Regarding my posting of a few month’s back, ‘The curse of metadata’, which was quite negative about the current state of web-related metadata initiatives in general, I thought perhaps I should provide a more positive counterweight so here goes.
For the past few years I’ve been working closely with geographic metadata, and one organisation that seems to be going in the right direction with this is OSGeo. They very much represent what is happening in the real world of graphical mapping on the Internet, and they are trying to improve the current status quo by promoting specifications and open-source software that is genuinely useful (and used) in this respect (e.g. Google Maps, GRASS and the GDAL/OGR libraries), as opposed to the rarified and labyrinthine world of the OGC, which seems to have little in common with what hackers are currently doing on the web with geographic data and maps (although OSGeo participants do recognise the importance of OGC and have many connections to the work being done by the OGC – the WMS specification, for all its faults, is one example of an OGC specification that fulfils a real requirement and has been widely taken up in the geographic software community).
The OSGeo approach is exemplified by the contents of the excellent O’Reilly books ‘Mapping Hacks‘, ‘Google ‘Maps Hacks‘ and ‘Web Mapping IIllustrated‘ (some of the authors of which are also participants in OSGeo).
OSGeo have produced the first of a series of online journals here. There is also an annual conference, FOSS4G, which concentrates on the aims of OSGeo.
Perhaps the most important thing that OSGeo is working on, from my point of view, is DCLite4G and the specifications for a geodata repository, including geodata metadata requirements. DCLite4G is a metadata standard that seeks, above all, to be one that can actually be used in the real world. It builds on Dublin Core, extending it so that it can be used to describe geographic information adequately, whilst retaining the clarity of Dublin Core. It also has the advantage that Dublin Core is already relatively widely-used on the web, so DCLite4G should be able to be integrated easily with existing data cataloguing applications. Compared to the ludicrously complicated behemoth that is the ISO 19115/19139 specification, DCLite4G is a much better bet for anyone who wants to predict where the future lies in this area.
There is currently no good open-source geodata repository application available anywhere, but the development work and research carried out by the GRADE project at Edinburgh University and the ongoing development of the open-source GeoNetwork application funded by the FAO, perhaps point the way (although in my view, GeoNetwork’s reliance on the OGC Catalogue Services and ISO 19115/19139 specifications are burdening it with unnecessary complexity).
My own view of OSGeo (and this may not accord with what they themselves might say, I’m not sure) is that the very fact that it exists at all shows up the inadequacies of the OGC and its failure to address the practical needs of the global geographic data community. OSGeo are stepping into the gap and providing a very useful service, and the fruits of their labours I’m sure will outlive and outgrow anything that the OGC produces.
Leave a Reply