Dr. Dobb's Journal, March 2006
(Grudging...) And other crafts have come up...if not to replace, then...occupy.
—Edward Albee, Box
In this art or craft of software development, thinking outside the box is a constant challenge, because it seems like it's just one box after another.
Comic-book writer Chris Claremont once said that when he was creating a new superhero, he would always ask himself, "Is there any reason this character can't be a woman?" It was a conscious effort to challenge the ubiquitous unconscious default, and it changed the course of superhero comic-book history. A small thing, but his own.
It seems to me that there is a similar question a developer might ask at the beginning of every software project: "Is there any reason this project can't be open source?" The question might be worth asking, even if you know in advance that your company's business model and your unwillingness to put your job on the line add up to an immediate "you're darned right there is."
I must sojourn once to the ballot-box before I die. I hear the ballot-box is a beautiful glass globe, so you can see all the votes as they go in.
—Sojourner Truth, letter to the World
I'm thinking of one particular software task for which economic or political forces seem to be producing a "you're darned right there is" answer. Unfortunately, it's a task that—it seems to me—absolutely must not be tackled with proprietary, black-box solutions. I mean the task of writing software for electronic voting machines and for tallying votes in elections.
The 2004 general election in the United States was a wake-up call, because many disturbing questions were raised—and remain unanswered—regarding the accuracy of machine voting results and the accountability of the process. But when off-year elections come around this November, the reliability of electronic voting probably won't be much better than in 2004, according to an October 2005 report from the General Accounting Office. The GAO cited lack of consensus on adequate certification processes. What's needed, according to the Electronic Frontiers Foundation, is clear: A verifiable paper audit trail and open-source software. EFF.org and BlackBoxVoting.org both have good information on the requirements.
I realize that there are other issues here beside the technical question of how to construct a reliable and accurate electoral system. There is the problem that not every player in the game necessarily shares this goal of reliability and accuracy.
Tracking this issue on a daily business, I am stunned to read about voting software written by convicted felons; officers of voting-machine companies getting involved in political campaigns even after the company officially prohibits such activity; State and local policies on voting that make it less fair, less representative, less trustworthy being put in place by questionable law or in contravention of the law; election results that defy probability; vote counts in which actual votes exceed the number of registered voters by orders of magnitude...But that's not the point I'm trying to make here. If people are breaking the law and trying to subvert the electoral process, they should be exposed and thwarted. But I'm trying to focus on the straightforward technological issue of how to ensure accurate and trustworthy electronic vote counting, something that every American who cares about democracy wants. And the GAO's gloomy prediction is just not acceptable. We need to solve the problem in time for this fall's elections.
When I retire, I'm going to spend my evenings by the fireplace going through those boxes. There are things in there that ought to be burned.
—Richard Milhous Nixon, Parade
Audit trails are problematic with electronic data. It's so easy to change the bits. On the other hand, you do have to know where all the bits are that you need to change. Say you've done something that you don't want some Federal prosecutor with subpoena power to find out about. Could you remove all traces of incriminating documents from your hard drive(s) in the hour that you might, if you're lucky, have before it's impounded by Federal marshals, and do it without leaving any sign of tampering with evidence? What about that damaging e-mail? More problematic, because removing it will likely involve both your computer and your correspondent's. Oops, and what about your ISP and his?
The bigger the slug, the bigger the trail of slime. If you're laundering billions of dollars of money through bogus defense contractors and nonexistent software companies, you'll leave a lot of electronic traces. Think about all those web sites you were so proud of, with their professional graphics and impressive pictures of the purported corporate headquarters, those dummy sites whose links all go to ground on dead or nonexistent pages. What could an investigator learn from them?
In December, I was following what looked like an important story about one such company. I'm no cyberdetective, but it only took me a few minutes to convince myself that the company in question was, uh, questionable. Faked-up web sites, references to company-produced software products whose names returned empty search results, no way to get to the identities of the company officers from the main web site. Of course, I ran a "whois." The domain name's registrant had a Network Solutions-listed address that lacked a street name. Not a lot of help there. The phone number looked real enough, but the fax number obviously wasn't.
