The intent maybe so, but the practice is exactly opposite. Not only patents are being used to effectively block progress in certain areas for years (because many researchers can't or won't pay license costs and us such unable to incrementally build on existing knowledge), but reading a patent if you work in similar area is the worst thing you can do - after you red it, all you work will be contaminated by suspicion that you have used patented IP without licensing. So if you suspect there's a patent in some area you work in, the best thing to either pay somebody to search for patents but never read them by yourself, or to ignore the patents and assume the risk. Otherwise you are exposing yourself to claims even if your work is 100% original - you'd have to prove it's original and familiarity with the patent would work against you. So it gives huge incentive to impede flow of information.

Additionally, if you read any of the software patents, they are written in intentionally obscure, outrageously dense and purposely obfuscated language, as to sound more generic and vague and capture more "space". They have tons of claims which repeat the same thing with tiny variations, they describe most mundane things as if it were huge novelty, and they are made as hard to read to a common person as possible without switching to Sanskrit. They never increase sharing of any information and never are written with this goal in mind.