All-Purpose Problem-Solving Machines
The computer that we are so familiar with today is the product of many brilliant minds and a long expanse of time. The Antikythera mechanism, an ancient Greek orrery retrieved from a shipwreck in 1901, is the oldest example of an analogue computer. As early as 1623, Wilhelm Schickard constructed a mechanical adding device. Just over 200 years later in 1833, Charles Babbage designed his Analytical Engine, an all-purpose problem-solving machine with a mechanical memory. George Boole discovered binary numbering and Boolean logic in the 1850s and in 1939 the Atanasoff-Berry Computer (ABC) was acknowledged as the world’s first general purpose electronic analogue computer.
Lloyd’s Register (LR) was an early adopter of computer technology and from the installation of their first computer in 1962, looked at ways that this remarkable technology could be used across the organisation. One way was to streamline the work of updating and producing the Lloyd’s Register of Ships. This publication, also known as the Register Book, has been published by LR since 1764, an extraordinary achievement. This book has been referred to in the maritime world as ‘the Shipping Bible’ as it helps to keep subscribers and users up to date on the ships they charter, insure or buy. The Lloyd’s Register of Ships was computerised from the 1975-76 edition, the first amongst the classification societies registers.
The Register Book
The earliest Register Book was unique and provided the template for all its successors. It contained the details of 4,118 ships, 2,000 of which were built outside the UK. Information included the past and present names of the vessel, the name of the master, the number of crew, the names of the owners, the number of guns carried, when and where the ship was built, and the classification given. The Register covered the years 1764-66 and was annotated in manuscript with updates as the information was received at the coffee house via letter. By the 1768-1771 edition, the Register was being updated weekly. From 1775 ‘posters’ hand-stamped amendments to entries in the Register using a wooden-handled metal block, a method that continued into the mid-20th century. Those subscribers that either did not or could not return their Registers to LR’s printing house, received a Supplement by post.
In addition, special editions of the Register Book were printed for the use of the London staff to enable them to keep track of all the changes made. The system was in operation for 150 years from 1834 until 1984 after which annotations were made by computer
Lloyd’s Register and the Computer
The computer proved one of the most useful tools of the 20th century for LR. It cut calculation times on mathematical problems for increasingly large and sophisticated ships from days to hours, or even minutes. The opportunity to introduce new regulations, based on mathematical analysis rather than experience was revolutionary. The computer decreased workloads and enabled plan and design appraisals for ships, industrial projects and offshore structures to be carried out much more quickly. It also made administrative processes across the organisation more efficient.
In June 1962, following some nine months of discussion and visits to manufacturers, LR installed its first computer: an IBM 1620 with a memory capacity of 120kB.The second-generation solid-state machine was specifically designed for scientific and engineering calculations and research and development.
I told [John] Murray and [Horatio Nelson] Pemberton that the Society had to convert itself from the quill pen era in which it was, where we used to do everything by hand
Bryan Hildrew
A True Pioneer
Joyce Cork joined the Society from school in 1950 and studied for her mathematics degree at night school. She later studied for a second degree in metallurgy. In 1962 she was asked to set up LR’s first computer department with Frank Brock, installing and operating the new system. A field in which she continued to work until her retirement in 1988. She had been one of the first members of the Engineering Research Department, which made extensive use of LR’s new computer facilities. Joyce wrote many of the programs for LR’s early computers, replacing the calculations that she had previously done by hand. Joyce’s mathematical talent, combined with her practical approach to problem solving, helped her to make significant early contributions in areas like fatigue in shafting and the low temperature characteristics of materials for containing liquefied petroleum gas and liquefied natural gas. She was also part of the pioneering team involved in developing the stress analysis and pressure testing of nuclear pressure vessels in the 1950s. She was instrumental in persuading many of LR’s more traditionally minded staff to renounce their slide-rules and adopt the use of calculators and would later go on to setting up desk-top computers for use in plan approval work.
