The purpose was to both incentivize invention, but also require detailed specs of that invention to be given to the public. In exchange the inventor gets exclusive rights for a limited time.

If trade secrets cause knowledge to be held from the public and then ultimately lost that would be worse.

Whether that’s still relevant today I don’t know. In cases with massive research R&D like drug discovery they still seem really important.

In that case a trade secret wouldn’t be enough to justify up front investment if the drug can be copied and sold as a generic for cheap along side it.

I'm not sure I understand how a patent system would ever encourage knowledge to be made public in a way that wouldn't have already happened.

Even with a patent system, if your innovation is not easily reverse-engineered then you have no reason to patent it. If you invention is easily reverse-engineered, then the knowledge becomes public effectively as soon as you release your innovation, and that is when a patent is most useful.

Even if your invention isn't easily reverse engineered, it could still be re-engineered by others once you've demonstrated the concept and the market. Trade secrets aren't guarantees.

Right but that's my point. I don't think there is clear evidence that it encourages innovation, and if the technology can be re-engineered from the end result then you are not really helping society with a patent system. If it cannot be re-engineered from the end product, I don't see a reason anyone would use the patent system.

Gotcha, but I guess the counter would be that you can't really know what someone else will be able to figure out. Some groundbreaking historical discoveries were made in quick succession by multiple people working independently; others weren't.

I think the system was not designed so much for end products, but for industrial technology that produced end products.