Developing and efficient file and folder naming scheme requires that you answers the question "why will I ever want to find the files again?" before doing anything else.
Most of my images fall into two distinct groups. The two have rather different naming schemes, each appropriate to nature of the group.
One group is my collection of family "documentary" images. These are sorted into folders labeled by date following the yyyyy-mm-dd format so that an alphanumeric sort is also a chronological sort. Withing these folders the filenames of the individual images begin with a serial number that starts with the date (yymmdd) that the image was added to the library (not date of image) with a unique two digit suffix. The two digit suffix is adequate since these are mostly scans of old film and prints I've never come close to adding more than 99 in one day. Most then has some description of who's in the image. This description often repeats in a large number of images so the begining serial number is necessary.
My other photography is essentially artwork. The date of a particular image is irrelavent. Naming of folders is generally a reflection of subject matter and/or location (e.g. ...\Key West\Historic Buildings\Lighthouse\...). I have some images grouped into portfolios (my series of porches I call "Quiet Porches") is in its own folder tree).
I do a little commercial work, here and there, and those images are sorted into folders by client and sub-folders by date.
One thing I do with images I consider important and likely to be searched for is to fill in a number of the IPTC fields in the file, particularily KEYWORDS, DESCRIPTION, the various copyright fields, and data and location created. I can then find images by searching the folders by these fields using various tools that are available (Adobe's Bridge that comes with PS primarily).