Sean Gillies Blog

The answer to all your questions is probably just NASA politics.

From a World Wind forums post a while ago, the NASA developer (there's just one) said that the original work was done in C# and Managed DirectX because that meant the project could be done in the very limited time and budget allowed while still reaching as many people as possible. "Going the C# route was the difference between finishing the project, and not finishing the project."

http://forum.worldwindcentral.com/showthread.php?t=7

The problem with a mono World Wind is that the reliance on DirectX keeps it from being cross platform. A Mono/OpenGL World Wind would be cross platform but requires a major rewrite...almost as much as a Java World Wind and Java is more widely accepted and more importantly it makes World Wind easier to integrate with existing java projects.

Is it a duplication of effort? Yep. Is a Java version going to be really useful? Yep. Will I have to rewrite all my plugins? Yep, but it makes it easier to integrate with other Java projects around my lab. Being java it reduces the entry barriers.

For example, I know that one of the mission planning systems for the Mars Reconniassance Observer is written in java and its visualization isn't as nice as the World Wind one. But integration with a C# WW would have been annoying so there wasn't much interest. A java WW would have been much easier to integrate and may have gained traction. A mono one? Not so much.

There's nothing wrong with .NET or mono but there are some advantages of java that outweighs the duplication effort IMHO. How that pans out in real life...well, we'll see.

Regards,

Nigel

ObDis: Not speaking for my employer

The short answer is yes.

But the longer and more ironic reason for the Java version is resources. :)

Just because a project is done wholly or in part by NASA does not mean they have infinite resources to do what ever they want. So if one year a group is willing to fund Chris Maxwell and company to do X, does not mean that in the next year funding will continue. So in this case someone else inside of NASA wants a World WWind style viewer for integration for Java (Eclipse) apps and is willing to pay for it. Just like in any other large corporation, development is driven by need. (Or a managers selling abilities)

With regards to the Mono question, as someone that has compiled WW in Mono, all of WW compiles and if you Mock the Direct3D code you can actually get a Window up with a large black rectangle in the center on Ubuntu, so I would say Mono is pretty mature. But what is missing once you overcome that hurdle is a 3D abstraction layer for using D3D or OGL in the application, depending on your choice. There are many OpenSource 3D engines that do this quite well and a few already running with Mono wrappers. Anselm Hook has thrown out the 'couple of months' number for a conversion.

And Tao, the OGL wrapper for .NET has come a long way in the years since WW has been developed so it might be much, much easier to do this now. In the past Summer of Code competition we actually had two submissions for making WW use OGL, but got prioritized down. Maybe we will accept them next year. So if the need becomes strong and someone steps up, it is technically feasable.

Viva la Open Soruce!

Simple answer: Funding.

DirectX managed mode actually gets slightly better performance than Java-JOGL. Anyway We currently have Tao-the OpenGL Mono wrapper in the Worldwind source tree.We use the Tao GLUTessalator for rendering polygons. Switching all the rendering code to OpenGL is a bit of work as anyone can imagine. The Java Worldwind project is in effect switching the renderer to OpenGL. Hopefully it will not be hard to port that code to Tao once someone has the initiative/funding/itch to do so.
Anyone from Novell willing to do some funding ?