I think the problem is that software / algorithm patents have broader implications and abuses than patents for chemical formulas. They are also being granted for topics that many would consider obvious to those active in the field. The issue isn't protecting valid and specific invention from copying, it's the ease with which software / algorithm patents can be used to sue / bully others with legitimate (not copied) creative inventions. I think the ephemeral nature of software makes it inevitable that a patent eventually gets approved by an examiner if resubmitted enough times. A bogus chemical patent is probably easier to prove bogus.