Patents with source code

If you solve the problem of going there in a non-obvious way you can get a patent, however putting the mining feature into claim 1 would be unwise. Claim 1 is normally the most important claim since it defines the invention in the broadest scope.

So claim 1 would be for a non obvious method for space travel.
Claim 2 would be the method of claim 1 wherein the destination is Pluto, unless there is something special about travelling to Pluto.
Claim 3 would be the methods 1 - 2 wherein you mine Pluto, unless the space travel absolutely requires you going to Pluto and do some mining there.

Claim 1 is an independent claim and claims 2 and 3 are dependent claims, depending on previous claims. In prosecuting the application you start with a wide claim 1 and then you bring in features from dependent claims to overcome objections from the Examiner.

These wide claims in a patent *application* is a frequent source of objections from the software community. The final patent tends to be much narrower in scope. The reason for this approach is twofold:
- get the widest possible scope of protection, and
- get a search report that covers a wide part of the art so that you can determine your position compared to state of the art and your competitors.

> In other words, will a patent be granted if the patent application causes »a person skilled in the art« to say »Hmm, yes, that would work once we get to Pluto«,

Correct. The amount of further practical engineering work required varies around the world. In Europe the person skilled in the art should not have to undertake "undue experimentations".

In the US it appears that a fair bit of experimentation is allowed if the story of Richard Feynman's patent for nuclear space propulsion is to be believed
http://www.myspace.com/richard_feynman/blog/332428072

> or must it actually be possible to do whatever the patent application claims in front of the patent examiner before a patent will be granted?

That is not a requirement since it would be impractical. For instance devices for oil platforms are unlikely to even fit into the Examiner's office. It is also unlikely that an Examiner would even want to see a component for a nuclear reactor core in operation (at this point I am tastefully avoiding commenting on specific Examiners...).

In the old days it was a requirement to demonstrate a working model before the USPTO but this was abandoned.
http://www.patentmodel.org/

Supposedly there is one example: if you make a perpetual motion machine the Examiner will want to see it. So does any patent attorney/agent who cares about his reputation.