
Unfortunately you have to stop before you start.
You would need to license the core files from Microsoft.
This would really only be true if the developer was trying to redistribute the files with the game. There are many open-source "community" editions of games these days that require the original game to be installed so it can use the art files from it. I haven't checked in the FSX eula to verify this, but chances are it wouldn't violate anything (I haven't found a game yet that it does).
One big hurdle would be the patents MS holds on FS. There are large parts of the sim that are covered by very vague patents. One would have to very carefully implement these parts of the sim under the guidance of a lawyer to ensure MS couldn't claim patent infringement.
So I noticed all the bad comments about Microsoft Flight and this idea came to me. We have a bunch of great developers here and I thought if we all worked together we could make our own flight sim. Using all the base FSX files and all the developer programs for FSX we could make a pretty cool sim. If our talented people did the complicated stuff and created the airplanes and we all made a few airports we would all be able to have a say in the sim.
Trust me, there are many people in the 3rd party addon business who have considered exactly that. I'm not trying to start rumors or cause wild speculation but I guarantee you that some people on this very forum are researching the possibility of this as we speak, and might even already be implementing it.
As an aside, making a flight sim is not hard. I have "made a flight sim" already in the course of my university studies. Making a flight sim that has the highest-fidelity flight dynamics, Crysis 2 style graphics, and the extensibility to create a huge 3rd party market is what we are really talking about. As we saw with Flight, if one of those things doesn't live up to the ridiculously high standards that this community demands then it is deemed as an instant failure.
The fact of the matter is that this is a business. Most of us are thinking about this like engineers, not businessmen. Is it possible for one man to build a flight sim? Sure. Austin Meyer did it (he was the only one working on X-Plane until the most recent version, I believe). The FSX/ESP engine is roughly 1 million lines of code, as stated by the folks at LM. I am staring at a personal project of mine right now, involving high-fidelity 3d graphics, that is sitting just shy of 100,000 source lines of code as calculated by the Visual Studio code metrics. I am a student, far from being a graphics/simulation expert, who works on this in his spare time. So it is completely possible that a small team could produce a realistic sim given enough time.
Unfortunately, those developers have to feed their families somehow. Taking on such an endeavor is an extreme risk and I think that's why many projects like this fail. Sure, they can find ways of getting funding, but things like this take time and the right mix of people. These people also have to get it right. They can't be learning as they go and must make the right design decisions from the beginning. Anyone who has worked in the engineering field for some time can tell you that it is possible to do a project like this on a tight budget with a small staff if they are all-star players. This is just an extremely risky business decision, and most people who would be interested in this type of field wouldn't be willing to take such a huge financial risk.
Sorry for the rant. I guess my point is just to give it time. I guarantee you that us "simmers" have not been abandoned by any means. I think the future is bright for simulation and can't wait to see what is coming a few years on down the road.