VeriSign fends off critics at ICANN confab (News.com)

I stumbled across this article on Business Week's web site. The sub-head reads:

The giant recorder of Web addresses sure sparked a fire when it redirected site-not-found messages to its own search engine. It's a bum rap

Their argument, I suspect, is ultimately based on the idea that ``On Sept. 15, Verisign unveiled a way for it to make money from these mistakes.'', even though they do offer other justifications. Their main one is the claim that it's OK for Verisign to capture mis-directed addresses because that's exactly what owners of one-letter-off domains are doing. They do note that the change broke ping and traceroute, but the last paragraph (page 2 of the article) says it's just a matter of consistency: if domain owners can do it, why not Verisign?

Because, first, that way lies a new form of spam, say I. The real cause of spam is that there's effectively no cost to send it. The only reason your metal mailbox contains only the amount of junk mail it does is that postal services the world over charge real money to send the stuff, even though it's less than what Jane Citizen has to pay to send a letter.

Verisign has, effectively, bought every unused domain for nothing, and are now spamming Internet users of all stripes because it's worth their while to do so. The off-by-one types have vastly higher costs, even if it's only $2/yr/domain, or whatever the current bottom-line registrar's cost is. Verisign gets off-by-one, off-by-two, ..., off-by-N domains for free.

The second, and vastly more important, reason is that one of the should-be stewards of the Internet has broken the basic principle on which the 'net has blossomed: dumb network, smart edges [1]. As examples, ping and traceroute might be written to handle "always find some domain" design, but they'd be more complex and much more subject to breakage as the misguided DNS-root-service provider thought up more ways to use and abuse the concept.

While some situations do call for a smarter network [2], we'll always need a dumb one, to show what can be done when the endpoints think up something new. That's what Verisign broke.

Best wishes,

Max Hyre