Before AAR UMLER

Index For This Page

This page was last updated on April 20, 2026.

(Sources include a wide variety of internet searches, including current and abandoned websites at Archive.org. The focus of this page is to use sources not previously readily available.)

AAR UMLER (Universal Machine Language Equipment Register)

American Association of Railroads (AAR) and UMLER (Universal Machine Language Equipment Register)

UMLER (Universal Machine Language Equipment Register) is the comprehensive, centralized electronic database for North American rail equipment, managed by Railinc for the Association of American Railroads (AAR). It serves as the primary registry for over two million pieces of equipment, tracking physical characteristics, maintenance, and ownership for railroads, shippers, and equipment owners.

(Read more about the start of UMLER)

Before UMLER

In the years leading up to the adoption of UMLER and related computer systems created by the larger railroads, such as COIN on Union Pacific, TOPS on Southern Pacific, and TCS on Missouri Pacific, one observer in 1966 noted that when a car left the a railroad's yard traveling to another yard on the same railroad, it was effectively leaving one "railroad" and going to another "railroad." Every process done at the first yard had to be repeated at the the destination yard as if the car had come from a foreign line. Every yard acted autonomously, and it was impossible to keep track of what car on a real-time basis.

TOPS (Total Operations Processing System) had been developed by SP, IBM and Standford University. UP took what it saw as the best parts of TOPS and created its own COIN (Complete Operating INformation) system. MP took its own best parts of SP's TOPS and created its own version called TCS (Transportation Control System) in 1968.

(Read the Wikipedia article about TOPS)

The existing communication system on each railroad was one in which contact with yard offices moved through the railroad's central office where the IBM or RCA or Univac mainframe computer was located. Each yard office waited patiently for an open line to transmit information that sometimes got garbled. The early computers began tracking cars using punch cards but at best showed only where a car had been twenty-four hours earlier. They provided history and statistics, but could not help make decisions on the spot or in advance. Empties got shuttled between yards and often arrived too late to meet demand. Yardmasters hoarded cars while steadfastly denying the process, and no one was in position to see how the local flow of cars affected the entire system.

The Railroads Act Alone

Finding pre-UMLER computer systems is like looking at a digital "Wild West" where every major railroad was racing to build its own proprietary language. Before the AAR standardized things in 1968, these internal systems were massive and often named with catchy, "space-age" acronyms.

The "Big Three" Pioneers

Southern Pacific: TOPS (Total Operations Processing System)
Developed in the early 1960s with IBM and Stanford, TOPS was the gold standard of the era. It was so advanced that Southern Pacific actually sold the software to other railroads (including British Rail). It allowed SP to move from "paper and telephone" logistics to real-time tracking.
Union Pacific: COIN (Complete Operating Information)
UP started with IBM 705 mainframes for accounting in 1958, but COIN was their first true operational system, developed in the mid-1960s using COBOL. A famous "quirk" of this system was that it couldn't handle leading zeros in car numbers, which forced UP to renumber its entire maintenance fleet to the 900,000 series in 1959.
Southern Railway: LDX (Long Distance Xerox) & Modern Data Center
The Southern was arguably the most aggressive adopter of tech in the South. They used an LDX network to transmit waybills electronically and built a centralized computer center in Atlanta. Their marketing slogan, "Southern Gives a Green Light to Innovation," was largely based on this early computerization.

The Eastern Giants

Pennsylvania Railroad: TRUCTRAC / Data Processing
The PRR was a heavy user of IBM punched-card equipment as early as the 1940s. By the 1960s, they had one of the largest data installations in the world in Philadelphia. They focused heavily on TrucTrain logistics (intermodal) and were founding members of the Trailer Train Company, which necessitated some of the earliest cross-railroad data sharing.
New York Central: Data Central
In 1963, the NYC opened Data Central, a $1.5 million computer-controlled communications switching center. It was the first of its kind in the industry, using Collins Radio computers to process teletype messages and track car movements across their "Water Level Route."
Norfolk & Western: Computerized Coal Tidewater System
N&W focused their early computing on the massive logistics of the coal piers at Lambert’s Point. They used IBM systems to coordinate thousands of hopper cars from the mines to specific ships, a "just-in-time" system long before the term existed.

The Western & Granger Roads

AT&SF (Santa Fe): PLUS (Programmed Logistics for User Service)
Santa Fe developed the PLUS system. Like TOPS, it was designed to give shippers real-time answers. They were also early adopters of the "microwave" communications towers that allowed these computers to talk across the desert.

