phc-discussions - escrypt naming (Re: [PHC] TigerPHS paper and code ready for review)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20140312024409.GA18880@openwall.com>
Date: Wed, 12 Mar 2014 06:44:09 +0400
From: Solar Designer <solar@...nwall.com>
To: discussions@...sword-hashing.net
Subject: escrypt naming (Re: [PHC] TigerPHS paper and code ready for review)

On Tue, Mar 11, 2014 at 10:11:48PM -0400, Bill Cox wrote:
> Naming is definitely a bike shedding event.  There just aren't enough
> good names to go around.  I googled Xcrypt and XScrypt, which I
> thought could be a decent alternative for Escrypt, the X coming from
> eXtended.  No luck... they're taken, in uncool ways to stomp on.

Yes, I had tried Googling these too.  Also transcrypt.  No luck.

We also need a name for the sub-block transformation in escrypt (which
plugs in a similar place like Salsa20 in scrypt).  I tried Googling
wxform (for "wide transformation") and pxform (for "parallel
transformation") - turns out these are taken. :-(  Should it be
"pwxform", with "pw" referring to "parallel" and "wide", and also maybe
being pun on "password"?  The natural name for the PHS would then be
"pwcrypt", but that's obviously taken many times and isn't specific.
Should the entire PHS be called "pwxform" instead?  But what to call the
sub-block transformation then?  We can't leave it entirely nameless at
least because programming languages mandate some identifiers there.

One relevant keyword is "scalable" - in terms of wider than scrypt's
practically usable range of memory cost settings, as well as in terms of
SIMD and instruction-level parallelism.  Any good way to include it?

I thought that "trans" could be it, in the sense of escrypt being
suitable as a replacement for a wider range of previous schemes
(including both bcrypt and scrypt), via it possessing properties of
multiple previous schemes.  But unfortunately transcrypt is taken.

Thank you for trying to help!

Alexander