Marhofn 196.11 - May 2009

Previous | Contents | Next

The Hills Database

Chris Crocker

Many readers will have come across the database of British hills, downloadable from www.biber.fsnet.co.uk since its launch in 2001. It is evidently popular, as it has topped Google searches on 'hills database' for some time. The downloads page gets about 20 hits per day, and had spawned 220 links at the last count. The brainchild of Graham Jackson, the database owes its origin to a task that must be faced by many walkers encountering The Relative Hills of Britain for the first time: how to calculate one's Marilyn total from ascent records for other lists. Graham's solution was to combine the lists in a database that could be queried to provide totals and other information. When I saw Graham's first attempt I was excited as it seemed he'd got a resource that would be useful to the bagging community at large. In those days one could obtain many individual lists from the web, but the only database was that on the University of Liverpool site, which was not downloadable.

I joined forces with Graham, and in September 2001 uploaded a Microsoft Access database, with embryonic notes. The RHB group was enthusiastic: there were suggestions for improvement, and help with data checking, so we had version 2 available within six weeks. We perhaps erred in not giving the database a specific name, but people were soon calling it the 'Hills Database' and we have stuck with that. In the early days we regarded the database as primarily a baggers' tool. We corrected obvious errors, such as incorrect grid references and map numbers in Munro's Tables, but were not concerned by minor errors in height or location. A frequent problem concerned hills belonging to multiple lists, where grid references or descriptions referred to slightly different locations, e.g. the Marilyn and Wainwright summits of Top o'Selside. Should the hill have one entry or two? We adopted a commonsense approach, but some early decisions have had to be overturned. In version 3 we decided the only way to overcome the confusion surrounding Beinn a'Chroin was to create separate records for the Munro, Murdo and Munro Top.

In 2005 we added 10-figure grid references from GPS units. This led to a change in objectives, since for the first time we were offering data more accurate than available elsewhere. Some other online sources were also publishing 10-figure grid references, but these were often derived from digital maps, which are not as accurate and often fail to identify the true summit. After some discussion we decided to embrace the concept of a database in a fuller sense and make accuracy a goal. This meant overhauling our existing data, since much published hill data is inaccurate (with the honourable exception of TACit Tables). It is a task still in progress. The most common error stems from grid references being rounded to the nearest 100 metres, as many list authors are unaware of the convention for co-ordinate systems (you should give the south-west corner of the square containing the feature, however many digits you quote). That problem is gradually receding as an increasing number of people contribute 10-figure grid references. Then we discovered that grid references from hand-held GPS units are subject to systematic errors arising from inaccuracy in the instrument's transformation from latitude and longitude. We have learned how to correct for them, so we now give co-ordinates to OSGB36 standard, in addition to measurements suitable for input to Garmin units. All this is fully explained in the database notes. That took some effort, but probably the most arduous task was overhauling the Wainwright Outlying Fells listing, in which we collaborated with Geoff Crowder.

In 2007, as a result of John Barnard's efforts, we added information from surveys. The majority are carried out by Abney level, and several of our contributors now use these instruments. Many surveys have caused a shift in location, often by a few tens of metres but sometimes by much more. John shares our views about provision of accurate hill data, so has become a co-author of the database and has created an Excel VBA version of it. By mid 2008 we were getting a lot of references in Wikipedia, and other websites were using our data, so we felt we should try to harmonise the various sources of hill data on the web to create a single source that the hillwalking community would accept as definitive. The survey data gives the Hills Database a unique advantage, but to gain wider acceptance it needed to be more comprehensive. A survey of users showed plenty of support for new lists and features, but the size of the task was daunting. We accepted some offers of help, e.g. to collate drop data, but to cope with database expansion we decided to increase the size of the core team. We have welcomed on board George Gradwell, a regular correspondent and expert on the Lake District hills, and Simon Edwardes, author of www.hill-bagging.co.uk. An early decision was to merge Simon's database with ours. Simon will take over the IT side from me, so it's possible the database will have a new URL by the time you read this. We are always ready to listen to suggestions, so please contact us if you have any views on how the database could be improved so as to gain wider acceptance - or indeed if you would like to help.

Previous | Contents | Next