which we called Opus using Multics technogy to be a new OS on one of the Honeywell small computers; I was the manager. The computer industry was changing and Bull bought controlling interest in the Honeywell computer business and Opus was cancelled. Some of my team stayed, there was a reorgination and layoff and I inherited a couple of other departments.

A new highup VP decided he wanted to form a high-level group called Corporate Software Technology to work on advancing a bunch of things so as to improve productivity and product quality. I was recruited by him and became Mission Manager of Software Renovation - this was looking at tools and technology to help enhance existing products. It was interesting, and frustrating - I worked with people in Billerica, Phoenix, Italy, and France - and scored a trip to Japan since NEC was a partner.

One of my colleagues in Corporate Software Technology was named the manager of an Applied Research group and I joined it. Our main project was working with a research group at University of Illinois and another one at Columbia to build a prototype software Inspection system called Scrutiny. We demoed it a various conferences, wrote a paper with several of my peers to present at the European Software Engineering Conference in 1993. We got a DARPA grant to do more work on this and there were thoughts about making it a product. But with many changes in the industry and upper Bull management decide to scrap our work. In 1994, another layoff came up and I volunteered.

I formed a one-man consulting company - JWG Software Systems. Bull was one of my clients and worked with my old group a bit. Then I branched into working on a number of patent infringement law suits (defense and offense). At my own choice this tapered off until I stopped doing it (it was a nice transition - worked 4 days a week, 3 days, 2, 1, and then 0).

The last Multics system was shutdown in 1992. There is still an active Multics cult; people who worked on Multics are called Multicians. We have a Multicians web site. It took a while, but some of us eventually got Bull to release the source code (and system tapes) and they are stored at MIT. Last year at the 50th Anniversary of MIT's Project MAC there was a Multics session and I chaired a panel session.

Several years ago a couple of people started writing an emulator of the hardware and the system now runs. There is lots of discussion on the email list and it is fun to hear from some old-time colleagues. Some more work is needed and there are various thoughts about a possible future.

The Supreme Court just issued a unanimous decision disallowing the patenting of human genes. It's about time; the idea of patenting something discovered in nature is so ridiculous.

For a few years in the recent past I was a consultant on a number of patent suits - both defense and offense. The patents were mostly computer software (my area of expertise) and a few hardware. In the course of this work I read over 1,000 patents and studied 100-200 of them in detail. In my opinion most of these patents should never have been issued because they weren't original ideas or they were obvious.

One of the main requirements of patentability is that the invention being patented is not obvious, meaning that a "person having ordinary skill in the art" would not know how to solve the problem at which the invention is directed by using exactly the same mechanism. That is pretty hard to check.

My understanding is that the patent examiners have very little time to review each application. And believe me it takes a lot of time because you have to relate the claims (the statements of what is being patented) to the often poorly-written design description which is needed to clarify what the claims actually mean. And then there is reference documentation and cited patents that are similar (but cited to show that this invention is something new) which should be checked and finally there is knowledge about the "state-of-the art" which is needed to judge whether the invention is actually new.

And then if a patent is issued and the holder sues for infringement a lot of sleuthing has to be done to defend the suit. I remember one patent that I worked on whose holder sued two companies for infringement. I thought the idea was clever but not really original and the patent shouldn't have been issued. The patent holder hadn't succeeded in making his own product.

Both cases were in Discovery Mode and thus I was given the source code and technical documentation of both products - there was a lot of it; if printed each wold have been ~1,000 pages or more.

It was interesting to note that one of the products was written in C and it was very messy code because the product had to be able to run on Macs, Windows, a DEC OS, a Novell OS and Unix and thus had to deal with machine and OS differences when compiled. How they ever got that mess to work is a good question. The other product was mostly written in Java which made it and was much easier to understand and I'd suspect was a more reliable product; it's design documentation was much clearer too.

I wrote some tools to examine the source code to look for potentially infringing code and think I found all the right stuff. I wrote up my findings, and went over it multiple times with the attorney. One case was going to go to trial and I didn't want to participate since it would have meant traveling to Chicago several times and hanging around with little to do; so I declined to do more work. My understanding was that neither case actually went to court (this is what usually happens), but there was some form of settlement. Both companies that were suing were bought by larger companies and I don't know what happened to the products.

