Marhofn 269.15 - May 2013

Previous | Contents | Next

The Hills Database version 13

Chris Crocker

After the effort of producing version 12, we had hoped for an easier life in 2012. No chance. Shortly after tidying up with v12.1, a new mapping resource appeared on the Geograph website. Like the vector map on the OS OpenSpace website, the largest scale shows many additional spot heights, particularly at cols. However, Geograph shows many heights that do not appear on the vector map - the converse is rare - or on 1:10k paper maps, and - very usefully - contours. The facility to rapidly compare mapping at different scales and read off grid references makes the new resource very easy to use. So Geograph offered the possibility of improving the data for many hills, particularly drop. We were already aware that some of our col data was probably suboptimal because the various third-party sources we had used frequently disagree.

How to proceed? Mark Jackson took a sample of 650 hills in region 34 and found 15% needed their data changing. A below average proportion, as it turned out; some regions approach 50% when you include revised col GRs. Jim Bloomer summarised the team's feelings: 'My judgement is that this (Geograph data) is the biggest issue to come along since I joined the team seven months ago'.

We bit the bullet and started a systematic review of all 7880 British hills in the database, setting a target of completion by April 2013. This was not limited to Geograph but included reference to old 1:10560 and 1:2500 maps for hills where data was missing or uncertain, using the resources offered by the National Library of Scotland and We gratefully accepted an offer of help from Dave Marshall but the rest of the work has been done by ourselves. As I write, we are on schedule to release version 13.1.

Inevitably, some tampering has resulted. The Humps have been the most affected with, as of January 2013, ten promotions, nine deletions, four replacements and five relocations. In contrast, the TACit lists and the Nuttalls have emerged almost unscathed. The reason is not hard to find: these authors consulted 1:10k maps extensively, whereas other lists are mostly based on data from smaller scales.

A casualty of diverting effort to the data review was delay to version 13. It was never our intention to let the database become so out of date. We intend to move to more frequent releases to reflect the current pace of change. Important changes will continue to be applied on the site as they occur, but the online and downloadable formats should stay better aligned in future.

Does the completion of the data review mean the team can now take a breather? I doubt it, as the OS are currently revising their 1:25k maps of Scotland. Some of the contouring has changed, and 5m spacing is more frequent. Spot heights may change too, as we believe the OS wants to make data consistent across scales.

It is unclear where next we will take the database. Some users have begged us not to add more lists. Others are asking for Tumps, Yeamans, Synges, island highpoints and so on. We may also have reached the stage where a list's inclusion has an effect on its popularity. How else do you explain the hill-bagging logs of some Subdeweys, as the list is not published elsewhere? We will consult the user base as always, but keeping everyone happy is becoming more difficult.

Previous | Contents | Next