Banning SW patents disables patent system?

One is (like you suggest with your five-point plan) to raise the bar for all patents in terms of patent quality. There have been and will continue to be different patent quality initatives. There's certainly a way to fight for patent quality and it's not as pointless as fighting for abolition. However, if you believe you can achieve a fundamental departure and do away with a double-digit percentage of the patents that are granted, it will become almost as hard to accomplish as abolition. Some progress can be made on that front, but you'll have to be realistic about how far you can get. With a reasonable level of expectations, it is however a more fruitful approach than abolition.

Similarly, one can also make progress on the enforcement side, concerning the rights of defendants in those proceedings etc. But again, that would require a realistic step-by-step approach.

Your second proposal is like "those patents are so bad we must do away with them entirely" (in terms of what kind of subject matter is patent-eligible, regardless of the quality of a particular application). That is the approach I once took but for the reasons I explained I have to recognize the fact that no one has ever achieved any legislation anywhere that would have been more restrictive on the scope of patent-eligible subject matter than previous legislation. The patent system is expansive and at the most one can slow it down or stop it at some point, but a trend reversal hasn't been achieved anywhere even though so many people have tried for such a long time.

Your third option -- saying that if there's copyright, there shouldn't be patents on the same thing -- is a point that anti-patent activists have often made but I never subscribed to it. As today's reality shows, it is possible to grant different IPRs that are relevant to one and the same product. But the idea that "if there's one IPR there mustn't be any others" doesn't work. For example, even if your software is copyrighted, you still have to respect trademark rights (another form of IP). Patents protect a different concept than copyright, and it's obviously possible to develop software that respects all IPRs. That's why Red Hat had to do some inbound patent licensing deals such as the one that gave rise to the article we're discussing here.

Also, the argument about making patents unenforceable against copyrighted content isn't useful because it won't happen for lack of political support. No matter which approach you try (unless it's a reasonable one to patent quality and to defendants' rights) you'd have to convince politicians to regulate an area of economic policy in a way that large parts of the economy support and there's absolutely no evidence that the downside of the patent system outweighs the upside (otherwise there would be more resistance by businesses).

The biggest problem is that patents aren't effective in achieving their stated objectives.

All other discussion - including the whole discussion surrounding obviousness vs nonobviousness - is futile. We can try to solve the software patent problem, but it will not solve the problem with patents on DNA. We can try to tell courts to use higher standards, but smart lawyers will find ways to subvert these standards.

Patents don't work. They have been proven to not work. They didn't work for steam engines, electricity, the telegraph, or telephone. For a while I thought research into medicine might be the one area that it actually achieves something but as it turns out, it doesn't work there either.

We need to abolish patents. Supplant it with a simple system where innovation is promoted by prizes, like the X-Prize does, if we truly wish the government to promote innovation.

Doing anything else does not actually solve the problem.