Re: Release plans

[Stephen J. Turnbull]

Alan Mackenzie writes:

 > What, exactly, are we getting so worked up about, then?

Well, I've been perplexed by the "no dynamic loading" policy for about
a decade now.  I know that Richard is not going to give an answer
except that he's fearful, uncertain, and doubtful about dynamic
loading, and so has decided to avoid it.  I was hoping you might
provide some insight, but you're giving me the same line.  I'm very
frustrated by that.

 > Right.  So do you agree the present issue is a matter of analysis and
 > judgement, weighing up risks and benefits?

Of course I do.  I see small benefits (measured in terms of freedom,
as I understand it) and much smaller risks (ditto).  That's precisely
what I've been arguing all along, but you and Richard assert that
*any* risk is too large.

 > > No.  I understand your point.  "Introducing a module loader could
 > > cause Emacs to become non-free."  That's scary.
 > 
 > Again, non-free "for whom?".  If you'ld've made specific reference to
 > people who couldn't or wouldn't take action against non-free add ons
 > themselves, I'd accept you'd understood my point.

I don't think there are *any* people who *couldn't* take action[1],
and I don't understand why "wouldn't" is a problem in the context of
freedom.  There are two kinds of people who wouldn't, those who have
considered the consequences and decided they don't care, and those who
just don't care.  Either way, why isn't it their place to decide, and
our place to provide them with the information they need to make
informed judgments?

And I still don't see how they "lose" freedom or Emacs becomes unfree.
In the first place, they *gain* capabilities that they did not have
before.  True, those capabilities do not come with the four freedoms
attached, but what have they "lost"?  In the second, Emacs itself is
still free, the users have the four freedoms with respect to it.

 > >  > and only allowing signed (by MS) Lisp libraries to be loaded.
 > 
 > > Again, the technical difficulties of accomplishing this in GNU Emacs
 > > are enormous.
 > 
 > Really?  A few defadvices in the right (wrong?) place would do the trick.

(ad-stop-advice)

It really is not that easy.  And yes, since it's supposed to be a
security feature, you do have to deal with savvy users like me who
would be able to handle a defadvice of ad-stop-advice.  The whole
thing would get blown out of the water with a million CERT reports.

 > > Remember, GNU Emacs's security features were designed by the guy who
 > > posted his password for somebody else's system in a public place.
 > 
 > What was somebody saying about ad hominem attacks not all that long ago?

It's not really an ad hominem.  It's an *allusion*, to the fact that
*by design* Emacs has no security features, except to protect itself
from crashing, and a little bit for protecting passwords (but a couple
of defadvices would defeat that, too).


Footnotes: 
[1]  At least, they have other, much more dire problems, as Richard
has been at pains to point out to you.  You *can* be free, you just
don't want to badly enough.