Re: Module naming and identification

On Thu, Apr 16, 2009 at 10:37 AM, Lex Spoon <[email protected]> wrote:
> Finally, I would slightly quibble about the discussion of DLL hell.  I agree
> that, e.g., choosing the module author's latest version of a module will
> inevitably lead to incompatible modules, rending of hair, and gnashing of
> teeth.  However, such a choice is subtly different from choosing a
> *distribution's* latest version of a module, which IMHO looks pretty good.


It seems plausible. The first question to ask about any such more
flexible binding is: What opportunities for attack does it create?

I dealt with similar issues in
<http://wiki.erights.org/wiki/Safe_Serialization_Under_Mutual_Suspicion>
in the chapter "Manipulating Authority at the Exits". Mysterious
summary: Literal integrity is easy; comity is hard.

I think the issue you're raising is best understood as a comity issue.

-- 
Text by me above is hereby placed in the public domain

    Cheers,
    --MarkM