Here's how the TLDs look so far:
.info had, and still has, a lot of potential, but so far hasn't seen that much active use. watergate.info, a comprehensive site about the Watergate scandal, is a rare exception; I think it should be looked at as the archetype of what a .info site can be. mta.info may be the most visibly promoted .info site, being advertised on signs in Grand Central Station. My own tatiana.info site is reasonably popular among fans of Tatiana, the Latin-American singer and TV host the site is about (she preferred the address herself over the chicadehoy.org address at which the site was first launched). But otherwise, you don't see much .info activity. The tying up of many good names taken by fraudulent sunrise trademarks pending Landrush 2 is one of the problems, and the permanent tie-up of some names with good generic potential by legitimate trademark holders (e.g., cat.info by the Caterpillar tractor company) is another.
.biz doesn't seem to have made much of a splash yet... I can't recall encountering any significant sites there, though a few must exist somewhere.
.name is perhaps the biggest disappointment so far. It has the biggest potential target market, all 6 billion individuals on Earth, but has much fewer registrations than .info or .biz. A lack of marketing to the target audience which is much less Internet-savvy than other domain registrants is the primary likely culprit, but problems with the registry figure in as well; even months after the launch, they're still only processing registrations and updates in batches every week or two, so the process of getting a .name site online can be an aggravating process even for Internet veterans, let alone newbies. Just getting the name servers updated is a big headache; you submit the update, wait two weeks to see if it makes it, then if it's lost or bungled you submit it again and wait another two weeks. And the whole business of having separate registrations for e-mail addresses of the form firstname.lastname@example.org, with a separate payment and a separate wait of up to two weeks for their activation, is another source of confusion and aggravation. My own attempt to activate email service for email@example.com doesn't appear to be effective yet, and I see no way in the .name WHOIS to check on its status. These problems really need to be addressed before .name can become really viable as a popular TLD for people who actually want to use it, as opposed to the speculative and protective registrations which probably amount to a large part of the current registrations despite the general lack of interest in .name by speculators -- enough people said "Hey, cool, I can cybersquat on britney.spears.name!", or "We must Protect Our Trademark by stopping anyone from getting a .name with "Amazon" in it anywhere, even if they are a female warrior or live on a South American river!" to likely outnumber the people who actually registered their own names in order to use them.
.coop, .aero, and .museum have had quiet, uneventful launches, with some minor criticisms (very small in comparison with the storm of protest over .info and .biz policies), but no big hassles. Unfortunately, there hasn't been very much use or promotion of these names either. Lots of famous airlines, museums, and co-ops have yet to see any need to register domains in their appropriate TLD, and those who did aren't really using them -- if they resolve to a website at all it's just the one they already have, and promote, under a different URL. Although .aero offers the logical structure of letting airlines and airports register under their reasonably well-known two and three letter codes (JFK, LAX, AA, DL, etc.), allowing a consistency in addressing that would be very helpful to frequent travelers if it were widely practiced, few have even bothered to sign up for those.
.pro has a signed agreement with ICANN now, but hasn't entered the root; I'm not sure what the holdup is at this point.
So, what can be done now to reverse the fortunes of the new TLDs? I propose that the domain industry should form a trade association that can launch a unified marketing campaign to encourage awareness and use of top level domains other than .com. The major funders of this association and its campaign could obviously be the .info, .biz, and .name registries, who have the most to gain by increased interest in "alternative" TLDs. The .pro registry can jump in once it launches. The noncommercial registries (.museum, .aero, .coop) should be invited in as members to provide moral support and greater diversity of represented sites, though they wouldn't be likely to kick in much money. The recently-spun-off .edu consortium might also become part of the campaign, as could the .org registry once it's separated from Verisign. .gov, .mil, and .int might be made honorary members to complete a sweep of the "non-dot-com" gTLDs.
This new association would then come up with clever slogans and ads to promote the concept of different top level domains, such as "Not just another dot-com!" They could highlight various sites, both commercial and noncommercial, that are useful or interesting and use addresses that don't end in .com. Creative references could be made to the silliness, empty hype, and big collapse of the "Dot-Coms" of the recent Internet boom era, and promote an impression that in .com sites the user will largely find silly marketing gimmicks, while the other TLDs give you a good idea of what type of site you're going to (governmental, informational, museum, etc.) and often have good content that's not just the Over-Hyped-Gimmick-Of-The-Week.
While each of the TLDs might not have the marketing budget to make much of a splash by themselves, a combined campaign might be able to do more. And others with a stake in the success of "alternative" TLDs, including site owners using those addresses and speculators hoarding them, might be willing to contribute to the marketing fund to make the campaign even bigger.
The campaign could culminate in an award ceremony for the "Non-Dot-Com Sites of the Year", where they give prizes in various categories.
Does anybody see any flaws in this?