Is the continued vitality of technical stock analysis an affront to the efficient market hypothesis?

Technical stock analysis (predicting future stock prices on the basis of past price movement) remains alive and well.  Just today, a Wall Street Journal headline referenced the "Hindenburg Omen"--a confluence of stock price data points that portend bad news ahead.  Price movement data is indeed an aspect of the efficient market hypothesis, since the weak version of the theory rests on that data being incorporated into market prices.  But we generally suppose our most liquid markets satisfy the semi-strong version of the theory.  That is, those market prices should reflect all publicly available information.  One may then ask whether past price movements should have any relevance to rational investors investing in such a market.  I suppose an argument could be made that price and volume movement might somehow reflect psychological trends of market participants--the "mood" of the market, if you will.  Nonetheless, it's hard for me not to think "superstition" when technical analysis comes up.

SJP