I think you have some preconceptions about the relationships between DT Pro and the Finder, rather than actual problems.
DT Pro is not intended to be a replacement for the Finder, nor is it intended that changes made to the database’s organizational structure are mirrored in the Finder, or changes to files (including file name) be mirrored in the Finder.
Those things can be done, in the sense that you can use a script to rename files in the Finder to the database document names, or that you can export the contents of your database to the Finder with the exported material organized as in the database.
I think that’s a strength, not a weakness. I don’t feel the need to reorganize the Finder every time I make changes in the database, especially as I can make organizational decisions in the database that are not allowable in the Finder. For example, I can use replicants in the database and there’s not a precise Finder equivalent for some of the powerful things I do with replicants. I can have multiple documents with the same name (with or without the same content) in a group in the database, and I sometimes have uses for that flexibility.
Here’s my own attitude to the Finder, vis a vis my database: Once I’ve captured material that was originally stored in the Finder, I don’t care much about the Finder.
That’s because I see the purposes of capturing material into my database as enabling me  to view and search the content of a variety of file types (that’s why I fell in love with the original version of DEVONthink back in 2002) and more importantly,  to use its artificial intelligence features to help me research and analyze the information content of my documents.
It’s that “purpose ” that guides me in the design and use of my databases. DT Pro’s ability to “see” contextual relationships among the documents in a database becomes more efficient with increasing contextually related content in the database. To put it another way, DT Pro is more likely to successfully suggest related documents when the “See Also” button is pressed, if there are related documents in the database. I recognize that fact in building my databases.
None of DT Pro’s “competitors” attempts to handle “purpose ” although many do address “purpose ”. So DT Pro becomes a much better research assistant for me than Mori, or SOHO Office, or NoteBook, etc.
I still find it amazing that DT Pro’s artificial intelligence features can run on consumer Macintosh computers. There are a number of “enterprise” software packages, most requiring “big computer” resources, that also do textual analysis in various ways.
So DT Pro doesn’t just index the text of documents. Spotlight does that. There have been lots of indexing programs around for years.
The artificial intelligence features do require an additional overhead requiring additional computer resources. I recognize that fact in building my databases.
So I deliberately construct my databases to have topically related content, so as to maximize the degree of contextual relationships of interest to me in the content while keeping the database sizes within levels that match the available resources of my computers.
Advantages of that approach:
Search speeds orders of magnitude faster, and much more useful, than Spotlight searches;
Fast and genuinely useful assistance by DT Pro’s artificial intelligence capabilities.
How I started my databases:
Back in 2002 I bought DEVONthink as the first program I had found that really worked to bring together my existing collection of text, PDF, HTML and Word files of significant interest to me.
My professional interests center around environmental science and technology, environmental policy and related laws and regulations, and international environmental science exchanges (especially support for graduate training). That covers a broad variety of disciplines.
Using the original version of DEVONthink I was limited to one database, and I had a 500 MHz TiBook with 1 GB RAM. I’ve been constantly adding to my original material, mostly via Web browsing and captures and, later, via DEVONagent searches. As the database grew, it stressed the resources of my TiBook but still allowed me to do things that couldn’t be done otherwise.
For the past several years my main database has hovered around 20 million total words in content. It had hit about 25 million words on the TiBook when I first started using prerelease versions of DT Pro and was able to start spinning off other databases containing materials topically unrelated to my main interests. I did that to reduce the load on the TiBook.
Now I’ve got a PowerMac G5 dual core 2.3 GHz with 5 GB RAM and a MacBook Pro 2.0 GHz with 2 GB RAM. My main database is about 22 million total words at the moment and still growing. As it’s my default database, I periodically “spin off” newly added material that’s more appropriate to one of my other databases.
Most search speeds are phenomenal, often running less than 50 milliseconds. “See Also” suggestions pop up instantly. And my main database remains a very comprehensive reference collection for my main interests and will continue to grow as new material is added. It’s a wonderful interactive research tool.
I’ve got other databases, some almost as large, that are equally fast and useful for their purposes. One, for example, was a volunteer project to help analyze the impacts of Hurricanes Katrina and Rita on the health care infrastructure in Louisiana. That was quickly built from DEVONagent search results (more than 10,000 results). Obviously, that database is completely self-contained, with no corresponding files in the Finder.
Note that I’ve chosen to Import (copy) material to my databases, so that I can easily move them to my MacBook Pro for use outside my office.
What if I had viewed DT Pro as a Finder replacement, and simply dumped the contents of my hard drive into DT Pro? If I had started that way, I think I would have been very disappointed with the results. I’ve got two 500 GB drives in my PowerMac, and many hundreds of GB in files. So I would have gotten a huge database containing lots of unrelated material, and I suspect I would be dissatisfied with its performance, even on my current computers. (Oops! It’s wouldn’t fit at all on the MacBook Pro.)
Question to myself: I’ve got very large collections of photos in iPhoto and a huge collection of music and videos in iTunes. Why would I want to dump those collections into DT Pro? On the other hand, I’ve got some photos and a video of a contaminated lake near Alexandria, Egypt that do fit into my main database on environmental topics, so I did link them in my main database. Topical interests – that’s my guideline.