LWN: Comments on "The JMRI Project and software patents"

The JMRI Project and software patents

t.brack — Sat, 09 Dec 2006 00:25:44 +0000

I have reviewed the thread, and it seems that a lot of people have missed the mark. The patent appears to on the application of Queuing Theory (specifically priority queuing) to commands sent to and received from a DCC command station, and not on the actual control of a model train through software.

A quick review of the patent reveals it to be (IMHO) nothing more a legaleze encoding of the application of queuing theory to the NMRA DCC specification which existed well before the patent was apparently applied for. Queuing Theory/Priority Queuing is definately prior art and goes back to the early days of computing. There is also nothing beyond the name that is unique to a Digital Command Controller.

Beyond the terminology and the encapsulation of queuing theory on top of the (public) NMRA DCC algorithms, there does not appear to be anything new or unique being patented. I guess if you sling enough legalese jargon around something as simple as this algorithm, you're bound to confuse lay people enough to get anything past a Patent investigator.

Copyright and Plagiarism are quite another story. Here we are embarking around a situation where something that is morally and ethically wrong may in fact escape the definition of "illegality" through technicalities of the law. At very least one would expect that the source of the copyrighted sources being used by a commercial package would be acknowledged.

I am aware that in 1976, at the Univerity of Waterloo a course in queuing theory existed that utilized model railroad equipment, but of course it would not have had a DCC Command Station.

Thanks,
Tony

The JMRI Project and software patents

coolgames — Thu, 12 Oct 2006 02:49:46 +0000

I am just recently made aware of this discussion but I will be trying to research my past effort to automate a HO train layout at DIGITAL Educational Services in Bedford , Mass at Crosbey Drive from November 1980 to October 1990.
I was the technical instructor for DPM - which was Distributed Plant Management which later became DEC Dataway. We were developing the curiculum to train on the RSX-llM+ for use in automated process control. We downloaded via a multicast the RSX-11S image that would have standalone instructions to control what it was DIGITALLY connected to such a automated factory floor. The host PDP-11/70 would tell the remote systems what to do over DECNET.

The HO layout was built for a fellow instructor who was previously a teacher in public schools and had an interest in model trains.

I will talk to my supervisor who works locally and may have been responsible for that layout's design and build. He may also know how it may have been used and where it ended up,

Need more explanation of "what we're looking for"

Fossils — Tue, 30 May 2006 12:48:24 +0000

Your comment about Software developers is spot on. Even basic programming course teach these techniques. Furthermore does KAM Industries honestly want us or the patent office to believe that he was the first in this regard, that no one else has used logic gates and/or multiplexing in electronic circuits controlled by computers and software.
Whats the definition of a "Digital Command Station" certainly not a term that can be specifically applied to a DCC controller in a train set.

What is the definition of "digitally controlled" in relation to a model railroad? what is the first command? an instruction issued by software or an electronic circuit be powered up and intialised.

The JMRI Project and software patents

Fossils — Tue, 30 May 2006 12:27:11 +0000

Of the claimed 7,000 copies of this software downloaded very few would actually be in use. Like me, they would have all good intentions but it will never happen. Always end up with to many other things to do and to many distractions.

This software could be used for any form of control or automation, the fact that it can be used for model trains is a bonus. The fact that it has been given a slant towards train control only serves to identify another use for it.

As i have said previously, done and done, way before KAM Industries put a Patent on what we had been doing for years beforehand.

My interests now lean more towards home automation and expanding on my car PC.

The JMRI Project and software patents

Fossils — Tue, 30 May 2006 12:08:08 +0000

Left out a bit, the electronics magazines included IC based interfaces and assembler code to interface to the various modules in the book and the train set.

The kit computers were used mainly for the trainsets and radio teletype decoders, both with code written in basic or machine code. Cheers Fossils

The JMRI Project and software patents

Fossils — Tue, 30 May 2006 11:52:18 +0000

Look at most existing electric trainset layouts or trackplan books and it is a natural and logical progression for these to be represented and controlled by a computer.

I am looking at the 1984/85 lima Railways catalogue:quote
Lima have developed a new multi-point control panel for fast automatic train routing, ... it goes on ... The control panel has interchangeable sections for you to map the layout of your tracks, and by using a special electronic pointer, trains can be directed automatically along chosen sections. End quote.

1984 - Electronics for model Trains by Ken Stone, Page 68 using computers on model railways - dedicated computer, circuit based on Z80 processor

1990 - Electronics for model Trains Book 2 by Ken Stone, Page 32 computing and model railways - upgrade to computer including addition digital outputs.

For me it was 1980 and my first Microbee,an Australian designed and built computer(Z80& later Z80/8088/68000),and writing a program in basic to control, switch and display a train layout. Microbee's were being used to control everything from hospital equipment to train sets. A detailed magazine was published monthly with hardware and software for the experimenters,including model trains. Specific hardware was also released which including the "microbee smart model controller", and there was also an "experimenters board" with detailed projects included for model trains. Microbee also had software for release which I saw operating on a private layout owned by one of the microbee engineers in Gosford. This layout was controlled by 2 Microbee computers, one managing over 2,000 switching operations with a colourful representation of the layout and switching operations, and the other an 11 track distribution "table" feeding the shunting yards by aligned itself with the main yard line to dispatch 1 of the 11 additional trains as it was called for. Prior to the production runs of Microbee computers there were a number of others sold in kit form and most based around the Z80 CPU.
Cheers - Fossils

What about full-size trains controllers as prior art

stuartward — Wed, 17 May 2006 11:34:36 +0000

There is a fully automated "Train" that runs between the North and South Terminals at Gatwick Airport. The north terminal was opened in 1988, so presumably the train was designed well before that.

From http://www.gatwick-airport-uk.info/gatwickairporthistory.htm
Work began on a second terminal in 1983. Her Majesty The Queen and The Duke of Edinburgh inaugurated the £200 million Gatwick Airport North Terminal in 1988.

and from http://en.wikipedia.org/wiki/London_Gatwick_Airport
Gatwick Airport Transit
The Gatwick Airport Transit provides free transportation between the North and South Terminals. The transit system uses transit vehicles that run along a 1.2km long elevated two-way track system. The transit vehicles are automatic driverless people movers each with three cars.

The JMRI Project and software patents

smpierce — Tue, 16 May 2006 16:52:47 +0000

On January 29, 1941 the Museum of Science and Industry in Chicago opened an 'automated model train' exhibit. Here's a picture of the designer/builder sitting in front of the controls which could be considered as an early form of a computer.

http://www.msichicago.org/exhibit/great_train_story/histo...

The JMRI Project and software patents

jlefevre — Sat, 06 May 2006 14:09:44 +0000

It was the early and mid 1980's. There was a series of articles in the MODEL RAILROADER
magazine. The system was called DIGITRACK. It allowed the independent direction and speed
control of up to 16 locomotives by sending digital commands over the rails to decoders in the
engines. The article was a complete how to build all of the parts. I and another guy I worked with
both built the complete system. As an electronics engineer myself I had fooled around in the late
1970's with using phase locked loop systems when those IC's first came out. I am using O gauge
so the noise issues were always a problem for me. I am using JMRI now and TMCC. I looked at
comercial software but the KAM stuff is WINDOZE CRAP and I am a Mac user as is Bob J and many
of the other major people in the jmri system. The KAM lawsuit is a sham. Since JMRI doesn't
compete with KAM for macintosh users (since their crap doesn't run on Mac's) then maybee that
is a way to put another angle on them. The fact it just _ happens_ to run on windozes since it is
platform non-specific java code is an accident. And what about the linus, unix, os-9, and other
users running jmri that KAM also is not available for. How can you claim to be monitarily
damaged when you do not even make a product that will work for different types of computers
than your software supports. Thats like Apple suing Dell because Dell installs a digital photo
orginizer software suite on its windoze boxes and Apple sells iPhoto as part of a software
package that won't run on windoze anyway.

forgot to mention the year : old in 1984

arcticwolf — Thu, 04 May 2006 15:47:57 +0000

I also recall reading about a system that allowed you to use your computer to control your model railroad in a Märklin catalogue - this must've been in the early 90s. I'm not a model railroad fan myself anymore, but if somebody else still has those catalogues or can find them on eBay, it might be worth looking into.

