screenLit

__Michael Larrabee__

Kevin Kelly's writing on the prospect of screen literature examines the growth of what they refer to as "visuality," which is similar to literacy, but in respect to proficiency and familiarity with visual media and being able to quickly generate visual content. The main issue they discuss is how do you become visually literate when the technology and art related to images and video is so new, while the methods involved in written word have had centuries to develop. Towards the end of the essay, Kevin defines a point that we have to reach before someone can have true "visuality" and it is the point where we would be able to search a video for a specific person, object or action in the same manor we can search a writing for specific words and passages. While this idea is great, with the way most media exists now, it comes across as wishful thinking.

Being able to identifies specific elements of images and especially video relies on two things. Either a human must manually tag the people, objects and actions in the media that are predetermined to be relevant to someone's search, which would limit the extent to which someone can reference elements of a video. This would be possible given a digital file that would be attached to the media that would store all of the tags within the video and would probably be visible to the user the same way we see tags in images on Facebook and other image based websites.

An alternative that Kevin mentions is the use of software that would automatically recognize faces, objects and so on. While this kind of software exists, it's capabilities are severely limited to being able to only recognize a small array of predefined objects that need to be programmed by the user in advance. Creating software that can have independent thought and learn based on previously recognized objects is what would be necessary to make the sort of tagging that Kevin envisions possible, but these are the same issues that robotics researchers have with attempting to create realistic artificial intelligence. So with either of these techniques we would have to wait for technology to catch up with this idea or for it to be manually done by content creators, there would need to be significant motivation for them to make their videos and images "searchable" through incentives such as money, but the reality is that most content found on the internet if free and the producers of the content have few ways to make money.

__John S. Rampersaud__ Kevin brings a great idea for programmers but the one major flaw is that we as citizens, we have privacy as Americans and for a program to be created like this would invade that precious area that we slimly have. Don’t get me wrong this is a very cool idea but once it crosses the line of privacy, his has went to far and most people would be against the idea.

__Will Gallagher @ Mike Larrabee__ Mike, I agree that it is wishful thinking to ever expect thinking AI that will actively search for pixel patterns, and it is just as unlikely that all content creators will accurately tag submissions. However, if a system in which any user could tag content were implemented, with plenty of checks and balances features it could be the answer to the question, "How do I find //any// video with this person in it?" I think it is reasonable to believe that we will see advanced tagging like this on websites like YouTube within the next five years.