Dear Team,
I'm wide open to other ideas for the support of robotic vision through tools already built into PostGreSQL. But you've already admitted to certain speed limitations...and robotic vision is going to require much more intense processing power. An MVD might allow the data stream to be quickly "re-arranged" during the recognition processing and then converted to RDBMS form later...after object recognition is "stabilized"...
I really would love to see little "Short Circuits" running around teaching themselves...ok? Just a dream of mine...
