Conclusion of Re: DFSG and GPL -- source retention

[Raul Miller <rdm@test.legislate.com>]

Jason Gunthorpe <jgg@ualberta.ca> wrote:
> And to understand why you accept that 'the place' is 'the internet' so
> 'the same place' is 'any internet site' and 'equivalent access' mean
> anon controls basically. I belive this has been an accepted meaning,
> you might want to ask RMS.

I've sent off a query.

Certainly, I got so used to people claiming that distributing KDE and
Qt on the internet didn't mean that they were being distributed together
that I didn't think to re-examine the concept in this context.

My feeling, in this context, at this point in time, is that the
information we use to configure apt defines the place where we're pulling
binaries from.  Which just underlines how immature our support for source
distribution is, and how weak our support for mirrors is.

> Anyone downloading a package from a binary-only mirror has internet access
> and thus equivilant access to a mirror that does carry sources. We do not
> need notices or require keeping packages for 3 years to satisfy
> binary-only mirrors. We do need to keep releases for 3 years to help
> binary-only CD vendors.

I guess this is good enough for the moment.

Thanks for clearing this up for me.


[switching from gadfly mode back to something approaching a design
viewpoint:]

I'd like to eventually see apt engineered to work from a list of mirrors
[perhaps always retrieved from the same location, though it would be
better to have the list always signed by the same key], with something
like netselect used to determine the order in which mirrors are polled,
and with some sort of "meta-Packages" files to indicate when a complete
"debian distribution" has been located.

Thanks,

-- 
Raul