When academia attacks
I went to a presentation by Susanne Boll this week that centered on the creation and capture of context in multimedia engineering. The methods that Dr. Boll presented to contextualize multimedia records, such as digital photographs, were interesting in that they approached the problem from the point of view of the machine rather than that of the user. The majority of the talk covered possibilities in the detection of context both during media capture and through analysis of data, an interesting alternative to user tagging when organizing (i.e. contextualizing) multimedia data.
I am still not convinced that this is a plausible method of contextualizing data, but I do hate tagging my images on Flickr and would prefer a more automated alternative, although I am not sure that such an alternative is possible. It takes many widgets to capture context automatically and those widgets will only record machine-measurable data. It also takes a lot of programming to analyze data for measurable traits like spatiality and those analyses will never to take into account emotive meaning. So the resulting contexts might not be so useful in terms of the way that people associate meaning and quality to pictures, but it will be interesting to see how her project turns out.
The most surprising part of the presentation, however, was when one of the attendees asked if it were feasible to create a generic context framework outside of the domain of multimedia that would be applicable to all types of data. My first and continuing thought is that the individual is an idiot. A generic context framework applied to data is a contradiction in terms. Such a framework would be the same thing as having no framework or an infinitely extensible framework, because it is the combination of object’s properties and its environment of usage that determines its relevant context. Context is contextual by definition, not something that can be planned for in all of its eventualities, and can only be structured within a given context, such as digital imagery.
Now watch me be the idiot ten years from now when I am a lowly serf in his coder plantation after selling myself into servitude to access to his marvelous context machines.
P.S. Contrary to what one of the other attendees brought up, context is not the same thing as metadata. It’s just not. Metadata does not take into account the currency of an environment. Booya, of course, will always be the sound that a shotgun makes.
P.P.S. If you are the idiot I was talking about and would like to defend yourself, just email your defense to me and I will post it for everyone else to see, all flames included. I prolly deserve it.
I am still not convinced that this is a plausible method of contextualizing data, but I do hate tagging my images on Flickr and would prefer a more automated alternative, although I am not sure that such an alternative is possible. It takes many widgets to capture context automatically and those widgets will only record machine-measurable data. It also takes a lot of programming to analyze data for measurable traits like spatiality and those analyses will never to take into account emotive meaning. So the resulting contexts might not be so useful in terms of the way that people associate meaning and quality to pictures, but it will be interesting to see how her project turns out.
The most surprising part of the presentation, however, was when one of the attendees asked if it were feasible to create a generic context framework outside of the domain of multimedia that would be applicable to all types of data. My first and continuing thought is that the individual is an idiot. A generic context framework applied to data is a contradiction in terms. Such a framework would be the same thing as having no framework or an infinitely extensible framework, because it is the combination of object’s properties and its environment of usage that determines its relevant context. Context is contextual by definition, not something that can be planned for in all of its eventualities, and can only be structured within a given context, such as digital imagery.
Now watch me be the idiot ten years from now when I am a lowly serf in his coder plantation after selling myself into servitude to access to his marvelous context machines.
P.S. Contrary to what one of the other attendees brought up, context is not the same thing as metadata. It’s just not. Metadata does not take into account the currency of an environment. Booya, of course, will always be the sound that a shotgun makes.
P.P.S. If you are the idiot I was talking about and would like to defend yourself, just email your defense to me and I will post it for everyone else to see, all flames included. I prolly deserve it.
0 Comments:
Post a Comment