But, even if you disable search, disable history, there's the fundamental fact that _anyone_ can record everything they see, easily and silently. You can't just have a private authenticated space, you need to be able to personally trust every single person you let in that space.
At that point, the features around archiving or search are a bit moot.
- We shouldn't bother checking for characteristics of credit card fraud at transaction time, because a determined attacker might get a fraudulent card through.
- We shouldn't bother checking IDs at bars, because a determined attacker might get a fake ID through.
- We shouldn't bother trying to prevent email spam, because a determined attacker might get a spam message through.
- We shouldn't bother making laws against recording people without their consent, because a determined attacker might do so anyways.
Please construct a more plausible argument than "it's ultimately hopeless". I'm willing to consider alternatives, but I'm not willing to consider fatalism.
I don't argue that "it's ultimately hopeless". I think, with effort, it's quite possible to create a pseudo-private safe space online.
I just think that the bulk of your benefit comes from vetting the intentions and judgement of your participants, and not from technological means like removing search.
If you're able to trust your participants, technological means can be like a simple fence/lock, keeping honest people honest, but don't otherwise add a ton of extra benefit.
If you're not able to trust your participants, removing search will help _some_, but it might not be enough.
If it was just "determined attacker can do a search, and use the results privately" it wouldn't be a big deal. But they can then spread the result to everyone else in the world in a way that almost negates the barriers.
The recording option is closer, but search can be done retroactively, which makes a huge difference.
Under US law anyways, it is absolutely not okay to republish a written conversation without permission under copyright law to do so, unless you fall under the "fair use" exemptions. It is safe to say that most forms of attack would not pass a "fair use" test in the courts. So if more people were to sue over copyright violations and demand the unmasking of who posted and/or those who operate, that would certainly go a long ways towards resolving that imbalance.
Sadly, it's a very expensive process, and it's very cheap to violate copyright.
If you permit full text search of content, you permit abusers to discover and harass your members, because they can just sign up to find out who posted something once full text search discovers it.
You're implying they're somehow inversely related. A site can be well represented even when it's users contents are not on display
If you want to regain the pseudo-privacy of physical spaces online, put your community’s conversations behind an authentication barrier and disable full text search of conversations.
Search engine indexing is what turns a pseudo-private space into a humiliating-public one.
It’s okay to let search engines index your forum’s existence, the sub forums it contains, and their descriptions. But do not let them index participants or conversations - either by subject, by participants, or by content. And do not offer full text search of post content to authenticated members. It’s okay to index keyword tags, but that’s it.
If you do this, you will regain the semi-anonymity that made the early Internet possible to enjoy. If you don’t, you will continue to suffer the trolls and abuse that full-text search enabled in the mid-90s (see also DejaNews, X-No-Archive: Yes, and Google’s purchase of DejaNews).
EDIT: If you truly feel that full-text search is so valuable that it must not be withheld, you have to do a lot of things to defend against abuse attackers - for example: charge money for search credits, deduct credits when they choose to reveal the text of results, warn users that their searches will be monitored for abuse, require users to be in good standing with paid membership and posting activity for at least 90 days, etc. Otherwise trolls will just use stolen cards to perform full content searches to identify users to harass and then report their findings back to a central forum. They may still do that after all the above criteria, but they’ll have to work excruciatingly hard at it. Yeah, they could manually scrape the site, but you can defend against that too (“you’ve participated on 12 days, so you’re allowed to view 12 days of old content” is a good simple test).