The credit-rating agencies have come under fire for being asleep at the wheel as the credit quality of mortgages deteriorated while the Wall Street securitization machine roared on.

As Roger Lowenstein detailed recently in the New York Times Magazine, Moody's and Standard & Poor's "are a central culprit in the mortgage bust, in which the total loss has been projected at $250 billion and possibly much more."

To the accusations of willful blindness in the pursuit of profits, a new culprit for the blunders of ratings agencies can be added: computer error.

Sam Jones, Gillian Tett, and Paul Davies of the Financial Times today deliver a stunning report on how a mistake in coding at Moody's led to triple-A ratings being incorrectly assigned to billions of dollars' worth of a type of complex debt product.

Moody's discovered the glitch early in 2007, the paper reports, but the products—constant proportion debt obligations—remained triple A until January 2008.

Felix Salmon calls it a "scandal." But The Wall Street Journal, strangely, in following the Financial Times, seems to play down the story nearly to the point of pooh-poohing it.

Yves Smith on the Naked Capitalism blog says, "while banks have rogue traders, it appears rating agencies have rogue computer models."

Tanta on the Calculated Risk blog is disturbed by the implication in the Financial Times report that Moody's may have tweaked its computer model to arrive at the same ratings as Standard & Poor's in order to keep business as a "second opinion."

Moody's says it is "conducting a thorough review" of the rating of the derivatives.