The Unique Identifier
In 1963 LR decided to use its new computer facilities to help produce the statistics on the world fleet that it had published since 1878 and to also help with ship research analysis undertaken within the Technical Records Office. Les Heminway was a statistician within that office and conceived the LR number (now also known as the IMO number). This unique identifier was assigned to every merchant ship recorded in the Register of Ships, regardless of class or flag and would remain with the ship throughout its life, irrespective of conversions and name changes, and never re-applied to another ship. It was originally a six-digit number. The base numbers used for the initial coding exercise for the LR number were those in the supplement to the Register of Ships for 1963-64, a year which had the Register commencing with 00001 A.A. COWAN, and ending with 39966 ZYRIANIN in the main Register and 4xxxx for subsequent entries in the supplement. This base number was prefixed with ‘5’ to make it a six-figure number, with ships ranging from 500001 to 54xxxx. All ships appearing in the 1963-64 edition of the Register were given a unique number, thus older ships were in the sequence 500001-539966, and new ships, together with changes of name, in the sequence, starting 54xxxx.
The 1969-70 edition of the Register of Ships was the first to show the check digit, making the LR number the seven-digit number known today.
Towards a Total Computer System
At sea the computer also made its appearance in the challenging marine environment. In 1969 LR classed the 210,000 tons deadweight tanker Sea Sovereign, the first ship to go to sea with an on-line computer handling machinery controls, cargo loading and navigation.
Rapidly programmes were developed and made available both in-house and to customers of LR. The IBM 1620 having proved itself, the need for greater capacity soon developed and an IBM 360/30 took its place in December 1966. This mainframe computer, designed by Dr Gene Amdahl, was first introduced in the mid-1960s. The System/360 was a family of machinery with upward compatibility throughout its range, so it was relatively cheap to upgrade. Continued development soon saw an IBM 360/40 at work in 1970. LR’s computer revolution continued with the commissioning of an IBM 370/155 in 1972, and computing facilities were made available in its offices in Yokohama, Gothenburg, Rotterdam and Madrid. In the following two years computing work grew by more than 20 times so the facilities were enhanced to the IBM 370/158 and linked with the PDP-11/45 used by Offshore Services Group.
With this new generation of computer came the ‘Nastran’ suite of programs originally developed for the American aerospace industry. This took LR closer to a total computer system covering all aspects from design, through survey and monitoring of individual ship records, to the production of the Register Book.
Computerisation of the Register Book and Statistical Publications
The decision to computerise the updating and production of the Register Book was taken in 1974 an initiative led by incoming chairman, Robert Huskisson. The opportunity was also taken to bring the Register Book back into step with the statistical publications, which since 1964 had been produced from Statistical Department’s computerized General Ship File. Computerisation helped make the increasing numbers of ships and data fields added to the Register more manageable and reduced printing time from nine to six months, now starting in March each year ready for publication in July. The first computerised edition, which appeared in 1975-76, was even more comprehensive, updated by monthly cumulative supplements.
The Register Book “take-on” as it became known, was a huge job by any standards, and an army of temporary staff was brought in to re-enter the data on the 70,000 ship entries that existed at that time. As there was no room for them in the Fenchurch Street complex, many were located at a building at Stone House Court, Bishopsgate. Many of these temporary staff would stay with LR after their temporary jobs concluded, some within the Register Book department and others in the Computer Department.
The information from the Register Book was entered onto forms, with some 94 forms needed for the different types of information contained within the Register for each ship. The forms were sent to operators who transferred the data to electronic form, before the data was transferred onto tapes using Datapoint 2200 machines. The data was batch processed and run overnight a few times a week on LR’s mainframe computer, thereby updating the master Register Book file.
The flag officers within the Register Book department put in many hours of overtime to carry out most of the proof-reading, fitting it in around their daily work of updating and verifying the Register Book information. The process took about 18 months, the edition for 1975-76 being the first time the Register Book appeared in a new, re-designed, computerised form. This was swiftly followed by the computerisation of the List of Shipowners and then the Casualty Returns.