Steve Atkins via mailop <mailop@mailop.org> writes:
>> On May 10, 2019, at 10:50 AM, Leo Gaspard via mailop <mailop@mailop.org> 
>> wrote:
>> Laura Atkins via mailop <mailop@mailop.org> writes:
>>> For victims of listbombing, COI isn’t an answer. In fact, much of the 
>>> problem with listbombing is COI mail.
>>> 
>>> How do you propose to address that issue?
>> 
>> Captchas are a way to force the malicious subscriber to spend human or
>> computer time breaking it (if captchas can be broken).
>> 
>> As a consequence, it might make sense to just accept it is the case and
>> require a javascript-generated proof of work before accepting the
>> subscription.
>> 
>> This way the amount of computer time required can be more easily
>> modulated, and the human is not faced with a hard-to-answer captcha -- I
>> don't know how you feel, but as a user I'd much rather have my computer
>> spin for a few tens of seconds before sending the form (especially as it
>> can happen at the same time as I fill it) than have to fill in a
>> captcha.
>
> Bad people have access to much more, and much cheaper, compute resource
> than good people.
>
> There are other things wrong with your suggestion but that's
> a simple thing to think about first. It also generalizes to many other
> poorly thought through "just use proof of work to solve email problems!"
> ramblings.

I hope you didn't assume I was saying it's a good idea. I was *only*
saying it was a better idea than captchas, as an answer to Laura's
question.

And I stand by this statement: the effect on bad people is approximately
the same as the one captchas have (well… maybe a bit less efficient when
the captcha is strong enough for requiring a captcha relaying attack?
but I'm used to failing answering those captchas as a human too anyway,
so might as well have a 10min proof of work…), and the effects on good
people are much, much less painful than the ones captcha have.

Basically, captchas are a POW for malicious actors, and a dice the user
rolls for the normal user. When computers were unable to do basic OCR,
they were a good idea. Nowadays, the user gets more pain than the
malicious actors, for whom even 1/10 success rate is enough to get lots of
forms through without questions.

_______________________________________________
mailop mailing list
mailop@mailop.org
https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop

Reply via email to