Lack of interest Proxy Detection

This suggestion has been closed automatically because it did not receive enough votes over an extended period of time. If you wish to see this, please search for an open suggestion and, if you don't find any, post a new one.

James

Well-known member
I think proxy detection should be implemented (even if it doesn't do anything) just so we know which users are browsing via a proxy.

The proxy could display an icon in the user's postbit/profile/membercard that shows that the user is browsing via a proxy.

It becomes difficult and tiresome when you have to ban users who create new accounts on your forum to evade a ban and I feel detecting proxies would help prevent this. Forums that do not allow proxies can ban these users whereas forums who do allow them can just leave them be.

Most people will probably sit here thinking "how can we detect proxies?" - Well you can use HTTP_X_FORWARDED_FOR and check if the referring URL contains "proxy" (or specify a list of proxies from a collaborative list - similar to IPB's spam DB)

*Like this if you agree - new prioritising method*
 
Upvote 3
This suggestion has been closed. Votes are no longer accepted.
I think proxy detection should be implemented (even if it doesn't do anything) just so we know which users are browsing via a proxy.

The proxy could display an icon in the user's postbit/profile/membercard that shows that the user is browsing via a proxy.

It becomes difficult and tiresome when you have to ban users who create new accounts on your forum to evade a ban and I feel detecting proxies would help prevent this. Forums that do not allow proxies can ban these users whereas forums who do allow them can just leave them be.

Most people will probably sit here thinking "how can we detect proxies?" - Well you can use HTTP_X_FORWARDED_FOR and check if the referring URL contains "proxy" (or specify a list of proxies from a collaborative list - similar to IPB's spam DB)

*Like this if you agree - new prioritising method*
I think a feature shouldn't be implemented if it doesn't do anything. The only "cosmetic" feature [that's a feature that has no function] should be design implementations. Also this has the whole you are guilty until proven innocent thing on it.
 
It doesn't at all. You are not guilty because you browse via a proxy. I use a proxy on some sites, I'm guilty of nothing by using this. As I said, forums that want to ban people for it can do so, forums that don't can ignore it.
 
There are perfectly legitimate reasons to use proxies, thus I see no real benefit in labeling such users. Also, if I were going to detect and ban certain proxies which are prone to abuse, I wouldn't do so from the application itself (i.e., XenForo) but rather (for sakes of an example) through a reverse proxy utilizing an rbl of some sort (though there are plenty of other ways (some arguably better) of doing this as well).
 
There are perfectly legitimate reasons to use proxies, thus I see no real benefit in labeling such users. Also, if I were going to detect and ban certain proxies which are prone to abuse, I wouldn't do so from the application itself (i.e., XenForo) but rather (for sakes of an example) through a reverse proxy utilizing an rbl of some sort (though there are plenty of other ways (some arguably better) of doing this as well).
Indeed there are legitimate reasons, nobody is doubting that at all. I just feel that a simple method to detect proxies would help forum owners, especially in cases of ban evasion.
Yes, IPs can be changed easily by your ISP and yes it can be useless to detect proxies in some cases. I'm not going to be ignorant and suggest that this is an admincp option because there's really no point in having such a small feature as an option, but if it isn't integrated I'd probably create it as a modification.

I didn't suggest a more complicated feature as you did because it's not worth all of that code just to detect a proxy server, I feel if someone is going to be simple enough to use an easily detected proxy then they should be prevented by an easily created script
 
I think proxy detection should be implemented (even if it doesn't do anything) just so we know which users are browsing via a proxy.
How do you plan to detect Tor? Military use it on a regular basis, as well ordinary people. A real hacker can crack into computers in Korea or Brazil and use them to launch abusive activities against your site.
Or even easier, he/she can use spyware, viruses, and other techniques to take control of literally millions of Windows machines around the world, without the need of any proxy. The technique you refer was popular in 1990, not 2010. :)
 
How do you plan to detect Tor? Military use it on a regular basis, as well ordinary people. A real hacker can crack into computers in Korea or Brazil and use them to launch abusive activities against your site.
Or even easier, he/she can use spyware, viruses, and other techniques to take control of literally millions of Windows machines around the world, without the need of any proxy. The technique you refer was popular in 1990, not 2010. :)

I use Tor myself, at times, and I definitely had it in mind when I replied earlier, which is why I don't see the need in visually distinguishing such users within the UI (as originally suggested). A user's activity on the site should be what draws your attention to them, not where they happen to be connecting from.

That said, it's quite easy to detect Tor exit nodes by using a service such as TorDNSEL (amongst others, which is commonly done on IRC networks where anonymous access is more widely abused). It won't catch all nodes, but does a fairly good job.
 
Try to do a dig on a ip-port.exitlist.torproject.org address. It does not return anymore a 127.0.0.2 response. That itself kills any possibility to detect the actual usage of a Tor proxy. Unless you block all torproject.org addresses and probably piss off some badass who will DDoS your site in no time. :)

Personally, I really don't care if people use proxies to visit my site. Very little of them have bad intentions, they are mostly visiting from countries where freedom of speech is prohibited. A client told me that he is having problems with users and some of them were banned. They kept on coming back, registering new names and posting furious remarks, flames etc. When I looked at his site, first thing I noticed was how arrogant was his staff, while the forums were ran in hitlerian style. Obviously, those banned users were furious for being kicked out without any explanation (don't we all know some company that does the same thing).

Anyways, a real suggestion would be to create a system that will allow users to become "invisible" to other users. In this way, a moderator can try to talk to the recalcitrant person, then make him/her invisible, if a solution cannot be found. The user has no idea, his posts are hidden, since they are visible to guests and himself. Again, he could create several accounts... the story never ends.
 
I merely used TorDNSEL as an example, but there are much better lists for detecting exit nodes. We utilize a few such lists (with a great degree of accuracy) on an IRC network I'm involved with to "mark" such users, so channel operators can easily ban all Tor connections, since such access is widely abused on IRC. Tor is not designed to be undetectable, but provide anonymous access, so it's not so hard to detect them yourself. But any method is unlikely to detect them all. I'm with you, in that I don't care if people visit my site using proxies.

Anyway, to get back on topic (somewhat), I agree, a better suggestion would be like the one you mentioned. But, indeed, nothing will be full proof.
 
Nothing is completely foolproof and I don't think proxies are such a major issue that we have to create a large script to check for proxies. Hence why I suggested a small feature :)
 
Anyways, a real suggestion would be to create a system that will allow users to become "invisible" to other users. In this way, a moderator can try to talk to the recalcitrant person, then make him/her invisible, if a solution cannot be found. The user has no idea, his posts are hidden, since they are visible to guests and himself. Again, he could create several accounts... the story never ends.
A Tachy goes Coventry?
 
This method wouldn't detect all proxies anyway. Some proxy servers don't sent the HTTP Forwarded For header. :)
 
What about less sophisticated web proxies which use Glype? Are there plans to have a solution which can detect web proxies usage?
 
What about less sophisticated web proxies which use Glype? Are there plans to have a solution which can detect web proxies usage?
Do they send any kind of identifiable header (or user-agent, in the case of web proxies)? If not, no, there's really nothing you can do without checking against some kind of massive proxy database, which really isn't worth the time to compile, keep updated, or check against. :)
 
What's that?
It is a global ignore. A person in Tachy goes coventry can make posts, and they will be viewable by that person while logged in - but no one else can see those posts.

There are a few problems with that system, such as when they log out, not being able to see their own posts. Even if they do not log out and notice it, if they are ever taken out of tachy goes coventry, all the posts suddenly show up, and then people respond. It is a bit awkward and I've never used it.
 
I was thinking this is only a few lines of code so it might be good to integrate it into the core. Of course, it wouldn't be foolproof, but it detects the simpler proxies.
 
Top Bottom