Northern Pacific / Great Northern / CB&Q (The Hill Lines)
Before their 1970 merger into Burlington Northern, these three were already collaborating on data.

Great Northern was one of the first to use a UNIVAC for car accounting.
CB&Q (Burlington) had a sophisticated "Compass" system that tracked car locations.
Northern Pacific focused heavily on automating their yard at Pasco, Washington, using early computer-controlled humps.

Summary of Names of Computer Systems

Railroad	System Name	Focus Area
Southern Pacific	TOPS	Real-time Operations
Union Pacific	COIN	Inventory & Operations
Southern	LDX	Waybill Transmission
New York Central	Data Central	Message Switching
Santa Fe	PLUS	Logistics & Shippers
CB&Q	Compass	Car Accounting

The interested observer will notice a pattern: most of these railroads were using different "word lengths" and character sets (like EBCDIC on IBM vs. Fieldata on Univac). The year 1970 seems to capture when the industry realized that having 10 different "brains" was becoming a liability.

B&O (1970)

The April 26, 1970, article from The Baltimore Sun Magazine chronicles the Baltimore & Ohio Railroad’s transition from early mainframe computing to more modern, integrated electronic systems. It specifically focuses on the retirement of the B&O’s first computer, the "Datamatic 1000."

The article captures the situation on almost all of the nation's railroads at the time.

The key points of the article include:

The End of an Era: The 60-ton Datamatic 1000, which occupied 7,500 square feet, first installed in 1959, was dismantled and sent to a local salvage yard, with select parts preserved for the Smithsonian Institution and the B&O Transportation Museum.
Technological Advancement: By 1970, the C&O-B&O operated nine computers (including Univac, RCA, and IBM models). These newer machines used transistors and integrated circuitry, making them significantly smaller and faster than their predecessor.
Operational Integration: The computers handled a wide array of tasks, including employee payroll, stores department inventory control, billing, and real-time freight car tracing. The article highlights that major customers could link their own computers directly to the railroad's RCA 3301 to receive shipment data.
Labor and Automation: At first, employees were hesitant about the automation, fearing that would eliminate jobs. In fact, about 1,000 clerical workers were retrained for computer technology roles, though the company acknowledged that some positions were phased out through "attrition" (retirement or resignation).
The Struggle for Industry Standards: A major hurdle was the lack of coordination between different railroads. Because computers used different codes and symbols, they could not communicate directly.
Standardization Efforts: Among the first challenges making direct communication possible was tracking car movement within and outside of the railroad. The article describes the implementation of the "KarTrak" ACI system—the red, blue, and black color-coded plates placed on cars to be read by electronic scanners—and the start of the TRAIN (Tele-Rail Automated Information Network) to track car locations nationally.

The following excerpts come from The Baltimore Sun Magazine, dated April 26, 1970.

Parts of the B.& O.'s old No. D-1000 were making their last trip: down the freight elevator of the B.&O. Building, on dollies through the lobby and into a waiting truck on Charles street for consignment to a junk yard.

Nine computers now service the 11,000-mile network comprising the Chesapeake and Ohio-Baltimore and Ohio Railroads, the sprawling organization which resulted from affiliation of the two railroads in 1963.

Five of the computers are on the 10th and 11th floors of the C.& O.-B.& O. data headquarters at Baltimore and Charles streets, making it one of the largest and most modern computer centers in the city. It also is one of the most advanced computer operations in the railroad industry.

Other major railroads in Baltimore are serviced by computers outside the city—the Penn Central by computers in Philadelphia and the Western Maryland Railway by a center in Hagerstown.

Compared with the D-1000, one of the C.& O.-B.& O.'s three Univac computers weighs about 15 tons and occupies about 1,500 square feet of space.

The C.& O.-B.& O. handles 4,250,000 carloads of freight over an area stretching from the Atlantic Coast to Chicago and St. Louis in the Midwest. The company must keep track of its 135,000 freight cars; run, service and make efficient use of 2,200 locomotives; devise schedules; maintain equipment and inventories, as well as manage the mechanics that would go with any company large enough to bring in $924 million of revenue a year.

A vast staff of clerical personnel handled the resulting paperwork in pre-computer days. Later, the clerical workers provided many of the personnel to man the firm's computers. About 1,000 former clerical workers were trained by the company in computer technology over the past 10 years, and about 150 of them were moved to Baltimore from jobs in other cities during that time.

A C.& O.-B.& O. salesman or customer anywhere in the country can call the computer service and ask for information on a specific car and within a few seconds get a reply from the computer.

