CircleID

It seems CNET has disabled it's domain name search feature.  The link is broken, and clicking on the Web hosting directories and tool: "Domain name search" that's currently here redirects to cnet's home page as well; other links on their site to this feature don't work either.  The actions of a victim or perpetrator?

Doing a metasearch through a service that then queries numerous other parties in order to determine whether a name is taken from a unique source seems like an odd way to go about name availability searching.  By definition, it is a query that doesn't require a metasearch from several sources.

Having tried the search with CNET for a fifteen-letter sequence that could in no means be a random coincidence, it appears that it too was registered two days later.

It's odd that people would think "things I type into some webpage on the internet" is held as confidential information by whomever might be on the receiving end of that information, but the moral of the story appears to be that if you want to know if a domain name is available, then check directly against the registry instead of performing a query that is several steps removed.

One wonders whether even Chesterton knows the source of the data it is receiving from any of several parties in the process.

Unfortunately many internet users aren't "savvy" enough to know exactly how this works. But that's where education comes in. :)

Let's do our part, guys. And do what we can.

I don't know how "savvy" one has to be about not communicating a purchase interest prior to being prepared to make a purchase.

For example, anyone can set up a webpage that says, "Thinking about purchasing stock?  Type in the ticker symbol here, and I'll tell you the current price."

Ummm… hey… there are web pages that do that as a "free service". 

Maybe I spend too much time at the track with other gregarious horseplayers who will be happy to chat about anything EXCEPT how they are betting in the next race....

Lemme know if you see anything you like on eBay, Dave.

As a domain name registrar (Namesbeyond.com), we have taken an approach that any request to determine availability of a domain name on our systems and Whois servers are "left alone" - i.e., no data mining and domain tasting.

Should be part of a registrar code of conduct.  In this case, the search is going through intermediaries and not registrars directly.

Should be part of a registrar code of conduct.

Agreed.

This can still happen even if a registrar believes it follows a code of conduct, though.  Some registrars outsource their 'name spinning' function (whereby a prospective registrant is provided with a number of additional variant names based on the input target name).  Those registrars might not even be aware that the name spinner is using the data.

Part of the difficulty for an end user is identifying a registrar, as opposed to a reseller, a webhosting company which will go through a registrar on behalf of its customers, or, in this instance a news site.  The search system used in the experiment, and which I also used to duplicate the result, seems like a brain-damaged approach to checking domain availability in the first instance (as a meta-search to determine a binary answer relative to a unique piece of information available from a single source makes about zero sense).

One of the other inefficient ways that people check domain availability is to type www.. and see if there is a web page there.  In that instance, the search defaults to say, msn.com, Google, or whatever default search is configured via a browser plug-in, so a third-party still has access to the data.  (Of course, if the ISP is running OpenDNS, then EVERY domain will look 'taken', and OpenDNS will know whether any unregistered domain names get an appreciable amount of traffic...)