Digital artwork storage

I'm in the process of re-organizing my digital art files-- all the photos of my sketches, paintings, art photography, embarrassing art school experiments, all the high school stuff I used to apply to college. I have original high-quality images, small watermarked versions of them, and working GIMP files (like open-source Photoshop). EVERYTHING from about 1998 onward. That's currently about 7,500 items (some of them duplicates and different versions of the same image), but that excludes all the stuff I have yet to photograph (lots of sketchbooks and work in storage), the many portraits I sold and forgot to photograph, and all the stuff on slides and film, which are sitting in a storage facility on another continent. I'm so glad we're all done with slides.

There must be people out there who are as excited as I am about proper categorization and cross-referencing of files, so here's my system: every item is named by its date (year-month-day) then title, then material, then any pertinent info (such as "sketched-in-Budapest" or "sold_to_So-and-So"), then dimensions. For instance: 2012-2-14Portrait_of_Anne_Bonney-drawn-from-imagination_charcoal_14x20. That way they're automatically displayed by order of date, almost like a visual diary.

Then all files are tagged with color coded tags. I can select "display by tag" so all the colors are displayed grouped together (and still in order of date/name), and pieces with more than one tag show up twice. Or I can scroll down the list and visually differentiate.
red=illustration and any sketches I can use to build my identity as an illustrator (@St.Rhinoceros on Instagram, btw)
yellow=art photos and photo-collage
blue=paintings and drawings (except portraits and figure drawings)
purple=figure drawings
green=portraiture
orange=other (sculpture, prints, odd projects)

Both "small" and "sample" versions of most original files exist ("small" are low-resolution images I can easily send and display online; "sample" images are small and have watermarks and/or labels as part of the original image). They're named the same thing as the original file but with  _small or _sample at the end so they're listed next to the original.

My main goal is to eliminate duplicates of the enormous original files, because my computer storage is almost at capacity and it doesn't need to be. I could store the images on an external drive, but I use them all the time and that would be inconvenient (but to be sure, they're backed up on a USB stick!). So if I need to create a file for a project, such as organizing images for a blog or an application for a gig, my plan is to make duplicates of only the small / sample images, and be confident that there is only one large original, and I know where to find it. I also hope to eliminate the process of creating a sample piece in GIMP only to discover I've already done so and stored the sample somewhere weird. I've edited the truly terrible pieces to be low-resolution so they take up less storage (and I cannot fathom ever printing them).

I'm torn, though, on the tags. I have the option of creating tags that have no color, just a title, in order to further categorize the work (in my case those tags would be several series titles, "good pieces," "sold" "destroyed" "missing" and several storage locations). But is this the best way to do it?

So, do you agree or disagree with my methods? Do you have any suggestions? And most of all, how do you store your own files?

The FBI's fingerprint cataloguing and storage facility, 1940s. Via MessyNessyChic, where you can find more photos of what amounts to a mind-bogglingly massive human computer made of women who were paid in war bonds which they were also encouraged to buy. They categorized the whorls and ridges of each print with a magnifying glass using the Henry System of fingerprint categorization. Each woman categorized up to 35,000 fingerprints per day, working 10 hours 6 days per week. Every soldier, person who'd been jailed, government worker and volunteer had prints on file. The facility held about 70 million prints in 1943. That's significantly more files than I'm dealing with right now.