I was completely tranced by the process of identifying names to faces in iPhoto last night. Since reading some sociological observations of why people will obsessively check email, i can see that the Faces module had similar qualities. There was a quality of intermittent rewards [wow, that's a creepy article about successful manipulation -- Christine & i spent time discussing the concept and Skinner and Cognitive philosophy's "debunking" of behaviorism....] that made it addictive: first, there was discovering the photos of people from years ago. That was an intermittent reward because sometimes there were photos i expected (not a reward), sometimes photos i'd forgotten (emotional jackpot). Then, as i developed the training set of images for the engine, there was the intermittent reward of having the training succeed and the system ask "Is this So-And-So?" -- very intermittently. I wonder if the engine is just effective enough to have the intermittent reward psychological trigger to engage the user in tagging hundreds of photos.
This morning i awoke thinking about the photos i'd run across that iPhoto "couldn't find." It corresponded with my memory of putting certain blocks of years on the removable Hitachi drive that dies so quickly. Ah, but the photos are still on the disk, i discovered! Now to figure out how to get the path names changed in iPhoto. Last night's experiment seemed to point to a one-at-a-time solution that is impractical. I woke this morning to the idea i could change the path names in an XML file but some of the path names are in binary files. I think just deleting the photos and reimporting is what needs to be done. Regrettable: because if there was any metadata investment, it's lost.
The fragile nature of the iPhoto database is what drives me nuts. I had saved off a "broken" iPhoto library hoping some day to do some data archaeology in it -- that, i think, is lost with Hitachi disk.
Thinking about that did remind me of all the uDig mapping stuff i did before the break: the path names on the new machine would be broken, too. Fortunately, that application seemed to be sensibly constructed for easy access to the different catalog and project files that specify the paths to the data sets. I fixed that this morning, i think.
---
This morning has been lovely with Christine awake and sitting next to me reading from her sociology readings for her science & technology course.
I'm about to bolt out the door. First, traffic patterns around San Mateo at 9 am have changed for the worse. Also, i've got to finish something that i only sort of finished before my vacation. Today is the day. Sheesh. Procrastinatix, me. (There is no intermittent reward with scheduling a software development project. Just intermittent punishment.)
This morning i awoke thinking about the photos i'd run across that iPhoto "couldn't find." It corresponded with my memory of putting certain blocks of years on the removable Hitachi drive that dies so quickly. Ah, but the photos are still on the disk, i discovered! Now to figure out how to get the path names changed in iPhoto. Last night's experiment seemed to point to a one-at-a-time solution that is impractical. I woke this morning to the idea i could change the path names in an XML file but some of the path names are in binary files. I think just deleting the photos and reimporting is what needs to be done. Regrettable: because if there was any metadata investment, it's lost.
[pennyroyal:TIA-more/library/iPhoto Library] me% grep 020424_cn.jpg * AlbumData.xml:/Volumes/TIA-more/library/iPhoto Library/Originals/2002/0204-X/020424_cn.jpg AlbumData.xml:/Volumes/TIA-more/library/iPhoto Library/Data/2002/0204-X/020424_cn.jpg Binary file iPhotoAux.db matches Binary file iPhotoMain.db matches
The fragile nature of the iPhoto database is what drives me nuts. I had saved off a "broken" iPhoto library hoping some day to do some data archaeology in it -- that, i think, is lost with Hitachi disk.
Thinking about that did remind me of all the uDig mapping stuff i did before the break: the path names on the new machine would be broken, too. Fortunately, that application seemed to be sensibly constructed for easy access to the different catalog and project files that specify the paths to the data sets. I fixed that this morning, i think.
---
This morning has been lovely with Christine awake and sitting next to me reading from her sociology readings for her science & technology course.
I'm about to bolt out the door. First, traffic patterns around San Mateo at 9 am have changed for the worse. Also, i've got to finish something that i only sort of finished before my vacation. Today is the day. Sheesh. Procrastinatix, me. (There is no intermittent reward with scheduling a software development project. Just intermittent punishment.)
Tags: