Congrats to the Google Earth team on the launch of GE5. The ocean feature is especially nice, though I’m not sure why anyone would be surprised. I remember stories from last spring (I reflected on this then, so I won’t repeat) about Google collecting the data, and this is the logical conclusion of that work. The water rendering is very clean, and frankly, the only thing it’s missing is sound effects and a dry towel. Mars is also very welcome.

Only thing I’m frankly not too excited about is the gx: namespace for new KML features. Why? Well, I (and others at MS) have been trying to encourage my friends on the various Virtual Earth teams (2D, 3D, 20D…) to more fully embrace and support KML, open standard that it is. Speaking only for myself, I think this makes it harder for MS and other companies to do so, not that I’m speculating on official future support either.

To be clear, it’s legitimate for any company to add custom extensions to its own implementations of KML and for its own users when it feels it’s the best or only option. and namespaces are the best way to do that. Also to be clear, I believe MS voted to support a recent Google-backed rule change at the OGC that would make these sorts of optional incompatibilities more fast-tracked to standardization, retro-facto. This is pretty much what I guessed we’d see as a result, though maybe not so readily. I would have hoped for a more collaborative approach to advancing the spec.

Anyway, it’s not earth-shattering, so to speak. In fact, the new work I’m doing has very little to do with any of this (I don’t ever want to do the same thing twice, in any event). So I’m quite happy to see Google continue to push the envelope in delivering an increasingly perfect mirror of the natural world, and an impressive structure for hanging all of your important geospatial data.

Congrats again to the whole GE team.


 [edited middle section for clarity]

4 thoughts on “Earth5

  1. The standards process for KML actually requires complete proof-of-concept implementations to have been tested in the wild before proposals for new KML tags will be considered. To be standards-compliant, this testing phase has to be conducted using an extension namespace. Google is just following the rules of the KML standards process. This seems like a much better situation to me than if they were to set a bad example by introducing new tags willy-nilly into the main KML namespace.

  2. Namespaces serve a valuable purpose. And I think I was clear about the process being legitimate. However, you make it sound as if Google is the unwilling subject of rules it significantly helped create, and that there is no possible way to advance the spec in a collaborative way.

    Arguing that this is much better than seeing Google pollute the global namespace is kind of silly, since no one suggested they do that either. Clearly, this method is much better than Google assassinating the pope as well. I wholeheartedly agree.

  3. The collaborative part of the process is still to come, Google have tabled the extension to KML, the Mass Market Working Group (MMWG) of the OGC was emailed details of the extensions yesterday and these will be debated at the upcoming technical committee meeting of the OGC.

    The KML standard can only be extended in this way, collaboration is key to the process, as co-chair to the MMWG I look forward to your contributions to the discussion.

Leave a Reply

Your email address will not be published. Required fields are marked *