"http://images.thetimes.co.uk/images/trans.gif" \* MERGEFORMATINET A 
database for old Times' sake


INCLUDEPICTURE "http://images.thetimes.co.uk/images/trans.gif" \* 
MERGEFORMATINET
THE full research potential of The Times is being unlocked, or perhaps 
unleashed. Complete pages and all the articles from our first 200 years 
will shortly be available online in a form that makes possible 
astonishingly sophisticated searches ó although to begin with only in 
subscribing libraries.
Gale, part of Thomson Learning, is scanning all the issues published 
between 1785 and 1985 ó nearly a million pages and some ten million 
articles ó to create The Times Digital Archive. To that is added 
technology that can read the original and sometimes erratic typesetting 
and locate words or phrases across the centuries. As well as editorial 
matter, all display and lineage advertisements and even the crosswords 
are being scanned.
In just a few seconds from entering a search term, the user is shown 
statistics of the number of occurrences, by category: advertising, 
editorial and commentary, news, picture gallery, business, features and 
people.
Beneath this are chronologically ordered thumbnail images of each 
relevant page, with a red border around the article in which the search 
term occurs. The border shows where the article is on the page, and this 
and the headline indicate its likely nature. A click will then produce 
either a larger image of the entire page or an onscreen ìcuttingî, 
reproducing the article in clear facsimile, which can then be printed.
The text of the entire two centuries ó a run of papers that fills a 
40-yard corridor with shelves three deep on both sides in the Wapping 
basement ó should be available by September. At present the years 
1880-1985 are online, so that The Timesís coverage of most of the 20th 
century can be trawled.
Search, for instance, for the word ìcomputerî, and the program finds 
early examples from when it used to mean a person who does calculations. 
The Times first recorded its use to denote a machine in November 1946 in 
a report of an impressively perceptive speech by Lord Mountbatten, in 
which he referred to the ìelectronic numeral integrator and computer 
(Eniac), employing 18,000 valvesî.
Describing the coming ìrevolution of the mindî, Mountbatten imagined a 
machine that ìcould even be made to play a rather mediocre game of 
chess!î And the report went on: ìIn the field of memory alone, it seemed 
likely that Man was to be provided with vastly greater and speedier 
access to the inherited knowledge of the ages than he was able to 
command at present.î The two terabytes of images and text in this new 
database certainly fulfil that promise.
But although the automated reader successfully identifies words printed 
in italics or hyphenated across lines, it is far from infallible. It can 
throw up a lot of gubbins, as a search for a non-existent word shows. 
For instance, ìboybî is said to appear more than a hundred times in the 
available years. The program is misreading words such as ìboysî, mostly 
in the small type sizes of classified advertising. Inevitably, then, the 
success rate with other search terms will be less than 100 per cent. 
Furthermore, misprints in the original are not utterly unknown.
Nevertheless, the technology is awesome, and the results are generally 
better (and clearer) than the equivalent searchable version of  The New 
York Times. The program can also do fuzzy searches, identifying words 
similar to a keyword ó revealing, for instance, that in the 1920s The 
Times referred occasionally to ìHerr Adolf Hittlerî.
Practically the only thing the program will not do ó and this is a 
commercial, not a technical restriction ó is convert the articles it 
locates to plain text that can be cut, pasted or e-mailed.


For possible interest, excuse the occasional lapses!

I wonder if map dealers advertised?

iain T.
Halifax, NS

The Times Digital Archive is the largest project so far undertaken by 
Gale, a company that specialises in commercial provision of large 
historical archives. Its next is to digitise the entire text of 
virtually every 18th-century book printed in Britain ó 150,000 titles.
Initially, The Times Digital Archive is being marketed to libraries 
(some 800 of which hold Galeís microfilm edition of The Times, from 
which the electronic version is largely being prepared). Online 
subscriptions start at £2,100. One must hope, however, that prices will 
fall in future as the initial costs are amortised, or that pay-per-view 
facilities will be made available. The sheer power and potential are too 
great to keep locked up.
One of the first subscribers was the Oxford English Dictionary, 
delighted to be able to search for earlier occurrences of words than 
those in its records. But every kind of history will be facilitated, 
including genealogy, biography, social history and the history of ideas.
Not only is the search facility unprecedentedly comprehensive, but it 
will also produce negative evidence, showing that something or someone 
was mentioned much more rarely than one would suppose, or not at all. It 
is easy to ascertain, to take a couple of literary instances, that 
Joyceís Finnegans Wake and Larkinís The Less Deceived were not reviewed 
on publication.
Such phenomenal tools will inevitably change the nature of research, as 
Google already has. There will be fewer opportunities to discover 
sources overlooked by everyone else, but all the more need to be 
imaginative in winnowing, making intelligent connections and deciding 
what evidence really matters, and what it signifies.
Meanwhile, for Times journalists there is a chastening sense of 
contributing day by day to a vast and enduring encyclopaedia.

By Jim McCue

THE TIMES May 09, 2003