Lost sources [was: Re: scientific paper in package only in postscript form non-free?]

[Mark Weyer <mark@weyer-zuhause.de>]

On Wed, Mar 16, 2011 at 07:39:58PM +0100, Francesco Poli wrote:
> On Wed, 16 Mar 2011 09:26:39 +0000 (GMT) MJ Ray wrote:
> > Sure, it should be - what happens if [the source] no longer exists?  That seems
> > quite possible for a years-old journal paper.
> 
> This seems to be a FAQ...
> 
> Well, if some form of that work no longer exists, it cannot be the
> preferred form for making modifications to the work itself.
> 
> One thing is when the author/maintainer uses a form of the work to make
> modifications (because he/she prefers that form), but does not make
> this form available to others.
> In this case, the actual source is being kept secret.
> 
> One completely different thing is when nobody has some form of the work
> any longer. That form cannot be preferred for making modifications,
> since it no longer exists.
> In this case, the actual source is the preferred form for making
> modifications, among the existing ones.

Is this your personal opinion or did you perceive it to be commonly accepted?
I am asking because I tend to disagree. The scenario I have in mind is that
someone takes a, say, GPLed work, modifies it, and distributes the modified
work without (what would otherwise be called) sources, claiming that he "lost"
the sources.
I always thought that such distribution would be in breach of the GPL, or
more generally of copyleft. After all, it is impossible to distinguish, from
the outside, between lost and secret sources. And if the
I-want-my-sources-secret person does not care about later modifications, he
might even really delete the sources.
It should not be that easy to weasel out of copyleft.

Best regards,

  Mark Weyer