KatolaZ <katolaz@???> wrote:

> All very good points, indeed, which unfortunately become automatically
> nonsense in the case of software. 17 or 25 years are the blink of an
> eye for hardcore 19th centrury industrial innovation, when the patent
> system was invesned, but correspond to several geological eras in the
> case of software or modern technologies. And this does not help
> innovation, rather hinders it.

Which is part of the "as currently implemented is broken" I've already stated.
Firstly, as you correctly point out, timescales for many (but not all) categories are far too long.
Secondly, the system, especially in the USA, is seriously broken in that patents are granted for "obvious" stuff or stuff that really isn't defined.

I still believe that if patents were only for a timescale relevant to the areas (probably no more than 5 years, another area for debate, in the computing world), and if all applications were properly vetted, then patents would be a positive thing.
On the flip side, regardless of what anyone thinks, getting a new drug from "I've had an idea about something", through research (where many idea turn out to be a dead end), trials (where more ideas end), approval, acceptance onto lists (cf NICE in the UK), and actually into patients typically takes decades. Against this, a 25 year patent on a new drug doesn't seem that unreal - at least in the real world where a business will only invest if it sees the opportunity for a profit at some point.