iPhoto metadata

March 16, 2004

Over the course of the last month or so I've undergone the arduous process of adding keyword descriptions to each and every digital picture I have in iPhoto. The idea being that when it is all said and done the library will be very searchable (the descriptors are also useful for creating "Smart Albums" — a new feature of iPhoto 4). I have keywords such as family, friends, florida, etc., but perhaps the most useful are my "weblog" and "web" keywords. I've done this so that in the future, when I change the layout of the site or decide that I want to make larger photos available, going through and figuring out which ones from each set I used on the web will be mindless. For example, if I increase the width of my content container and want the pictures I used inline with the blog to match the new width, all I have to do is search the library for "blog" and then resize those pictures accordingly (same for the pictures in the photos section). Nevermind that iPhoto (yes, even v4) is the absolute worst for resizing pictures and that I have to do it through third-party software. While on the subject I should also mention that I haven't found iPhoto 4 to be too much faster than the previous version (except for launch time).

I knew going into this that it would be time-consuming monkey-work and that I would have a hard time stopping myself from coming up with evermore descriptive meta tags. I'm the kind of person that, if given enough time, would probably have keywords such as "shoes" (to highlight all pictures with shoes in them) or something equally absurd. You get the idea. Fortunately, sometimes the earlier tags actually speed up the addition of new tags. For example, let's say that I want to label all of those pictures that include a particular friend that I've met in California. I would just search for both "california" and "friends" and then weed out the pictures from the results. Obviously none of this is groundbreaking, but it has helped me further organize my collection. Though incredibly daunting at first, it becomes much easier after you have a satisfactory keyword base in place.

What really scares/excites me is talk that the next Mac OS X filesystem will have metadata functionality. Again, this concept is nothing new and has been implemented before (*cough* BeOS); Microsoft's upcoming (*cough* never) Longhorn OS is rumored to run a similar database filesystem. The options and usefulness of such a system are endless. It could include ID3 tags from MP3s and senders/subjects from e-mails and filetype descriptors and author information and... you soon realize that the implications of an all-inclusive descriptive database of every file on your computer are very powerful. Gone will be the days of multiple search apps (e.g., e-mail search, file search, etc.) — only one application will be needed — in the case of Mac OS X, the Finder. Hell, you could probably get rid of deep directory structures as well — just build them on-the-fly as needed, or have the OS create 'smart' directories defined by user-specified rules. All of this says nothing of the smarter interaction between programs that metadata will allow. Ugh, I just realized that I'm starting to get rather tangential to the title of this post and can probably keep writing indefinitely; I'll digress (but expect much more on this topic in the future).

You should follow me on Twitter here