How about this? Err... not really
* All patents subject to compulsory, non-discriminatory licencing. (Patents exist to reward inventors for sharing things, not just thinking of them.)
+1 at AC: "Why would I invest my time and effort creating the widget, if I don't then own it?"
The main point of obtaining a patent is to deal in the invention as the patent owner sees fit. Where's the economical incentive to go through the process if they can't? One upstream from that: where's the economical incentive to spend on R&D if the results and the economic advantage they may confer can be grabbed and exploited by any competitor for free?
* Patent Office, not holder, sets royalty rates for aforementioned licencing, unless holder wishes it to be licenced gratis. (Prevents effective lock-out by setting stupid rates.)
A Patent Office's business is to examine and reject/grant IP rights, not value the economic contribution of inventions to which they relate (which may or may not be proportional to the amount of R&D invested)
* Patent examiners are paid according to what they reject. (Reduces temptation for examiners to grant frivolous patents.)
Patent Examiners are Civil servants paid according to scale. Anyhow, the last I checked (daily at the coal face), Patent Examiners do not grant 'frivolous patents'. They may have done in the US, but that time is well and truly gone, post-KSR. Thankfully (grudging admission, LOL!).
* No prototype => work of science fiction, not a proper patent application.
That's fine in a first-to-invent jurisdiction like the US, not so in a first-to-file jurisdiction (aka the rest of the world).
Read up on 'enabling disclosure': if I describe how to make a new and inventive widget X in my patent specification, and you can make one exactly like it just from reading it (which is the legal test), why should I be astrained to material costs in building a prototype as well if there is no need to? If widget X does not perform as described and claimed, then the disclosure is not enabling and the patent is invalid by reason of same. This is current (EP/US) law & practice, btw.
* Term of validity runs from date of filing, not date of award. (Prevents "submarine patents".)
Patent term generally runs from the date of filing. The US and some remote jurisdictions sometimes "top up" the term, usually because the USPTO has taken longer than average to prosecute the application. Only bio patents get specific extensions (not that long either) because their commercial exploitation timeframe is quite small relative to the term (time between application made and revenue actually starts).
Submarine patents pretty much ended when the US adopted the "publish-at-18-months" rule already in place everywhere else. There are exceptions (as for all good rules), which are a very small minority (usually relevant to weapons tech).
* Patent is annulled in event that it becomes a legal requirement (e.g. if someone holds a patent on a low-pollution / less hazardous version of a process which ends up being the only way to comply with a newly-introduced environmental / safety law). (Prevents effective privatisation of law.)
Provisions for compulsory licensing, to be used in situations you depict, are already in place in the legislation of most countries to remedy this issue (see e.g. 'Crown Use' in the UK).
* Demonstrably-independent invention is not breach of patent.
Pre-filing date (or reduction to practice in first-to-invent US), it's called a confidential disclosure, which by definition is not shared with the world in the manner a patent application is when published (which is why it does NOT invalidate a patent with a later filing date). Rights of earlier confidential independnent devisors are preserved in most jurisdictions (they can carry on infringing without liability, but e.g. can't license out).
Later devisors should improve upon the disclosed invention, that's the whole idea of the system. Why reinvent the wheel?
* Courts get power to annul patent and charge both sides costs, in event of a dispute. (Gives highly effective weapon against trolls by threatening to cut off revenue stream if abused.)
They already do. At least in most European jurisdictions. Assuming you mean "a Court can invalidates a patent found not new/inventive during a countersuit for revocation" (standard knee-jerk reaction to infringement proceedings), and a Court charges cost to the losing party or apportions costs to both parties for an 'in-between' decision (patent valid but not infringed, patent valid and partially infringed).
Mine's the one with the Rule 71(3) Communication in the pocket.