The ECNs contend that an array of special interests profit at the expense of investors in even the most mundane exchange-directed trades. Machine-based systems, they argue, are much more efficient, because they speed up the execution mechanism and eliminate the need to deal with an intermediary.

Historically, the 'market' (which, as noted, encompasses the totality of stock trading on all exchanges) has been slow to respond to technological innovation. Conversion to all-electronic trading could erode/eliminate the trading profits of floor specialists and the NYSE's "upstairs traders."

William Lupien, founder of the Instinet trading system and the OptiMark system, has been quoted as saying "I'd definitely say the ECNs are winning... Things happen awfully fast once you reach the tipping point. We're now at the tipping point."

Congress mandated the establishment of a national market system of multiple exchanges in 1975. Since then, ECNs have been developing rapidly.

One example of improved efficiency of ECNs is the prevention of front running, by which manual Wall Street traders use knowledge of a customer's incoming order to place their own orders so as to benefit from the perceived change to market direction that the introduction of a large order will cause. By executing large trades at lightning speed without manual intervention, ECNs make impossible this illegal practice, for which several NYSE floor brokers were investigated and severely fined in recent years. Under the specialist system, when the market sees a large trade in a name, other buyers are immediately able to look to see how big the trader is in the name, and make inferences about why s/he is selling or buying. All traders who are quick enough are able to use that information to anticipate price movements.

ECNs have changed ordinary stock transaction processing (like brokerage services before them) into a commodity-type business. ECNs could regulate the fairness of initial public offerings (IPOs), oversee Hambrecht's OpenIPO process, or measure the effectiveness of securities research and use transaction fees to subsidize small- and mid-cap research efforts.

Some, however, believe the answer will be some combination of the best of technology and "upstairs trading" — in other words, a hybrid model.

Trading 25,000 shares of Lucent stock (recentquote: $2.80; recent volume: 49,069,700) would be a relatively simple e-commerce transaction; trading 100 shares of Berkshire Hathaway Class A stock (recent quote: $88,710.00; recent volume: 450) may never be. The choice of system should be clear (but always that of the trader), based on the characteristics of the security to be traded.

Even with ECNs forming an important part of a national market system, opportunities presumably remain to profit from the spread between the bid and offer price. That is especially true for investment managers that direct huge trading volume, and own a stake in an ECN or specialist firm. For example, in its individual stock-brokerage accounts, "Fidelity Investments runs 29% of its undesignated orders in NYSE-listed stocks, and 37% of its undesignated market orders through the Boston Stock Exchange, where an affiliate controls a specialist post."