Re: tl26 pretest start

On Mon, Feb 23, 2026 at 07:57:11PM +0100, Norbert Preining wrote:
> Hi
> 
> > I get your point.  Nonetheless, could signify be an option worth
> > considering?  The system adopted by OpenBSD consists of providing the
> > public keys that will be used for the next two releases in each
> > release.
> 
> And how is guaranteed that those keys are not tampered with? What kind
> of check is done on the authenticity of the keys?
> 
> Assume a evil mirror changing the key, the files, whatever. signify
> whould be more than happy with those files. But gpg would show that
> the key is not signed by anyone, in particular not by Karl or me.

Surely I wouldn't say otherwise, which is all what so-called web of
trust is about.  That said, what do think of the way Ted Unganst, the
author of signify, put this:

> There are no key servers for signify. No web of trust. Just
> keys. The good news is the keys are pretty small. As
> demonstrated. We can stick them just about everywhere, and we
> do. They're on the web site, they're on twitter, they're on the top
> side of CD. 56 base64 characters. You can read it out loud over the
> phone in under a minute. Wide dispersion makes it harder and harder
> to intercept all the ways you may get the key and increases the risk
> of detection should anybody try some funny business.[1]

> > I raise this question because when one installs TeX Live, the script
> > asks whether to reuse the settings from the previous
> > installation. (When there is one.)
> 
> There are surely ways to improve but I don't see how signify (which I
> checked, too!) is actually improving on the situation.

The idea behind signify is that of trust-path, instead of web of
trust.  That's why I came up with the idea that a TeX Live
distribution N could include N+2 keys.  No more.  The chain is
obviously not as robust, I completely agree with you.  But it's much
simpler, so it's all about accurately assessing what is really lost in
terms of security.

I just find the discussion interesting, actually!

Best,

-- R.

Footnote:
[1]  https://www.openbsd.org/papers/bsdcan-signify.html