Condensed from a paper by Ian
Buxton, Department of Marine Technology,
At the first Ship Databases Workshop held at
1. Creating a register of ship databases.
2. Developing a Global Ship Numbering system (GSN) to provide "unique identifiers" to distinguish different ships of the same name.
3. Facilitating creation of global ship databases.
A website was created documenting several dozen existing ship databases ranging from Lloyds Register’s Seabase to shipwrecks in Scottish waters.
A Working Group was set up at the first Workshop to review how a Global Ship Number system could be developed.
For modern ships, the Lloyds Register number printed in their Register Book (which is also the International Maritime Organisation/IMO number) serves such a purpose, remaining with the ship throughout its life. This seven digit number is unique, only ever being allocated to one ship. LR started allocating numbers to ships in the mid 1960s, now amounting to some 150,000 vessels of which about 90,000 still exist. However there have been several hundred thousand ships which never had an LR ID, essentially all vessels which did not exist after the mid 1960s, most naval vessels, vessels under 100 gross tons and non-seagoing vessels.
Ideally every significant ship should have a unique ID, which can then be used both as an identifier to show which of the 800 vessels which have been named ANN is being referred to, and to link different ship databases by means of a common field. Thus a database on say ships built in a region should be able to relate to another database recording vessels wrecked. The first database does not need to hold information about shipwrecks, nor does the second need to hold information on the ship construction. As long as both store the Global Ship Number (GSN) as well as the name of the ship in question, they need only a pointer to each other (or a manual input of GSN) to find further information. This means that the creator of any ship database should include an 8-digit numerical field to store the GSN, in addition to any record numbering system the database uses.
The key requirements are that the GSN should be unique, that
it should be capable of application to all ships, past and present worldwide,
merchant and naval, and that it should be easily created by authorised people
without risk of duplication. (Such conditions rule out such suggestions as
using the British Official Number). It is not necessary that the number contain
any ‘intelligence’ or information about the ship itself, such as year of build
(which in any case is not always known for old ships); that is the function of
the fields in specific ship databases. The Group concluded that the GSN should
be an 8-digit integer, as 6 is not enough to cover all possible ships, while 7
could be confused with the LR ID. The GSN needs to be associated with just
enough ship information to show that the right
A GSN Index is analogous to a telephone directory: to find
the number of Mr Smith of
Given that several hundred thousand GSNs will be required, the most practical way of producing them is to derive them from existing ship databases. Creating the core information for all these ships from scratch would be a huge effort requiring large resources which are simply not available. However a number of large database owners have agreed to allow data from their records to be used to used to populate such a GSN Index. For example, the British Shipbuilding Database includes 70,000 British built ships which were built before LR IDs were assigned. Such a database can be used to ‘populate’ the Index entries by exporting the relevant fields. The GSN can then be formed from the BSD Record Number/ID with one or two additional leading digits added to make up the eight.
The proposed numbering system is (with recent revisions supplied by the author):
1xxxxxxx Reserved for future use.
2xxxxxxx Ships derived from the British Shipbuilding Database, i.e. British built merchant and warships generally post-1850.
3xxxxxxx Ships derived from Marbase, generally
ships over 1000 tons grt or deadweight, worldwide.
4xxxxxxx Ships derived from Lloyds Register database, i.e. merchant ships over 100 GRT existing in mid 1960s when LR numbers started, or later.
5xxxxxxx Available for other Class Societies and number assigning authorities who have computerised databases, e.g. with modern (non-oceangoing) vessels like inland waterways craft, fishing vessels and larger leisure craft.
6xxxxxxx All other merchant ships without an LR
ID or on BSD or Marbase etc. In practice, where records already exist, these
might be moderate size databases like the Canadian Mills List or Paddy Watson's
tankers. These ten million numbers could be allocated in smaller blocks to
database owners willing and able to export core data, with the second/third
digit indicating the actual source, e.g. 61xxxxxx, 62xxxxxx.
7xxxxxxx Since it appears that little
structured work has been done on pre-20th century wooden sailing ships, it
would be useful to have a completely separate block and encourage database
creators and populators. (Due to the quirks of 19th century Lloyds Registers,
there is a major problem of avoiding duplicates, as ex-names were not always
given).
8xxxxxxx Naval or offshore vessels not already on any of the above, both current and historical. In practice, the largest number of ships are likely to be modest sized vessels from major navies like U S Navy. It appears that little has been done in the form of comprehensive databases for warships other than those currently in service, e.g. LR or Janes.
9xxxxxxx Used for dubious entries or potential
duplicates, to be checked as further information comes to light.
The Editors would assign number blocks to contributors, develop mechanisms for checking of duplicate entries, and convert the table fields exported by contributors into the format required for the GSN Index. Duplicate entries need to be eliminated, not just because the same ship may be on two of the populating databases, but because a ship launched under one name may be the same as that completed under a different name (common with speculation ships).
The Working Group agreed to develop such a pilot GSN Index, for demonstration at the second Workshop, populating it with a sufficiently large selection of ships so that users could try it and feedback their responses.
Three database originators in the Group agreed to provide test data on some 6000 ships with some 12000 different names, i.e. including renamings: Marbase (Erik Bergman), Lloyds Register (Chris Dean) and British Shipbuilding Database (Ian Buxton). Sufficient fields were exported to enable ships of the same name to be identified from: Type; Tonnage; Year of Build; Place of Build.
A single file of ships was created for the pilot scheme with a query mechanism that allowed users to search by ship name to find the GSN, or by GSN to find all the (other) names that the ship held.
Making available a list of unique GSNs is an essential first step towards any global ship databases. A comprehensive Index covering all of the world’s quarter of a million significant ships will allow a user to determine whether a particular ship existed, as determined by its Name compared with its Type/Tonnage/Year and Place of Build, and find its GSN. If all ship database creators then included the GSN for each vessel they record, further information may be found on the required ship.
Thus progress that needs to be made includes:
1. Finalise the form of the GSN Index
2. Expand the GSN Index in stages to cover all significant ships.
3. GSN Index then becomes the ‘first port of call’ to determine if a particular ship existed and to find its GSN
4. Creators of existing databases gradually add the relevant GSN to a new field in their own databases
5. Users search all likely databases using the unique GSN to find further information on the ship of interest
6. Ultimately, link such databases by GSN to allow searching of each, e.g. through the Register of Ship Databases and the Internet.
In this way a network of global databases can be created, with coverage of ships built, owned, operated, chartered, engined, crewed, war lost, wrecked, etc. This seems more practical than trying to produce one single mega-database derived from dozens of different ones, each with different formats and degrees of accuracy in their information. But although Items 2 to 6 are now technically possible, we should not underestimate the large amount of effort and cooperation and time and money needed to turn such goals into reality.
To return to ship identifiers paper click here To return to main Maritime index click here