[my opinion here is not necessarily that of my employer]

Tom, this is an outstanding hypothetical or rhetorical scenario, and you’ve hit a pain point that new gtlds will likely experience.

I would comment not on a specific TLD so much and more towards the oncoming thundering herd of new GTLD proposals that are on the horizon.

New top level domains are going to compete against some fairly well populated encumbent top level domains for mindshare and ultimately traction in the form of use in email and browsers.

If we look at the two rounds of ‘testbed’ TLDs that have been introduced, we’ll probably see a multitude of business models or to be more technocratic, uses of these new namespaces by the public. 

It is safe to state, from what I have understood from 32 of the new applicants that are in the wings, some will likely be introducing wildcards as part of their business model—in reasonable ways—where it might be a core component of why a person would adopt use.

NDAs restrict me from speaking in any way other than generally about how the wildcards play in to use, but the applications have some very attractive use cases which are fundamentally enabled through wildcarding, and would benefit internet users.  Without a wildcard, these benefits get stripped away, and these TLD applications get homogenized into sterile .com knockoffs.

I would venture to say that it will be all about getting quickly to the most relevant information, product, or service, and I hope that there is reasonable, dispassionate, and fair review of the new TLD applications that will be untainted by the events of 2003.

If a given TLD is adding a valuable service using a wildcard, then it might hopefully have a chance to operate with one.

What if .COM had been introduced with Sitefinder?

First of all, people like Jon Postel and Paul Mockapetris would have complained about whoever implemented it not getting the concept.  And when somebody tried to telnet or ftp or smtp to a mis-spelled .com domain name, and the DNS system gave them the IP address of a server that only implemented Port 80, their connection would have failed, just like it fails today, but they’d know the name of the person who mis-configured the DNS server and could reach him or her by phone or email - or could reach him by walking down the hall to his desk, because hopefully nobody would have provided a service like that to the world without having a friendly user try it with all of the common protocols first to see if it broke anything.

On the other hand, there was no World-Wide Web back then, much less an HTTP 1.1 that would know *what* mis-spelled domain name they had typed, and even if there had been, the blatant attempt at commercialization of misspellings would have been a violation of the Arpanet Acceptable Use Policy.  If the misguided individual who had misconfigured the DNS server had also put an email server on the Sitefinder box, it might have been configured to return “5xx   user unknown - .(JavaScript must be enabled to view this email address)” or might have been a then-normal open relay which would realize that .(JavaScript must be enabled to view this email address), queue it for delivery, look up missspellled.com in DNS, get its own address, and do something appropriate or inappropriate about loop detection… Would wildcards in .travel and .museum threaten the Internet’s stability? IP packets would continue to be delivered just fine.  One big problem is that wildcarding interferes with a popular anti-spam technique.  There probably has been some spam with SMTP envelopes or From: lines claiming to be from .(JavaScript must be enabled to view this email address) in order to evade crude spam detection techniques, though most of it was probably caught by Bayesian filters or IP-based blocklists.  If it became a popular enough technique, mail from .museum and .travel would start to be frequently blocked or high-weighted by spam filters, like mail from .biz and .info often is already, and the .museum folks would either give up on wildcarding or else do something clever like advertising strict SPF records for wildcarded DNS results. Being able to experiment with different kinds of DNS policies and technologies is really valuable, and techniques that work well under 2LDs or 3LDs may deserve TLDs of their own.  Unfortunately, the big upfront price that ICANN charges to apply for TLDs means that very few organizations can do it except for direct short-term profit-making purposes.  It does keep out the riff-raff, and keeps out the half-thought-out academic experimenters as well, but there’s a lot of potential value that hasn’t been created because of it.  Meanwhile, the managers of .COM did abuse their position to break correct DNS behaviour without even warning the community first, though ICANN finally told them to stop.