I just finished writing and filing my own - for normal entities, a normal patent is around $2000 dollars to file with the USPTO, for "small entity" inventors its $600. But that's before the ~two years it takes for the USPTO to review your case and the costs of fixing and refiling. With rush fees, and to get the USPTO to actually review your application in this lifetime (just kidding.. well kinda), add another 10k.
If you want to go with a lawyer (and these days, unless you're willing to literally spend months drafting your application and researching previous claims, it's really hard to avoid) you'll spend anywhere from $5k for a basic patent to $20k for something building on existing tech for a "real world" product.
Whereas copyright is a) instantaneous and self-evident (you don't need to "file" for copyright) and b) actually a much more powerful protection. With patents from small inventors, you have a lot more at stake (enough that it was worth it to go through that process), a lot more has already been invested (so there's more to protect), but simultaneously protecting it against bogus claims is also more expensive.
I know it's a common strategy for patent trolls and extortionists in the US to count on the fact that the defendant cannot afford to defend themselves and would lose money even if they win the case. So small developers usually settle before going to court even on bogus patents.
Is it a different when applied to patents than to copyright? So on which cases are the defendant actually awarded its costs? Which ones are not? What's the difference? Is this case in the article maybe an exemplary case that might change how judges award defendants with its costs? Could this help set patent trolls back from suing small developers?