If that patent hadn't been issued I and the attorneys wouldn't have earned any money; other than that nothing else in the world would be different. Patenting things that shouldn't be patented and then dealing with the subsequent law suits is a big industry that keep lots of people employed. There are people/companies who buy patents and then sue for infringement - they are called Patent Trolls. It is estimated that patent trolling in the US resulted in 29 billion dollars of costs in 2011.

While reading The Statues that Walked* by Terry Hunt and Carl Lipo I saw this quote at the beginning of the conclusion chapter.

Finding the occasional straw of truth awash in a great ocean of confusion and bamboozle requires vigilance, dedication, and courage. But if we don’t practice these tough habits of thought, we cannot hope to solve the truly serious problems that face us - and we risk becoming a nation of suckers, a world of suckers, up for grabs by the next charlatan who saunters along.

Carl Sagan - The Demon Haunted World: Science as a Candle in the Dark, 1955

*a new book about Easter Island that offers a very different explanation of what happened using good scientific methods.

No, not me collapsing. I'm reading a fantastically interesting book: Collapse - How Societies Choose to Fail or Succeed by Jared Diamond. Diamond is a professor of Geography at UCLA. (He won the Pulitzer for another book: Guns, Germs and Steel which I'll read as some point)>.

In this book he writes about a number of societies that encounter a host of problems - some of which result in failure. I was drawn to it because of his analysis of Easter Island. Thus far I've read about Montana (not failed), Easter Island, Pitcairn and Henderson Islands, the Anasazi, and the Mayans - more failures to come as well as a lot of political-econmic analysis.

He takes a very scientific approach using information from many other scientists and includes his own analysis as well. He clearly expresses his opinion but allows for the situation that there are many opinions and some contradictory data so he might be wrong. He does an excellent job explaining the science involved. An example of this is where he discusses Dendrochronology by which a climate whose rainfall varies a lot over large time spans (not true for all areas) allows scientists to make a pretty valid estimate as to when a tree used in a structure was cut down as a means to date a structure.

I loved this sequence of sentences near the beginning.

"So that readers will have some advance idea where they are heading, here is how this book is organized. Its plan resembles a boa constrictor that has swallowed two very large sheep."

He talks about five factors:
• damage that people inadvertently inflict on their environment,
• climate change,
• hostile neighbors,
• decreased support by friendly neighbors,
• how the society responds to problems.
He uses this framework to discuss each place that he analyzes.

In the Montana chapter he talks about the effects of mining, deforestation, lack of rainfall and distance from markets that set the stage for a large set of problems. On the other hand the fact that it is so beautiful there has attracted a large number of wealthy people to buy land and built fancy homes (but not be residents who pay state taxes). This causes land prices to go way up so farming becomes difficult - note that here is where the estate tax comes in because when the children inherit a family farm that barely has enough income to keep alive they suddenly have to pay a large tax and have to sell the farmland for development. This community is still healthy primarily because of financial support from outside - government and the out-of-state land owners who also get houses built and use services. But the potential conflict between old-timers vs new-comers could destabilize the place. (Full disclosure: This is an overly simplistic description of his analysis.)

For Easter Island his analysis is that its collapse was due basically because the people overexploited its own resources. At the time there were no friendly neighbors or enemies and it looks like the climate didn't change so he believes those factors can be ruled out. Essentially deforestation made fishing impossible because they couldn't build canoes, the population was too large for local farming on bad soil (deforestation made it worse) to get enough food and as a result there was so much strife between the rival tribes. Essentially a massive civil war resulted in the killing of most of the people. Of course a bit later there was the hostile neighbors problem where the Peruvians kidnapped about 1/2 the population to be slave labor and later when some returned they brought smallpox.

Of course you can see the parallel to some of our near future dilemma. We are messing up our environment, the climate is changing and may well be more hostile, each country has friendly neighbors and hostile enemies. Currently we are not responding well to many of these problems. The earth is not going to be saved by friendly neighbors from nearby planets and not many people can migrate to them. So we are going to have to figure out what to do, or...

I like to drink root beer out of bottles with a straw but most straws are too short. Last night at Four Burgers in Cambridge I got two straws, took out from my pouch my trusty nail clipper (which had come apart and had to be repaired), trimmed a slot at the end of one straw so that it could be crimped a bit and inserted in the other straw to create a double length straw. It worked although the flow was a bit constricted; it didn't leak.

As my brilliant idea for the day maybe I should patent it - though I looked and found 12,960 patents with the word straw in them - I glanced at the titles of the first 150 and they were almost all about drinking straws. This is the first public disclosure - I have a year to file :).