Re: Blacklisting methodology

We actually discussed this on IRC earlier today and these same problems 
came up.   In light of that, we decided to  cancel  the banning task and 
instead enhance the approval process in JLCP.  This will be much easier 
and not such a pain for everyone.

-Matt

Geoffrey Knauth wrote:

> On Thursday, May 8, 2003, at 16:21 US/Eastern, Matthew Schmidt wrote:
>
>> Hey guys.  I'm working on banning people from JLCP via the admin 
>> console and I was wondering if anyone had a nice way of blocking 
>> IPs.  We can get the IP from the request as it comes in and check in 
>> our database to see if its blocked.  My question is how far up do we 
>> want to block.  If someone's IP is 216.27.54.20, just blocking that 
>> exact address won't work for a dynamic range.  Do we need to block 
>> all the way up to 216.27.*.*? or just 216.27.54.*.  Your thoughts?  
>> We are also blocking by username, but this doesn't do a whole lot of 
>> good since someone could just create another user with a different > 
>> username.
>>
>> -Matt
>
>
> Let say, hypothetically speaking (I hope), you wanted to block me.  
> The problem is that my ISP changes my IP address every once in a 
> while.  So if you block that ipaddr, in a few days you'll be blocking 
> someone else, probably an innocent person who just happens to live in 
> the same town.  Now, I doubt more than 1 o 2 people from my town would 
> visit JLCP, just because the population in north central PA is not as 
> technology-minded as, say, the folks in the big cities or Silicon 
> Valley.  But let's say the address you want to block is coming from 
> MIT:  would you then want to block out a whole segment of that campus?
>
> Maybe you could do something like this:  when someone creates an 
> account, they have to have a valid email address.  You acknowledge the 
> account creation, but mark it as tentative until the new person 
> replies to your email test, confirming they really exist.  That is the 
> way many things I sign up for these days (mostly mailing lists) seem 
> to work.  And is JLCP so different (in terms of it being a community 
> vulnerable to abuse) from a mailing list?
>
> Geoffrey
> -- 
> Geoffrey S. Knauth | http://knauth.org/gsk
>
> _______________________________________________
> JL-Volunteers mailing list
> [email protected]
> http://cato.usermagnet.com/mailman/listinfo/jl-volunteers