The basic problem concerns application hosting services (AHS). In contrast to a webhost, which provides a virtual host machine on the Internet which the client configures with the desired applications, an AHS provides the application directly. Typepad is an example which provides the Movable Type application, or Smugmug which does photo galleries. Such services are proliferating because, for clients who really only want the provide application, they are cheaper and enormously easier to set up. From a junk point of view, the key features are that such services generally provide the ability to put up user generated HTML content in at a domain address for free.
While I agree with PT that being netsearch engine friendly and trusted are desirable characteristics of AHS targeted for colonization by junkers, I think that is secondary to them providing a source of free domains. I certainly see a lot of junk from obscure hosting services (such as u-blog.net and phlog.net). It is also the case, as I noted, that hosting services that don’t protect themselves against this kind of colonization will ostracized from the larger community1. From a junker point of view, even if a particular AHS becomes such a pariah that is is useless, there are always more of them waiting to be used and discarded at no monetary cost to the junkers.
What are the likely responses to this? One I have seen in a few places but I expect to become much more common is requiring a valid credit card even for free services. This service as a form of confirmed identity. It is not that one can’t fake that, but it would require engaging in a much more serious criminal act that is easier to prosecute and it would be trivial to detect the re-use of the same identity for large numbers of service instances.
I don’t think we’ll see much of a move toward invite only communities, as those don’t do very well unless the sponsor is already well known (e.g., gmail and Google). Not that such things won’t exist, but they won’t be a significant factor in the overall AHS market. Or, alternatively, the meaning of “invite” will be diluted into meaninglessness just like invitations to get a credit card2.
Internal policing will be important, but I am not sure how effective that will be given the scope of the problem. Absent some sort of gatekeeper software, no economically feasible staffing level could keep up with junker softbots registering new service instances.
Weblog defenses will continue to improve, although at this point my defenses are handling new variants with very little tweaking (I did end up banning the .fr domain). But this is limited as well, because of the fact that the point is to allow anonymous and unexpected communication.
Ultimately, the solution will be expensive (in some way — money, time, effort, latency) identities. As noted, credit cards can serve as a proxy for that today, modulo the problem of potential clients being unwilling to provide the information3. Any such identity system doesn’t have to be perfect, it just has to raise the bar high enough to make junking infeasible.
1 And even that might not be enough, as there’s no reason that the colonized host can’t also be target for junk from the same hosting service.
2 One does wonder if the junk market is sufficiently lucrative that, should credit card numbers become an important identity system for AHS, it would be worth it for a junker to sign up for every card he could, as that would not involve fraud.
3 This will be less of a burden as more people shop online and become used to providing that information to obtain online services.