I would be interested in knowing what people who don't want to use, or can't use, the normal tools for whatever reason do for this sort of thing as well.
This is what I do. Disclaimer: I'm a programmer so a lot of this is done with scripts and assumes you are willing to do things to the underlying filesystem. And it will probably only work on Unixoid machines. This probably will not work for you, but there might be ideas here.
First of all I try to make sure that the EXIF dates make sense. This is trivial for images which originated digitally, but for images that originated on film but were scanned much later it's not trivial. One option (which I often do) is just to give up, but for archival purposes I think you want to get things right if you can (or as close to right as you can). ExifTool is your friend for doing this.
Then, once I have a bunch of images with sane datestamps, I use a script I wrote called
pig (photograph ingester) which will look at a bunch of files, and making use of the date information in them, move them into a directory structure which looks like
.../Masters/YYYY/MM/DD/<image>, making the files read-only (and optionally the directory structure above them read-only once I've decided there will be no more images for that period).
pig uses ExifTool to do the work of course.
This is the master catalogue of photographs, organised by date: these get backed-up in lots of places (I think I have four different mechanisms and six locations for backups). It's now easy to find things if you know when they were taken.
Now if I want to catalogue things by some other way (subject or whatever) then what I do is have one or more other directory structures whose toplevel has a name which tells you what sort of things it might be. So I might have
Projects/Appleby/... for a project on Appleby fair. Those directories get filled with symbolic links to appropriate image files in the
Masters directory. Edited versions of the images end up in these directories.
And that last step happens ... mostly by hand at present. I'm hoping to find an application which is sufficiently automatable (but sufficiently small and non-intrusive) that I can mark the images visually (none of mine are huge enough that this is a problem) and then automate the creation of suitable links.
Why I keep the master catalogue in a structure I invented. I simply don't trust proprietary tools not to occasionally destroy all my images or tie me into some opaque structure which I then need to escape from when they decide they'd like to remove all my money so I can using the tool, or that the tool is no longer worth supporting. (Yes, I used Aperture, not going through that again). By using a storage structure I constructed and which relies only on the Unix(oid) filesystem I can be sure that I will be able to maintain my images essentially for ever, at the cost of some convenience. I have nothing against LightRoom, or whatever, I just am too paranoid and too interested in understanding how things work underneath to use it.