Here’s an alternative approach that may simplify things.
I do most of my downloads from scientific journals and other sources as rich text captures of the desired text and images, as that lets me eliminated unwanted content such as ads. More importantly, I want to maintain any hyperlinks contained in the article. But that’s another story.
Most of the PDFs that I capture are stored in the Files folder inside the database package. Some are Indexed from files already on disk. Either way, the approach discussed below works.
When I rename files, I almost always do that inside my database. I will select the title of the article, for example and use the contextual menu option “Set Title As”. This doesn’t change the file name displayed in the document’s Path field in the Info panel (if it did, that would break the path to externally linked files, not a good thing).
So, just like you, I have many thousands of documents in my database that have perfectly understandable names, but which often have cryptic file names.
If I wish to send, for example, PDFs to a colleague, I prefer sending files with those understandable and appropriate file names. I do that by exporting the desired documents from my database. Here’s how:
 I create a target folder on my Desktop to receive such files. Why? Because when I use File > Export > Files & Folders in my database, an extra file named dt_storage.plist is also created for each export. That’s useful if I want to import exported material into another database, as that extra file contains metadata information that makes sense to the other database. But I don’t want those “extra” files cluttering up my Desktop, so that’s the reason for the target folder.
 Select the desired file(s) for export and press File > Export > Files & Folders. Choose that target folder to receive the material.
Look at the name of the exported file. It’s the one I assigned to the document in my database! It’s not cryptic (I hope) and my colleague will have some idea as to what the PDF is about when that new file name is read. Problem solved.
Personally, I don’t care (except in some exceptional cases) about any correspondence between the files and folders in my Finder, and the contents of my database. Most of my databases are topical, and are designed so that I can easily move them from one computer to another without worrying about externally linked files. They are self-contained.
Most of my files never had any visible existence in the Finder. That’s true of tens of thousands of rich text, HTML and WebArchive downloads from the Web. Those go directly into my databases. That’s also true of the PDF files created when I scan, OCR and save to a database directly from paper to PDF. For example, I’ve scanned, OCR’d and sent to my databases more than a thousand pages of paper documents, using a beta of an upcoming DEVONtechnologies product.
Note 1: I make it a practice to backup my databases, usually using the DT Pro Scripts > Export > Backup Archive routine. That runs quality assurance routines on my database, then creates the smallest possible compressed and dated archive file, which I save to an external drive, another computer and/or to a CD or DVD disc. I don’t have stability problems, but a hard drive could always fail. Although I backup my computer drives, I provide even more safety for my databases, as they contain the files that are most important to me.
Note 2: My files are not captive to my databases. Any time I wish, I can export them to the Finder. Not all databases (especially some in the Windows world) make this easy, and some make it almost impossible.
Note 3: In the current release of DT Pro there’s a bug that sometimes creates a link file instead of the actual file intended to be exported. That will be corrected in an upcoming release. It’s already corrected in the beta I’m using.