The JMRI Project and software patents

Wol — Thu, 04 May 2006 14:44:21 +0000

It's not about what the other person gained, it's about what you lost.

Let's say A is selling his software at $10 a CD.

B copies it and gives it away.

A is entitled to go after B for the number of copies he gave away, times $10. The fact that B made nothing is irrelevant - A lost a load of sales (and hence lost a load of money), for which he has a right to be compensated.

(The thing about this case, however, is it appears it was B's software in the first place, and A was threatening to sue, so B has sued first.)

Cheers,
Wol

The JMRI Project and software patents

philips — Thu, 04 May 2006 11:12:29 +0000

The case defies logic.

Guy was giving the software freely to anyone - w/o any kind of profit.

So what damages can the plaintif claim?????

Talk disruptive innovations...

Using a model railroad to teach digital process control

grouch — Thu, 27 Apr 2006 16:43:01 +0000

Using a model railroad to teach digital process control by John W. McCormick, State Univ. of New York, Plattsburgh. Published 1988, citations back to 1979. Full PDF available for download.

The JMRI Project and software patents

NRArnot — Thu, 27 Apr 2006 11:07:11 +0000

I recall reading many years ago that a lot of pioneering work was done in the early days of computers (early 60s?) at MIT, in particular in an organisation called the TMRC (Tech Model Railroad Club).

It still exists, see http://tmrc.mit.edu

Could be a very valuable place to start asking questions!

forgot to mention the year : old in 1984

macc — Thu, 27 Apr 2006 11:07:06 +0000

i.e. it was already shoved in the corner then because
the institute had moved on to air traffic control

In Germnay starting in the 80ties you could buy a lot of
modular digital control stuff for modelrailroads from
Märklin and Fleischmann and probably some others

Magazines like Elektor or ELV probably had project
in that direction as well. I will have to look this up.

One computer-controlled train example:

NRArnot — Thu, 27 Apr 2006 11:02:33 +0000

The London Underground's Victoria Line, opened in 1968, was designed to run under automated control. Officially it was never run without a "driver". Unofficially there has been at least one instance when the driver (or ratherm, manual supervisor) got out of the cab without disabling the automation, and the train took off without him, no harm done.

http://en.wikipedia.org/wiki/Victoria_Line

What about full-size trains controllers as prior art

macc — Thu, 27 Apr 2006 10:56:49 +0000

the "Institut für Flugführung" at TU-Braunscheig, Germany

had a large Modelrailway to demonstrate general concepts
of ( distributed ) traffic control.

The JMRI Project and software patents

biolo — Wed, 26 Apr 2006 14:08:12 +0000

I don't know if this qualifies, but I remember as a kid in the 1980s having a controller which talked to little circuit boards onboard the trains themselves, allowing multiple trains to be independently controlled on the same track. I can't remember the name of that product, but I do remember it was a cheaper alternative to the Hornby Zero 1 system (Wikipedia entry). That system was replaced by a digital one, Digital Command Control in the 1990s, which according to Wikipedia was an open standard!
I seem to remember that you could have multiple controllers connected to the same piece of track, in fact the controllers could be bolted together for just this application, to my mind taking this even closer to the patent.

One computer-controlled train example:

eru — Wed, 26 Apr 2006 11:34:57 +0000

Apologies for following up myself, but I just now remembered one very well-known (at least in Europe) case: The Docklands Light Railway in London.

Accoring to the info on this page: http://www.tfl.gov.uk/dlr/about/infopack.shtml, (follow the DOWNLOAD FREE INFORMATION PACK (PDF) link), construction started in 1984 and it was opened to the public in 1987. It was computer-controlled right from the start.

