The computer is not a person and what it does only matters insofar as you may infer that the owner of the property programmed it to do what the owner intended.
As you admit, the property owners did not intend those documents to be accessible. So the only relevant question is: would a reasonable person infer that documents which could only be accessed by editing a URL (by "tricking the HTTP server," if you insist on anthropomorphizing a dumb machine) was intended or not intended to be accessed?
I think most people would assume that documents that can only be accessed by editing an ID were not meant to be accessed. And that really is the end of the analysis.
As you admit, the property owners did not intend those documents to be accessible. So the only relevant question is: would a reasonable person infer that documents which could only be accessed by editing a URL (by "tricking the HTTP server," if you insist on anthropomorphizing a dumb machine) was intended or not intended to be accessed?
I think most people would assume that documents that can only be accessed by editing an ID were not meant to be accessed. And that really is the end of the analysis.