Tim Cook wrote:
> 
> What should I do?
> 

A radical suggestion - Forget about naming and structuring the images as 
files.

Just give them arbitrary sequence numbers: 1.jpg, 2.png ...

Then tag the files with the metadata that makes them findable.  Use 
either file tag fields or create a database with the FileId, Artist, 
CreatedDate etc.  Even add a thumbnail.

Write something to suck in your existing collection from a directory 
structure and turn the path data into metadata attached to the sequence 
numbered files.

Now you can search them using any decent database front-end, search 
engine or custom application.  Reclassify them by changing the metadata 
rather than moving files between directories.

Trying to keep things in a structure by conflating names, dates, 
descriptions etc quickly leads to the problems that you are hitting.

What about a painting, piece of music or novel that has multiple 
creators?  Which name goes first?  Should you order the structure levels 
by Period, School, Style, Media, Artist?  What about if the same artist 
painted in different styles or media?  What about when one work includes 
or is derived from another?  Duplicate names, multiple works on the same 
subject by the same artist (Van Gogh's Sunfowers for example).

Consider if for example some of the files are linked to web pages, used 
as resources for renders etc by path and name.  Then you correct a typo 
in the artist name or add a date of death.  Suddenly all of the links 
from multiple unknown places are broken.

Now of course if you use a file somewhere the reference to 43857.png 
isn't particularly meaningful.  So instead use a meaningful variable 
name to hold the name - RockyPlanetTexture = '43857.png'.  Or include a 
comment generated from the metadata where you make the initial reference.

"Names are fickle, Identifiers are permanent"

Post a reply to this message