What about full-size trains controllers as prior art

eru — Wed, 26 Apr 2006 11:26:04 +0000

So the patent talks about model trains.

In the normal universe (as opposed to the looking-glass world of patents), it is obvious to anyone that controlling model trains with a computer is not much different in principle than controlling real trains (if anything, model trains should be much simpler).

So when was the first fully-automated rail system (including subways, trams, airport terminal trains, automated carts running on factory floors, whatever) designed? I am pretty sure I have heard of such things being done before 1986.

The JMRI Project and software patents

ttonino — Wed, 26 Apr 2006 08:31:50 +0000

My Java instructor - while a bad teacher - explained to us that Java descends from set top box and military control applications. The latter is not so far from a model railroad. Depending on what is claimed, this may make the invention trivial (like using a screw for what it is designed to do).

I also remember publications by Sun (don't remember the name of the technology) touting some Java feature to contrl independent entities - getting an ambulance to the place of a car crash is an example I remember. I do not know about the date of this one.

Why not threaten them with cold war?

copsewood — Wed, 26 Apr 2006 08:03:48 +0000

If you look closely at the retaliatory action that has been taken, this seems to point out the dangers inherent in trying to make claims for licenses to use patents which are of questionable merit - especially if false claims were made about lack of prior art or lack of knowlege of such when the patent was obtained.

Why not threaten them with cold war?

MarkVandenBorre — Tue, 25 Apr 2006 21:57:13 +0000

A relatively small firm with actual product sales is very vulnerable. Why not make cold war with them? Surely someone somewhere must have a pile of braindead dusty patents ready to nuke them with if they keep on harassing!

Fast, clean and cheap. And guaranteed to gain you respect and publicity in the free software community. What are you waiting for, patent holders?

Need more explanation of "what we're looking for"

dwheeler — Tue, 25 Apr 2006 17:39:03 +0000

I agree, we need more explanation of what we're looking for. The right to create blog entry on this refers to the KAM letter. I looked at the KAM letter, and the letter focuses on this: "Claim 1 of U.S. Patent No. 6,530, 329 claims a method of operating a digitally controlled model railroad comprising the steps of: (a) transmitting a first command from a first program to an interface; (b) transmitting a second command from a second program to said interface; and (c) sending third and fourth commands from said interface representative of said first and second commands, respectively, to a digital command station."

I would love to see a more detailed posting on WHAT prior art, exactly, they're looking for, so we could all help more directly. But the above may be a start. Identifying many previous model train controllers might help to at least some extent, since one of those may very well have the other properties they're looking for.

The notion that you can patent multiplexing when you apply them to model railroads is ludicrous, and points out the problem in the patent system: It fails to work rationally with software. It's like a painter being told he can't paint, because someone patented the idea of painting big buildings. Software developers go to school to learn a large set of well-known techniques and then apply the techniques to each problem they face. Multiplexing is one of those well-known techniques. The idea of applying a well-known technique to a well-understood problem is immediately obvious to any practitioner of software development, and thus the patent should never have been granted in the first place. Frankly, the fact that JMRI was first should be enough, right?

The JMRI Project and software patents

iabervon — Tue, 25 Apr 2006 16:29:11 +0000

Could we have PJ's explanation of the patent, so we know exactly what we're looking for? One common problem with asking for prior art from experts in the field is that reading patents is somewhat tricky, and it's easy to find previous inventions which fit criteria which are non-normative or incomplete. So there's a patent on a nail clipper, for example, and people say, "I've been clipping my nails with a nail clipper for 30 years!" But the patent is actually on a way to make nail clippers with one fewer part than anyone's done before, but people trying to read it tend to quit reading while the patent is still describing things everybody knows about nail clippers to provide the necessary context to be able to state what's different, and so they think they find a lot of prior art, because they're only reading parts that the invention shares with the inventions that it improves upon. So it would be helpful to have a plain language description saying what would be necessary to invalidate this patent and what would just be considered a different invention.