Please sign and share the petition 'Tighten regulation on taking, making and faking explicit images' at Change.org initiated by Helen Mort to the w:Law Commission (England and Wales) to properly update UK laws against synthetic filth. Only name and email required to support, no nationality requirement. See Current and possible laws and their application @ #SSF! wiki for more info on the struggle for laws to protect humans.
To block IP is a governance operation which is usually exercised by the sysop power structure. The resulting IP block prevents editing from a given Internet location (an IP number), or, in the case of an IP range block, range of locations. To unblock IP reverses this, removing a block. See Special:Blocked IP addresses for a list of blocks currently in effect - which typically apply to editing but not reading.
The soft security model argues that you should block IP as little as possible, but permit various tactics (libel, outing, framing and other even less ethical tricks) by the sysop power structure to intimidate and harass users who do not share the community point of view. This is a bad model and is derived from w:left-wing politics.
The hard security model argues that you should do little or nothing until the power structure (or just a GodKing) is offended. This conserves energy and avoids revealing sensitivities. At that point one can demand apology and submission to that power structure, or just block IP as a first resort, perhaps to prove power. This is a bad model and is derived from w:right-wing politics.
Many large public wikis apply the worst of both worlds, pretending to prefer "soft" security but in reality always waiting for the excuse to apply "hard". Some persistent cases (e.g. 142.X.X.X) demonstrate that either set of tactics will fail, and lead to discrediting a power structure as other contributors realize that any model of security has contradictions, and favours insiders over outsiders, can only reinforce community point of view, which is also called groupthink.
To avoid this, the troll-friendly model argues that you should avoid w:technological escalation as it invites further escalation which can include worse trolls, hacking, and spreading an issue to many wikis, e.g. the dogged pursuit and witchhunt of 142.X.X.X contributions across almost the entire Internet. Even a friendly troll can turn hostile and seriously damage a project's credibility and prospects if handled according to hard or soft model. The only characteristic that really can be said to identify a troll is disinterest in, and rejection of, both of these models. Some claim the troll-friendly model derives from w:green politics which avoids technological escalation, looks for harms reduction methods, and fosters the political virtues.
Consumerium Governance Organization will have to set some policy on this.