I don't mention the name of the company because the story could still be bogus, and my research was far short of investigative journalism. I bring it up just to make the point that it's not that easy for the crooks to cover their tracks now that we all pretty much live in cyberspace.
Any man forgets his number spends a night in the box.
—Donn Pearce, Cool Hand Luke
Which suggests two observations and two opinions.
The first observation is that privacy seems to be more and more illusory. That cell phone call you just made pinpoints your location to within a few hundred feet. The credit-card purchase narrows it to right here. You send e-mails and place cell phone calls and offer up your credit-card numbers and passwords and rent DVDs and check out library books with insouciant indifference to what you are revealing to the prying eye. Yet, even as the very notion of privacy seems to dissipate, some seek to discern a right to privacy in the U.S. Constitution.
Another observation: Bloggers and other amateur newshounds are challenging professional journalists, often doing a better job of finding and disseminating important news stories. Boxed in by their success at gaining access to the powerful, celebrity journalists seem to have fallen victim to Stockholm Syndrome, like kidnap victims defending their captors. Professional outsiders like Josh Marshall and bloggers who spread the word even if they don't break any stories do something that the mainstream media stars can't. Or don't. Bob Woodward, Judith Miller, and Viveca Novak all seem to have made the fatal error of forgetting to put truth and the reader first. Meanwhile, all the big newspapers are in financial trouble, challenged as much by Craig's List as by bloggers.
The decline of privacy and the rise of citizen journalism are huge changes, but I offer these meliorative opinions: It seems to me that the fact that privacy is becoming harder to maintain should have nothing to do with the constraints placed on government interference in our private lives. And it seems to me that citizen journalists and professional news analysts can perform complementary tasks if they focus on their respective strengths. I just hope these things sort themselves out soon.
One thing a computer can do that most humans can't is be sealed up in a cardboard box and sit in a warehouse.
—Jack Handey, Deep Thoughts
Whether Microsoft can sort out its dilemma is an open question. And whether it's an open-source question is another question.
Steve Lohr of The New York Times seems to think that Ray Ozzie has been tasked with bringing Microsoft into the era of software as services. It's possible that Ray's role is more limited, being merely to implement the already-articulated and well-under-way dot-net technological plan, while Bill and Steve work on the harder job of reinventing Microsoft to be the kind of company implied by this technological vision.
Either way, Ray has his work cut out for him. Imagine that you have been hired as the CTO for a company that produces huge, feature-laden apps and you are tasked with shifting the focus to producing tools to allow others to produce small, focused apps that compete with and probably better serve the changing market than the company's fat apps. A tricky message to give to the troops.
Ray's exhortation to the troops can be Googled by anyone interested. In the document, he admits that he isn't saying much that Bill hasn't already said, except for Ray's kind words about the open-source paradigm.
For all Ray Ozzie's virtues, he can't ever have the kind of hearts-and-minds influence at Microsoft that Bill Gates has. But he has as much formal authority as Bill and Steve give him, and that's shaping up to be a lot. If he's charged with more than the mere technological challenges in shifting Microsoft's direction, he'll need it. Because the software-as-services vision will involve, for example, not just educating the sales force about new products, but more or less rewiring their brains. A company the size of Microsoft trying to change directions has all the maneuverability of a jumbo jet on the runway.
Here's a thought, Ray. Nobody in technology has survived more disasters or reinvented himself as often as AOL founder Steve Case; He's the real comeback king. And you know what Steve is proposing for AOL/Time Warner? He thinks that the best thing to do would be to break it up.
Calvin has an obsession with corrugated cardboard boxes, which he adapts for many different uses.
—Wikipedia, on Calvin and Hobbes
DDJ