As the dust settles over last week's manic-depressive stock-market frenzy -- in which the Dow Jones took a 554-point free fall one day and then recovered most of it the next in its biggest trading-volume day ever -- the technological reckonings are piling up. We've been inundated with stories of glum investors who couldn't buy or sell fast enough via their online trading services. Another staple of the coverage has been articles marveling that, despite the record volume of 1.2 billion shares on the New York Stock Exchange and 1.3 billion on the NASDAQ, the exchanges' computer systems actually worked and kept up with the frenzied demands of traders -- most of the time, anyway.
No one seems to catch the underlying schizophrenia here. On the one hand, we're shocked when complex network technology doesn't work. ("What? You mean my cut-rate online brokerage hasn't built a system efficient enough to handle my every need during the busiest stock-market day of all time?") Yet we're also stunned when complex network technology does work. ("Wow! The New York Stock Exchange's computers don't crash the way my Pentium does!")
In truth, most reporters remain touchingly naive about the strengths and weaknesses of the networks that now run our economy. Like many of their readers, they share our cultural predilection to view computers as essentially perfect, infallible systems -- despite our daily experience to the contrary. And again like much of their audience, these correspondents tend to get flummoxed when systems don't work, and they aren't very good at fixing the blame in any particular direction.
For one example, look at this remarkably chummy interview with Charles Schwab on CNNfn, where interviewer Lou Dobbs allows the leader of the biggest online brokerage to get away with this vague explanation of why so many of his customers couldn't get through last week to make their trades: "Well, Lou, it reminds me of Mother's Day when everybody wants to call their mother, and simply there are not enough lines available to allow every person to call their mother on that given day."
Schwab remains warmly fuzzy about whether the lack of "enough lines" was simply a congestion problem on the wider Internet, which he couldn't be held responsible for, or a capacity problem at his company's own computers, for which his customers would have the right to be mad. The distributed nature of the Internet gives it much flexibility and strength; for companies in Schwab's shoes, it also means never having to say "I'm sorry" -- you can always blame glitches on the Net.
On the flip side of Schwab's bland assurances, there is the kind of tub-thumping outrage offered by MSNBC's Barton Crockett under the headline "No Excuse for Broken Brokerages": "Web traders of the world, get mad as hell and refuse to take it anymore. You don't have to put up with the traffic jams that overwhelmed some online brokerages last week. And don't let anyone tell you otherwise. Reason? Web sites can, indeed, handle big volume surges. All they need is enough computer and system capacity."
Well, sort of. Certainly, the brokerages that best served their customers last week were those that had smaller numbers of customers and extra capacity -- that's what a survey at TheStreet.com discovered. Sure, having enough servers and enough bandwidth to handle extra customers is a necessary prerequisite for any online business that wants to handle crisis-level traffic. But it doesn't guarantee good service. As any Webmaster can tell you, customers may have difficulty accessing your servers even when traffic at your site isn't unusually heavy if there's a problem somewhere else on the Net between the customer and you. On a day like Oct. 28, when Wall Street went nuts, sites all over the Internet found that their visitors were experiencing slowdowns -- even sites that had nothing to do with stocks.
The lesson of last week's stock gyrations, the financial gurus tell us, is that the financial world is now a seamless global system -- and when markets on one side of the globe, like Hong Kong, drop, markets everywhere else will be affected. What journalists and their readers need to understand about Internet technology is that it, too, is a vast, interdependent system. This gives it many fabled advantages -- like "scalability" (it's easy to expand) and "redundancy" (it has lots of backup routes when a main one fails). But this also means that it will slow down when traffic "spikes" -- guaranteed. The only brokerage that can honestly promise its customers that they will never experience the "World Wide Wait," even in a crisis, is one that relies on a private network rather than the public Internet.
GOOD READS: In the Nov. 5 Wall Street Journal, Don Clark analyzes the goof-ridden saga of the Microsoft Network in "How Microsoft lost cloak of invincibility while getting online." It's been painfully obvious for a long time that MSN had big problems, and Microsoft's many 90-degree turns in transforming its online network from proprietary mall to Internet service provider to proto-interactive TV service have been well chronicled (we took a skeptical view last March).
But the Journal piece provides embarrassing new details about Microsoft's "humbling." Remember those ugly snafus in the company's billing system? It seems that "billing software was the responsibility of a single programmer who had no experience in the field."
Maybe this juggernaut isn't so terrifying -- but no one is counting Bill Gates out yet. Microsoft is notorious for sticking with bad products through one costly upgrade after another until, finally, everyone's using them. Windows was once a disaster area, too.
Shares