What helps most is to say what, specifically, you would do with SQL Razor, SQL Maesto and Navicat that you cannot do now in SQL in 9. Even just a list of the top five things you would want is very helpful.
Most of my data has geotagged photos and it is common to revisit them and add notes as photos record things other the subject they were taken for.
I've long been an advocate of enabling a new data type in tables, where each record could have a raster field (I'll call it a "photo" field for clarity in this post). You could then have a photo in each record instead of having an image per table with each record being a tile in that image. After all, if a single record can efficiently hold 2GB of coordinates for a geom or a single tile could efficiently, easily, be 1024 x 1024 pixels, the storage architecture could easily handle a photo, even a "big" photo megabytes in size, per record.
We've not done that to date because there is much work to be done to fully articulate what is already there. Much more important are filling out all the vector editing capabilities, filling out all the basic raster editing and manipulation things, merging in the LiDAR super-fast structures and such. Only then could we, I think, peel off some time to do "photo per record."
The issue with that is not so much infrastructure, although that certainly requires work, it is all the accessory facilities that people will instantly require to work with a new photo per record type. For example, you want the ability to browse and to display, to edit, to batch load and unload from many different formats and storage (traverse folder hierarchies, harvest all the usual tags, maintain orientation, etc.). None of that is rocket science but there is a lot of it.
You can also easily spend a lot of time on things not central to the Manifold community. For example, just yesterday I complained about the truly awful GUI for most audio players. You have Windows stuff insisting it wants to show and order my mp3 files by their track name and by tags that Windows thinks I should prioritize, not by the file names by which they are known across dozens of devices. I found myself complaining to a friend that I wished Manifold had an audio type so we could store an MP3 per record and then write a simple app to display and to manage those as desired.
That's nonsense, of course, but not complete nonsense as with web-based interfaces one can think of casting all sorts of data into geographic context. As you walk about Paris and you get near a point of interest your smart phone Tourist app starts telling you about the scene at which you point it, or as you are driving down highway 5 in the central valley of California, just wondering if you can hold out until the famous In-N-Out about halfway between the Bay area and LA, your tablet starts telling you the history of oil discoveries, what all those pumps are in the fields, a discussion of the San Andreas fault, or whatever.
Or, a utility crew is working at a site where there is a break in the water main and the supervisor calls up a video shot on cell phone by the crew that last dug up that spot.
If you have rich media associated with spatial coordinates, all that is possible, and even easy.