Several recent happenings have me thinking about the transient nature of the Internet. We're all used to clicking a link on some Web page or in the results of a search and getting an all too familiar "Not Found: The requested URL blahblahblah was not found on this server," when something that once was found, now is lost. What we don't anticipate is the sudden, jarring disappearance of an entire, high profile, commercial Web portal. This has happened to me twice in the last few weeks.
Two sites that I relied on, one that I used several times a week, and the other countless times each day, disappeared without so much as a puff of virtual smoke.
Back at the end of January, the Disney people let it be known that they were going to shut down their portal site, Go.com, swap common shares of Disney stock for their Web stock and lay off 400 people. Now I could give a rat's patoot for Disney and for portal sites in general, but the key element of Go.com that kept me coming back day after day was their wise purchase and incorporation, prior to their launch in late 1999, of Infoseek and its terrific Web searching abilities.
What Infoseek had going for it that none of the other kids on the block had back then was the wisdom of including search narrowing in its user interface. Simply put, when you search on "dogs" and find 3 million hits, you could click on a button to make your next search look inside of the results of the first search and so on, narrowing your search as you went. You'd get 3,275,185 for dogs, then within that 18,516 for Dachshunds, 1,672 for breeders, and 69 for Kansas, leaving you with a very useful search result.
Of course you can do this on most search engines by manipulating a complicated and arcane search syntax, but Infoseek didn't expect that from its users. Ironically, the search engine that I ditched in favor of Infoseek, AltaVista, has the closest thing to this feature now.
Amazingly, when Go.com first launched as a beta test, taking over for Infoseek, they left this feature out! Now I don't have any illusions that it was my distressed e-mail that got it restored, but happily, restored it was. Frankly, I could never understand why every search engine didn't have this feature.
One day last month, Go.com was just Go-dot-gone. It now bounces to the prosaic go.com/dynamic/Portal where one can read "about the new site" where you'll learn that search services are now provided by goto.com, a wholly unrelated contractor, who in fact sued Disney in 1999 over an early Go.com logo.
Deja vu all over again
Less personally inconvenient on a daily basis, but with much more profound ramifications is the sudden and total disappearance of Deja.com, the online Usenet newsreader and archive.
Though Usenet has been around since 1979, Deja.com, an outgrowth of the older Dejanews.com, had only archived postings back to 1995. After a couple of years of trying to hit upon a profitable business model, Deja put the word out that their archive was for sale. On Feb. 12, Google.com bought it and immediately did the unthinkable: They shut it down.
Google replaced Deja.com with an inferior user interface with access to the mere eight months of Usenet news they'd archived themselves. They explained that they're reformatting and reorganizing the Deja archive, but they could have reasonably been expected to leave the old service in place while they prepared whatever they envision and a superior replacement.
This sequence of events illustrated some disturbing facts. Usenet represents an enormous repository of collected wisdom. Of course, it also represents an even larger repository of infantile blather and unimaginative porn, but so be it. This collected wisdom is incredibly valuable as a tool and also as an enormous chunk of the Net's historical record. Relying on a single, commercial enterprise to properly steward this archive is a big mistake.
Fortunately, some people have seen the folly in such reliance. Members of the so-called "open source" community have plans for a decentralized Usenet archive, hosted perhaps by university servers and free to everyone. Two keys to this becoming a reality are Google's acceptance of the wisdom of donating the archives, if not in fact the programming source code, and overcoming the enormous technical hurdles associated with housing and serving such a large and heavily-accessed data set.
Historical society
The Deja.com debacle also points up the general lack of importance that's been placed on preserving Internet and Web history. Someday people will want to know what the state of the Web was back in '93, '99 or 2010. The time for Internet archeology is now.
As Web sites fall by the wayside, and they are properly maintained and updated, the Web equivalent of "I Love Lucy" and "Lux Radio Hour" reruns are not being kinescoped for the museum of Webcasting. Someday this will seem tragically short-sighted.
To learn more about the open source community, check out Open Source Development Network at www.osdn.com For more information about search engines check out searchengines.com. If you're interested in learning more about Usenet news, try cnet.com/internet/0-3805-7-1564164.html (these links are a bit easier to follow from this column in the news section of theMagonline.com).
No comments
Commenting is turned off for this story.