I’ve been talking quite a bit over the past few weeks about why I think Google Earth will become the default GIS viewer in the next year, but I thought I’d also post about what might keep it from coming that viewer.
- Support for projections – Sure one can always change the projection of files, but I shouldn’t have to do that. GE should be able to handle different projections on the fly. ArcGIS has done this for years and I think it is a requirement for any GIS data viewer. Because most of our work is for the U.S. Department of Defense, we deal in UTM and State Plane most of the time. I’d just rather not have to worry about changing projections (or have an ArcGIS extension do this for me).
- Customization – I’ve seen some really nice attempts at creating an interface inside GE (the National Geographic example is great), but I’d like to have an API exposed so I can add buttons and forms so I can add or subtract features I don’t need. Google Earth is simple, but sometimes I need something even more simple.
- Printing – I know you can upgrade to a paid version of Google Earth that allows better printing, but I’d like to be able to set up templates that would standardize how maps are being printed. Some will say this isn’t what Google Earth is about, but I say any competitor to ArcReader/ArcExplorer should be able to print maps as I want them to be.
- Advanced Query Tool – As we load up more complex information into Google Earth, we’ll need a better “Find” dialog to get at this data. I’m not sure how Google will view this since they are all about “I feel lucky”, but I know many Engineers and Planners who will want this capability.
- Metadata – I’ve said this before and I’ll say it again. We need better information about what the acquire date of the satellite imagery is and the source of the road information. I don’t want phone calls from clients telling me that the photo is out of date when at the bottom of the screen it says “Copyright 2005 Google”.
I know what you are thinking as you are reading this and I agree. Google Earth is a “geoviewer”, not an analysis tool. GE is about a 70% solution toward a great GIS tool, but its limitations are showing up in how I’d like to deploy it for my clients. I can see many cases when Google Earth is all they would need, but I suspect unless Google Earth opens up more on many of the above points as well as their planned updates to the datasets, I’ll probably be looking more toward ArcExplorer (I say this without even having a test drive so I reserve the right to say I don’t like ArcExplorer in the future) which will do many and more of what I want/need in a GIS viewer rather than Google Earth on its own.
More Google Earth Wishlist Items.
- From Matt Perry
- Support for other file formats- importing shapefiles, sde layers, wms, wfs would prevent forcing users to convert their data (and consquently have to maintain two datasets). Another route is that some real-time conversion utilities will arise that will let you add these disparate data sources with some server-side logic allowing you to maintain the data in it’s native format while serving it up as KML on-the-fly
- Support for large vector datasets – I recently converted a 60,000 feature line shapefile into KML and the display totally choked. I cut it down to 600 and it was fine. The stock Google road layer doesn’t even render properly at all scales/angles. (At least not on my machine). If GE is going to be for real, they need to gracefully handle big vector datasets.
10 Comments
GE does NOT want to be a GIS viewer, but it’s going to be. ESRI ArcExplorer WISH it could be “the” GIS viewer, but is faaaaar away… Where is the new AE right now??? Where are all the Arcweb services??? In Dangermond’s and your dreams…
-C
Chris, did you read my post or did you just look at the title and reply? I’ve been saying for weeks that Google Earth will be the default viewer of GIS information. This post is of what needs to be done before it will be able to become that viewer, not that ArcExplorer is going to be better. I said AE could become the viewer I use for my clients, but only because I’ll need features that GE doesn’t have.
Another issue with GE is getting the client installed on computers. Right now most DoD clients of mine don’t have permission to install GE, but they can install ArcReader.
Right on. GE has some limitations that prevent it from being a “real” GIS app.. but boy is it cool!
I would add two things that need to happen before GE is the real deal:
importing shapefiles, sde layers, wms, wfs would prevent forcing users to convert their data (and consquently have to maintain two datasets). Another route is that some real-time conversion utilities will arise that will let you add these disparate data sources with some server-side logic allowing you to maintain the data in it’s native format while serving it up as KML on-the-fly
Support for large vector datasets:
I recently converted a 60,000 feature line shapefile into KML and the display totally choked. I cut it down to 600 and it was fine. The stock Google road layer doesn’t even render properly at all scales/angles. (At least not on my machine). If GE is going to be for real, they need to gracefully handle big vector datasets.
Those are really good Matt, not sure why I didn’t think of them. I’ve added them back to the main post so others can see them.
As a fellow GIS dude working for those in .mil land I would like to pass on a little tidbit that will definitely keep Google Earth from becoming a GIS viewer of choice (at least for those of us working in DOD).
The local IT guy informed us today (right before he made us remove GE) that the Army has just decided to ban Google Earth on its network. Apparently the amount of bandwidth used and lack of admin requirements on the installer (one of the best features if you ask me) were big problems for the all-controling IT types. Any computers found running GE during (frequent) scans will be promptly booted off the network.
I seriously hope he is wrong on this, as our roll out of .kml based data sets was just beginning to pick up steam.
-Mike
Hi James,
Another showstopper for me is that some of the imagery that I’ve seen in Google Earth doesn’t appear to be registered accurately. If I cannot overlay vectors and have them line up with the imagery it makes Google Earth pretty much useless for more advanced applications.
Another issue, encountered today, is that there is no way to filter data based on attributes. (Actually, I’ve yet to see any way of including attributes in KML.) The example was: Show me the location of events that occurred between two dates.
Regarding capture dates for the imagery, I suspect that Google is not making this public for privacy reasons. Knowing the capture date gives you a whole lot of context.
Andrew
I’ve wondered about the registration problem too. Take a look at the screenshot in this post.
James,
That spatial offset looks very similar to what I’m seeing, which is about a 20-25 metre shift. Not good if you are trying to display building footprints, street centrelines, etc.
Andrew
I haven’t had too much time to figure it out, but I guess it could be a couple things. Not knowing enough about the base data from Google it makes it very hard to test it.
I might try and figure out what the x/y shift is and see it if changes over the globe.
As much as I would like to see GE as a default GIS viewer, I agree it probably won’t be for the many reasons already listed. The biggest hurdle of all are the all-controlling IT departments. Web-based apps, no matter how primitive, are huge precisely because they bypass that layer of bureaucracy. But to see wishlists of ideas speaks volumes about GE’s public mindshare so immediately after its “debut.”
I think we’re really waiting on a 2nd generation AJAX website.