- From: Garrett Wollman <wollman+semantic-web@bimajority.org>
- Date: Tue, 24 Jan 2006 00:36:10 -0500
- To: Richard Newman <r.newman@reading.ac.uk>
- Cc: semantic-web@w3.org
<<On Wed, 4 Jan 2006 11:47:20 +0000, Richard Newman <r.newman@reading.ac.uk> replied to my message about the difficulty I was having in generating the sort of metadata I wanted for a photo-gallery application: > (See "The author's existing work", [1]) I found this to be very helpful. > This is actually quite fair; you haven't expressed the knowledge that > people with the same name are the same people, and the SW would be > unwise to make that assumption. And, of course, I don't have that knowledge, either. It's a conclusion that only makes sense in the context of a specific collection; it's not globally true, even within the user community of my specific application. (Well, actually, I suspect my user community can be counted on my fingers, but I don't want to /a priori/ limit myself in that way.) >> [I wrote:] >> a) Technical: how the photo was taken, at what resolution, in what >> orientation, etc. I am mostly not concerned with this, since it is of >> no value to my application. > Though I do cover it in my iPhoto work, because the information is > there. I turned out that I actually needed height and width, so I used N. Walsh's exifi vocabulary for that. (I could extract all the other information, but it would bloat the RDF with lots of information that I have no particular use for.) >> b) Temporal: when the photo was taken. This is easy to accomplish and >> the choice of representation is obvious. (I used dcterms:created and >> represent the date in DTF.) > Good. I necessarily used archivedOn and lastModified properties, > because those are what iPhoto gave me. Well, the date of capture comes direct from the EXIF data, so it should be readily accessible. (I currently retrieve the EXIF data I care about once, at transfer time, and store it as static comments and XML elements in my gallery description file. This is clearly suboptimal but it works.) I think I'm mostly comfortable with the structure now, although an angel loses its wings every time the RDF validator asks graphviz to make a diagram of it. Tim's Tabulator does a reasonably nice job of browsing the structure, although it also points out areas with room for improvement. (Since all the improvement has, in my current architecture, to be written either directly in XSLT or even worse in an XSLT-generated Makefile, there's a limit to how much effort I'm likely to put in.) I'm certain the experts here will find many more things wrong with it. You can see a sample (from one of the galleries that I've already annotated) at <http://gallery.bostonradio.org/2005-06/boston/index.rdf>. -GAWollman
Received on Tuesday, 24 January 2006 05:36:34 UTC