Well,
Geomatics Atlantic 2010 has wrapped up and I must admit that it was a lot of fun. It was good to get a chance to get out and talk with all of the people who use our software, take notes, and talk about the past and future. There were a lot of great paper presentations, and I was honoured to be selected as one of those presenters. I really enjoyed presenting and I'd love to do it again sometime in the future.
During the conference, the
OSGeo Atlantic Chapter had it's first meeting where we selected our first representative and co-chairs. Following the meeting I got a list of
CARIS employees who would be interested in joining the chapter and I got quite the response. It's good to see CARIS employees taking the initiative to support the free and open source software community for geospatial, that is
OSGeo. We use a fair amount of open source geospatial software in our products and we're proud of it.
Metadata?!Lately I've heard another resurgence in discussions about metadata in regards to spatial data. So I figured I'd do a quick summary of my thoughts on the subject without getting into any ISO details that seem to confuse the overall idea.
If you have had any experience in web page creation, you must have stumbled upon
Meta data profiles. These are often used by
Web Crawlers as they move about the web, gathering information for
search engines to index the World Wide Web. Web Crawlers gathering data is the first step in allowing your search engine to know what exists on the web and ranking the content. Currently the geospatial community doesn't use bots for this type of discovery because, well, it's tough to figure out what stuff is. There are a lot of proprietary spatial formats and they don't lend themselves for discoverability. That is why there are now XML files that help better describe spatial datasets, and they're put into
discovery portals so that people have a one stop shop to find what they're looking for. Now that pretty much puts the Geospatial Community into the mid-1990s where there were more search engines than you can shake a stick at. I personally don't care much for this, but whatever... that's where we're at until everyone finally just uses Opengeospatial Web Services that exchange spatial data in a more descriptive way. (Many thanks to the
OGC for trying to make this happen!) This will leads to a more
Semantic Web where people can find spatial data easily on the web using existing search engines. Because let's face it, everyone uses the web to find stuff so let's have it work with the current infrastructure. If we were driving around downtown looking for data, we'd attach big flashing lights and loud horns to whatever we wanted people to find.
It just makes sense. But enough about discoverability...
The other big case for metadata is to accurately describe your data. Sure enough, those same scary ISO specifications exist with standardized ways to represent what something means. This helps software more than it helps users, since everything is standardized and
should help with system interoperability. If you think we're describing data for discoverability, you're wrong. I talked about that earlier. Pay attention.
I'm talking about the fact that a lot of these various spatial data formats cannot accurately describe their content well. This is a big deal for many organizations since they want to correctly and clearly describe their data for various reasons, some of which are legal. So in this case, if I'm to distribute a spatial dataset to some other party, I want to make sure that metadata goes with it. Whether the metadata be inside of the exchange file or along with it, it doesn't matter. My legal responsibility is fulfilled and I did my best to describe what was provided. In my opinion, this is the first and most important goal of metadata. Again, with Opengeospatial Web Services, this isn't a problem. The OGC has taken steps to make this more and more of a non-issue.
When data is shared with suitable metadata to complete it, you can accurately describe your map. A map in this case is whatever you're seeing in your favourite GIS product after loading well described data. If you have data loaded from various web services, at a glance you can quickly get information on your map. This is a simple and effective use of metadata. To an end-user, they don't even have to know what metadata is. They'll see an "info" link on their map page and see information about their map. It's just that simple. But first, we need to accurately describe our data.
So don't get too worked up about metadata. Definitely don't get too concerned with discoverability. It's happening and I bet it's not going to happen the way we currently expect it. Instead, focus on correctly describing your spatial exchange formats. The rest will fall into place.
My Opera
