I did write a C program that strips HTML tags from a text file.

In other words, each photo session is given a meaningful name, and I put the year and month in front of that for directory sorting purposes. Sometimes, I go one level deeper where the ‘event’ is “2007 01 India Trip” and then under that are separate collections based on each of the major cities/sites visited… as if I am to look for an Indian photo I will know where to go, and it saves having too many higher level directories. By the time I get to 75,000 photos I will have to put some of the 1990s to 2004 photos in a year-based higher level directory.

This approach means that there is no need to access any external database or spreadsheet to work out where a photo is. For most sessions, photos are left as just the auto-numbered filename, but if it is a great shot, I put an explicit filename on the photo, such as “Great shot of Claudia jumping Princess 12oct08.jpg” and when I access that directory, I can easily see the few great ones, without looking at thumbnails. Also, for truly great shots I COPY the image to a separate higher-level directory “Claudia’s favourites”, as it is never worth re-diving into 50,000 anything to re-find something that is within your top-200!!!

Using a well-organised directory level organisation beats the hell out of any photo-organising software, as that software may well go out-of-support, yet your JPEGs will always be readable. However, best to put explicit description in the filename, not elsewhere, so that when people have a copy of that file, you get the description ‘locked’ with the file.

When I do cut-down sizes for emailing say, I create a “/sqzd” directory under that event, use Irfanview to create 1024×768 (say) images at 40% compression, then work through them, deleting the squeezed versions of any not worth emailing, till that directory contains the top-5 or whatever of the session… then I email just those, and leave that squeezed directory in place, as the total space is just 5x60kb in total, and who knows someone may ask for it to be emailed again…

In fact I’ve set up an imaging bureau and believe that with things like medical images (or scans of medical paperwork), nothing beats having the ‘medical record number’ followed by ‘surname’ followed by ‘first names’ and optionally followed by ‘date of birth’ on ALL images. As that way, if one is searching in a hundred years time, one will still easily find all material for that particular patient, and none other… whereas all databases and other required search intermediaries may have been long lost. I’ve been guest-lecturer to national archivists associations etc – the real risk is not losing the information, but knowing how to access it… like all the tapes kept which no longer have readers to retrieve the information.

Having said the above, I also use Diskeeper (home edition US$29) in WinXP, so when dual boot is brought up in WinXP every now and then, Diskeeper does a full-automatic background task defragmentation of the 2TB of external NTFS drives, putting all directory information together etc, to optimise directory opens and file searches. Diskeeper is the best commercial disk optimizer… and the effect is to fully-offset any delays caused by using long filenames.

However, while I differ with the author on the best way to organize photos, let me suggest that the article is great for its description of how to use script to do detailed automated directory operations…
Graeme Harrison (prof at-symbol post.harvard.edu)