The Web GIS Duel: Final Thoughts
This is a continuation of Mark Balwanz’s blog posts on his creation of web mapping sites using both ESRI and Geoserver. Today he will talk about his experience creating the site using open source technology.
Over the first three parts of this blog series (here, here, and here), I have laid out my project plan, walked you through my ESRI implementation, and my Open Source implementation. During this fourth and final blog I plan on sharing with you my overall thoughts on both implementations in regards to what I liked and disliked. As I mentioned in Part 1 of this series, everything I share here is just my opinion and is based on this one project. I also want to point out that most of my previous experience, both academic and professional, is based in ESRI and is probably shaping some of my opinions.
I will start by sharing my opinions about my experience building the ESRI version. I found the ESRI build to be quite simple thanks to the enormous amount of information that can be found online. Since the entire stack is from the same company the integration between ArcGIS Desktop, ArcGIS Server, and the API was seamless and very easy to navigate. Publishing data to ArcGIS Server was a breeze thanks to the publishing tools that they have included in the desktop software. I also found it extremely easy to pull ArcGIS Server map services into the application that was built with the API. None of this should come as a surprise since these things were built by the same company and are meant to work together. Another positive I found with ESRI is their documentation. The ArcGIS for Developers website and specifically the JavaScript samples and API reference page contained everything I needed to build my application. The online community of users is quite large as well which provides a tremendous wealth of information within forum posts. I think one of the huge benefits that the ESRI implementation provided was the ArcGIS Server capabilities; the identify task, geometry service (although I did not use it), and many others. I found using ArcGIS for Server to be much easier than GeoServer when it came to writing JavaScript code to query the services. I really did not run into anything with the ESRI implementation that made frustrated me. Of course, if I was building something like this for a company I would have to factor in the cost of the systems as well. The ESRI stack has the advantage of being built by a single company with a single vision of how everything should fit together, and that is something that by its nature the Open Source stack will never have.
So as you can tell I enjoyed my time building the ESRI version, but was really excited to see how the Open Source one would compare. I think I mentioned this in Part 3 of this series, but I was once again very impressed with QGIS. I found it very intuitive and the amount of plugins that you can add is very impressive. The GeoServer Manager plugin made the publishing of WMS to GeoServer a trivial task. As mentioned above, I do prefer ArcGIS Server over GeoServer, but that being said the GeoServer manager page is very easy to use and seems to be well thought out. The open source JavaScript libraries were also quite impressive. Although I ended up switching from Leaflet to OpenLayers 3, I still came away impressed by how easy it was to use Leaflet was and will be looking to use it in the future. The one problem I had with OpenLayers was how hard it was to find accurate information about OpenLayers 3, online. Most everything I found was for OpenLayers 2 and that does not migrate very well to OpenLayers 3. Even the book I bought for OpenLayers 3 had some code that is not included in the library anymore and therefore did not work. The samples page on the OpenLayers 3 website was helpful, but overall I was not impressed by their documentation. Another area where I felt was lacking was the integration of GeoServer and OpenLayers. I found it complicated to perform relatively simple spatial queries against a GeoServer WMS from within OpenLayers. What made this more difficult is that I had to search through two sets of documentation to solve my problem (GeoServer and OpenLayers 3) rather than just one. I do think that some of these problems would probably decrease the more I used GeoServer and OpenLayers, but better documentation would make it easier for people to jump into the Open Source world. After completing this project though, I am pretty confident that the Open Source world has the capabilities to match ESRI and the fact that the software is free is a huge benefit. Additional work to more seamlessly integrate these Open Source projects would go a long in making them more user friendly.
As someone who has spent most of his GIS life working with ESRI, I have to admit it was a little uncomfortable to move into the Open Source world. However, being uncomfortable was a good thing as it pushed me to learn a lot of new technologies. I think Open Source can be a great way for organizations to start using GIS as there is a lot less upfront cost and with a little research you can find the Open Source project that best fits your need.
I hope you have enjoyed this blog series and have maybe learned something that you can use. Please feel free to leave comments if you have any questions and thanks for reading!