2. Design notes

The heart of an image management system is the catalog number, a unique identifier for original images. For many years preceding the design of this system, my catalog number had this format:

YYYYMMDDxFF

YYYYMMDD

The date as a four-digit year, two-digit month, and two-digit day. Note that either the MM part or the DD part may be “00” if the month or day is unknown.

x

In most cases, this separator will be a period (“.”). However, to address the problem of multiple shots on the same day with the same frame number, it may also be a lowercase letter.

While indexing older slides, my method was to use a period for the first or only slide with that frame number on that day, then assign letters a, b, … as duplicates appeared for cataloging.

However, recently a better practice emerged: use a period here for the first or only roll of film (or digital camera) used on that day, then assign letters for additional rolls or cameras. This makes it easier to see the actual shooting sequence on a given day.

FF

For film originals, this is the frame number. For unmounted film, this number appears in the margin of the original; if the numbers 1, 2, … appear exactly on the frame boundaries, the alternate sequence 1A, 2A, … is used, but the trailing A is omitted from the image's catalog number.

For most all the older mounted 35mm slides, this catalog number appears in ink on the actual slide mount. Holding the slide in its correct orientation, rotate it around the horizontal axis, and the catalog number is written on what is now the top edge. Typically the word “Kodachrome” (or whatever) will appear right-side-up just below this edge.

Digital cameras typically assign a four-digit number to each frame. My Canon A530, for example, names each image “IMG_nnnn.JPG”, where nnnn starts at 0001 and eventually wraps around. Assuming that one never shoots more than 10,000 frames a day, this number plus the date should uniquely identify a frame for a given camera.

I also try to avoid long, featureless strings of digits nowadays, as they are hard to read. The Tech Computer Center's ticket system, for example, assigns each ticket a number of the form YYYYMMDDhhmmss, but the resulting lump of 14 featureless digits can be hard to parse, especially in a small screen font. There is an ISO date standard that recommends dashes as separators, so a full timestamp would look like “1949-09-05T14:30:15”.

Here, then, is the current general form of the catalog number for originals:

YYYY-MM-DDxFFFF

where the normal character for x is a hyphen (“-”) instead of a period. The use of lowercase letters for the x separator is still supported, for disambiguating film frames with the same number on the same date, or for disambiguating frames from different cameras with the same number.

An XML-based bird image cataloging system

2. Design notes