Such calls go to a group of seven employees, each seated in front of a telephone, a keyboard similar to that of a typewriter and a small television screen. Immediately after the identification number of the car is typed, about a dozen lines of number and symbols fill the screen and give a coded and abbreviated history of the car for the past 10 days—who owns the car, what's in it, where it's been, where it's going and the route it has taken.

About 200 of the C.& O.-B.& O.'s biggest customers are linked directly with the computer—an RCA 3301—and their computers receive information on their shipments each morning.

Another 3301 computer in the same room is a message switcher, receiving information from 750 stations, relaying messages to the proper stations and keeping copies for headquarters. Two other 3301's in Huntington, W. Va., also route messages for the company.

The three Univac computers in the B.& O. Building handle the company's payrolls, inventory control, billing and about a dozen other chores, including checking all car movements against a master schedule and reporting on those which are late.

Two smaller IBM computers keep track of cars and movements in yards in Chicago and Baltimore. The Baltimore terminal computer is in Camden Station.

The Western Maryland Railway's computer center, like the C.& O.-B.& O., also controls inventories, keeps track of cars and traces them and handles the billing or revenue accounting.

The Penn Central adds to these functions the handling of passenger trains and ticketing. Metroliner passengers in Baltimore can obtain reservations on the high-speed Washington-to-Baltimore trains and have them confirmed nearly instantaneously.

While computerization has spread throughout the railroad industry, the spread has not been coordinated. In billing each other, some railroads simply exchange rolls of computer tape on which the billing information is stored. But for the most part the computers on the nation's railroads cannot communicate with each other. The codes, symbols, and information contained in the memory banks of one computer system are not compatible with the data in another railroad's system.

One step toward standardizing computers, however, is now evident on about half of the nation's railroad cars. It will be on all railroad cars by the end of this year. This is a metal plate about 2 feet high and 1 foot wide on which 13 colored stripes contain the coded number of the car, identify the owner, describe the car, and provide a double check of the information to make sure it is correct.

The red, blue, black, and black-and-white dotted stripes were first tested 10 years ago [1960], but it wasn't until 1967 that the American Association of Railroads agreed on a standard code and the identification plaques began going on the cars.

Electronic scanners, similar to television cameras, can read the color codes and transmit the information to railroad offices and computers. About 70 of the scanners are now in use throughout the country in pilot testing programs. C&O-B&O expected to acquire its scanners in June and to begin immediately trying to make computerized use of the information the scanners pick up.

Railroad identification is now done teletyping information from terminal to terminal and hand-carrying lists of the cars on a given train.

Another cooperative venture—the Tel Rail Automated Information Network (TRAIN)—has been organized by the American Association of Railroads for exchanging information on railroad car locations, which will aid computer operations and uses a computer itself.

Beginning January 1, all railroads began notifying an AAR computer in Washington of each railroad car coming onto or leaving a railroad system. The association uses the information to alert railroads of potential car shortages in sections of the country.

The system is expected to be in full operation by the middle of this year.

Far from being a completely automated system, the computer operations rely on countless bits of information accurately translated into computer language by thousands of people.

The key word is accurately. A mistake of a digit or a transposed number results in a wrong connection, much the same as in a mis-dialed telephone number. The data flowing into the computers is about 97 per cent error-free.

Ten clerical employees must still untangle the remaining errors in tracing lost boxcars alone. An undetermined number are involved in sending out corrected bills and straightening out inventories, the railroad said.

The Need For UMLER

That Baltimore Sun newspaper article actually catches the railroad industry at a critical "missing link" moment in the development of UMLER (Universal Machine Language Equipment Register).

UMLER was born out of the exact problem described in the Baltimore Sun article: "the computers on the nation's railroads cannot communicate with each other... the codes, symbols, and information... are not compatible."

Here is how that specific era connects to the UMLER system and its predecessors:

1. The Predecessor: The "Official Railway Equipment Register" (ORER)

Before UMLER, the industry relied on the ORER, a massive printed book published quarterly.

The Problem: As the 1970 article notes, railroads like the B&O were handling millions of carloads. By the time a book was printed, car dimensions or capacities might have changed.

The Transition: In the mid-1960s, the Association of American Railroads (AAR) realized they needed a "Machine Language" version of this book so computers—like the RCA 3301 and Univac mentioned in your text—could talk to each other.

2. The Birth of UMLER (1968)

Just two years before the Sun article was published, the AAR launched UMLER.

Universal Language: Its goal was to create a standardized magnetic tape format (referenced on Page 8 as "exchange rolls of computer tape") so that a B&O computer could understand a Penn Central car's physical specs instantly.

