[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [tor-dev] prop224: zero bits in addresses

And 10 years down when 20 bits is easy, you're going
want shrinking it again, or along any interim update cycle.
This is going to upset downstream parsers such as
web indexers that expect matching fixed length / pattern.
or that have to write zero [de]fillers.

ex: [a-z2-7]{16}\.onion, we now see subdomains and uppercase
patterns posted and resolving beyond this original pre224 spec,
where same is hardly widely documented and known.
ie: most untrained users think 0-9 is valid, perhaps even
some full UTF-8 charset.

Rather than trying to shorten crypto hashes or key encodings
for own sake, or for silly human reasons least important of which
is memorization belonging in another layer, consider actual
buffers in apps, shell input, 72 char file width, etc.

Prefixing for some anti-dos sounds nice but also
goes the way of Moore, and you can only soft
increase it without hard banning users with old
code and onion keys.
tor-dev mailing list