Ordnance Survey Central Government conference @ Leeds

I was fortunate enough to attend the Ordnance Survey’s Central Government conference in Leeds yesterday. I’d not attended one before. The event was pretty well attended by groups representing central gov’t from around the north and beyond. Two presentations, in particular, drew my attention.

Can you freestyle?

One was from John Carpenter who gave a great presentation on Product Development of OS data product over the next period. It’s pretty interesting how the OS product set has evolved over time, especially given the Phoenix Project ‘stuff’ that’s been going on over the last few years. For the OS to reach ‘product nirvana’ where all products are derived from a single capture-base will be a triumph, and John admitted that they are not there yet. However, it’s an ambitious target given the wealth of historic information and the immense legacy the OS have got, so it can’t be any small task.

The item of interest there surrounded the ability to render information from a single code base, at a variety of scales, but still to be relevant and appropriately accurate, whilst not ‘confuddling’ any data that is to be placed on top of it. As ever, this is the issue: historic paper maps aren’t designed to have anything more than coloured pins inserted on them. Digital maps are designed to be manipulated to have a wealth of information placed on top of them.

How do they do that? Well, VectorMap is a good stab at it. I was interested in the OS’s take on it that they didn’t want to supply style files to consumers in order that they use it: they want the user community to come up with styling themselves, especially if they want to place it below their own business data. In my experience, the business community likes to have styling done for them. In larger organisations, they like the idea that it conforms to a common look-and-feel, and will no doubt ask the OS to provide their own – they won’t want to change it. I suggest that, as much as it would probably pain them, the OS provide a set of different styles for the layer (somewhat like Tom Tom and their mix of rendering styles, I guess).

Web services and Government

Tim Martin gave an excellent presentation on web services. I was interested, though, in the audience reaction to how it’s possible to consume OS data from On Demand or via OpenSpace. The reaction – well, I’d describe it as pensive. I was trying to work out why though – the idea of consuming data from a variety of map sources is great and isn’t really new now. ESRI promoted the GeographyNetwork something like ten years ago – On Demand is clearly a large development from that, but the idea is the same.

Maybe a lot of us are still in the desktop safety zone, with data in close proximity to us. That’s highly understandable, both for political and for ‘warm fuzzy’ reasons as well. Tim’s presentation provided the audience the belief that the total cost of ownership in the new web services, cloud-based world is inevitably cheaper. I’m inclined to very much agree. However, getting this change into a number of public sector organisations is a large shift in culture. It also comes at a time when, unfortunately, public sector budgets are under threat. Although it makes sense to move to centralising data and to consume it direct (rather than hosting yourself, where possible), the cost of change, I think, may be the off-putter.

There’s also the threat that web services themselves take away some of the role of the GIS department in an organisation. Being able to download chunks of data, as opposed to applying for it to your GIS department does indeed knock someone out of a role. However, I’d argue that that person is then freed up to do more analysis and business work – stuff that’s of better value to the business anyway. And potentially less arduous.

So, in summary, it was a good occasion and I’m glad I went to it – I came away with a number of ideas and thoughts as to how this will go, though it may take time to make the move to a service world within central government. We shall see.