The "Tape" Era: In 1970, UMLER wasn't a "live" cloud database like it is today. It was a physical master file on magnetic tape (like the ones Mrs. Claudia Collins is holding in the photo on Page 9). Railroads would mail or courier these tapes to the AAR to update the national register.

3. The "KarTrak" Connection (ACI)

The Sun article described the "red, blue, black, and black-and-white dotted stripes" (Page 8). This was the KarTrak ACI (Automatic Car Identification) system.

The Goal: KarTrak was intended to be the "eyes" for the UMLER "brain." The scanner would read the label, look up the car ID in the UMLER file, and instantly know if the car could fit through a specific tunnel or handle a certain weight.

The Failure: As the article hints (mentioning a 97 percent accuracy rate), KarTrak struggled. Dirt and snow on the labels caused too many misreads. While UMLER survived and evolved into the digital backbone of Railinc today, the KarTrak color-bar system was abandoned by 1977 in favor of the RFID tags (AEI) used now.

4. TRAIN (Tele-Rail Automated Information Network)

The article also mentions TRAIN. While UMLER tracked the characteristics of the car (length, weight, etc.), TRAIN was designed to track the location. Together, they formed the first real-time national "operating system" for North American rail.

UMLER Made Mandatory

The "forcing" of the independent rail giants to cooperate was a decade-long battle of economic survival vs. corporate ego. The AAR didn't just ask nicely; they used a combination of federal pressure, the threat of bankruptcy, and a newly created "technical police force."

Here is the breakdown of how the AAR broke the internal systems (known as "silos") of the Pennsy, UP, and SP:

1. The "Common Language" Crisis (1964)

The turning point came in 1964 when the AAR formed the Data Systems Division (DSD). Its sole purpose was to stop the "Tower of Babel" problem. At the time, if a Southern Pacific car went onto Union Pacific tracks, the UP computer literally could not "read" the SP's electronic waybill because they used different coding structures (like the EBCDIC vs. Fieldata issue).

The Stick: The AAR leveraged Interchange Rules. They basically told the railroads: If you want to get paid for your cars being on another railroad's tracks, you must report that car's data in a format our central computer can understand.

2. The 1968 UMLER Mandate

The AAR realized that the Official Railway Equipment Register (ORER)—the printed "Bible" of car specs—was too slow for the computer age. In 1968, they issued a mandate: UMLER (Universal Machine Language Equipment Register) was no longer optional.

The Enforcement: Every railroad was required to submit their fleet data on magnetic tape to the AAR's central "Data Center" in Washington D.C.

The Penalty: If a car was not registered in the UMLER master file, it was effectively "invisible." It couldn't be legally interchanged, meaning it couldn't leave its home tracks. This forced even the most stubborn roads (like the Pennsylvania RR) to comply.

3. The "TRAIN" System (1967–1970)

While UMLER was the "database," TRAIN (Tele-Rail Automated Information Network) was the "network."

Federal Pressure: The Interstate Commerce Commission (ICC) was breathing down the AAR's neck because of chronic freight car shortages. The government threatened to take over car distribution if the railroads didn't fix it themselves.

The Compromise: The AAR built TRAIN as a central clearinghouse. Railroads had to report every time a car crossed a "gateway" (like Chicago or St. Louis). For the first time, a railroad had to "snitch" on itself regarding where it was holding other companies' cars.

4. The KarTrak/ACI Gamble

The 1970 Baltimore Sun article you transcribed mentions the "red, blue, and black" stripes. This was the AAR's attempt to automate the enforcement of cooperation.

The Mandate: In 1967, the AAR made ACI (Automatic Car Identification) labels mandatory for all cars in interchange service.

The Result: This was the first time the industry agreed on a universal physical "barcode." Even though the technology eventually failed due to dirt and maintenance issues, the legal framework of having to label every car the same way was a massive psychological victory for industry-wide cooperation.

5. Why They Actually Gave In: The Bottom Line

Ultimately, it wasn't just AAR rules—it was interline settlement.

Railroads were losing millions of dollars because they couldn't track which "foreign" cars were on their lines, leading to massive errors in per diem (daily rental) payments.

The computers in the Sun article were finally powerful enough to prove to management that standardization of car-tracking data would directly result in more revenue due to faster interline billing.

The "smoking gun" of this AAR mandate was the "AAR Interchange Rules, Section L" from the late 1960s. That's where the legal requirements for UMLER registration and ACI labeling were codified into industry law.

###