Re: Negative Summary of the Split Proposal
You asked for comments about the text, not the issues, and I tried to
honour that.
Jason Gunthorpe wrote:
> ----------- Cut -------------
> This text is a negative response to the non-free split proposal currently
> being deliberated by the Debian Project. It attempts to summarize the
> concerns voiced by some developers. Only concerns that were left largely
> unresolved by discussion are presented here as 'facts' and observations
> about what other consequences the proposal will have. Please see the
> Debian-Vote archives for the full discussions.
I had to read the third sentence three times to understand what you meant
with it. I suggest a comma after "here".
> It is Only the Start..
> ~~ ~~ ~~~~ ~~~ ~~~~~~~
[...]
> and some of the Archive Masters have indicated that they would
> give the new non-free upload queue a lower priority than non-free software
> currently enjoys.
As one of the current archive maintainers, I can confirm that I have
no interest in maintaining an archive dedicated to non-free software.
If "some of" means more than one, then by my count this makes it
unanimous.
It means that new maintainers will have to be found for the non-free
server, which you may want to mention under "Cost".
> Too Specific
> ~~~ ~~~~~~~~
> The proposal does not set any sort of general rule but addresses a single
> problem with a single site. Even after voting the question will still
> remain if anything should be done with non-us, the web site and other
> places the contain references that may confuse users about the separation
^^^ should be "that"
> that exists. It addresses a single technical problem and it does not
> address the larger issues of Debian's support of non-free software that is
> not part of The Debian Distribution.
After this section I would add a new section:
Too Vague
~~~ ~~~~~
Implementation of this proposal is not straightforward, and is left
largely unspecified. This makes it hard to understand the full effects
of the proposal. It is not necessarily a problem at this stage, but it
should be rectified before a General Resolution is proposed.
> Does it Really Matter?
> ~~~~ ~~ ~~~~~~ ~~~~~~~
> Ultimately the proposal is based on the assumption that users are/will
> be confused by the proximity of the non-free packages to the main
> archive. This is really impossible to prove either way so it has
> represented a point of contention. Several people suggest that
> non-free.debian.org is no more clear than debian/dists/unstable/non-free
> and that tools like APT (which already indicates if the package is
> main/contrib/non-free) make the change all but pointless.
Is this true? I've never seen APT indicate that.
You may want to write out "main, contrib, or non-free", because otherwise
you use slashes as path separator and as alternatives separator in one
sentence.
> Cost
> ~~~~
> As stated the proposal will require a separate machine for hosting the
> new non-free archive with bandwidth able to handle an excess of 3G per day.
Probably best to begin this with "The proposal, as stated,". Otherwise
you imply that this text has stated the requirement.
> This will require Debian to solicit sponsorship and purchase hardware or
> place more load other machines to support this increase in bandwidth.
> In all likely hood the new non-free.debian.org machine would also contain
^^^^^^^^^^^ should be "likelihood".
> a mirror of the main archive and serve as a master.non-free.debian.org.
(Do you have evidence for this likelihood? I see no reason for it.)
> Since we would be removing non-free from the main archive an alternate
> mirror network would have to be arranged and new mirrors setup and our
> users informed of them. With nearly 200 mirrors of the main archive this
> is a non-trivial loss that will be time consuming and difficult to recover
> from. Many people will have to be involed and some mirror masters may not
^^^^^^^ should be "involved".
> agree with our discision.
^^^^^^^^^ should be "decision".
You may want to mention here whether the mirror coordinator will want to
do this work :-)
Richard Braakman
Reply to: