Re: DRAFT: debian-legal summary of the QPL
On Fri, Jul 23, 2004 at 11:18:28AM +0200, Sven Luther wrote:
> On Thu, Jul 22, 2004 at 04:45:07PM -0700, Josh Triplett wrote:
> > Sven Luther wrote:
> > > On Wed, Jul 21, 2004 at 09:05:40AM -0700, Josh Triplett wrote:
> > >>Sven Luther wrote:
> > >>>On Mon, Jul 19, 2004 at 12:01:57PM -0400, Brian Thomas Sniffen wrote:
> > >>>>luther@debian.org writes:
> > >>>>>Well, simply configuring your SVN/CVS/ARCH/Whatever archive to spam upstream
> > >>>>>with every change done should resolve all the issue. Or maybe giving him
> > >>>>>consultation access would be enough.
> > >>>>
> > >>>>Spamming upstream is not enough. You have to provide one on request,
> > >>>>even if you just sent one. Additionally, now you're suggesting doing
> > >>>>away with the ability to make private modifications.
> > >>>
> > >>>Bullshit, you have provided it before it was asked, so where is the problem ?
> > >>
> > >>Do you see anything in the QPL that says the original developer can only
> > >>request your changes once? They can ask twelve times a day if they
> > >
> > > Well, whatever is the problem ? You provide it to them, and if they ask you
> > > again, you either say, sorry, i sent it to you already, and haven't got a
> > > backup copy, would you like the latest version perhaps ? If you already
> > > fullfilled the request before you are asked, where is the problem.
> >
> > >From the QPL:
> > > c. If the items are not available to the general public, and the
> > > initial developer of the Software requests a copy of the items,
> > > then you must supply one.
> >
> > Where in there does it say that you may refuse to supply a copy if you
> > have already provided one?
>
> Where in it says you have to ?
Where in it says that you don't? For my part, I can't see how either
interpretation is more plausible than the other.
In this sort of situation we *can't* take the easy interpretation; if the
least friendly plausible interpretation is non-free, we have to go down that
path. Optimism is dangerous.
> Please let's not forget common sense.
Common sense is being used. We're approaching this from a pessimistic,
protection perspective. You're approaching this from a "it has to stay
free" perspective.
> > >>want, and you have to comply; there is nothing in the license that says
> > >>otherwise. For that matter, do you see anything in the QPL that says
> > >>the original developer has to compensate you for the costs of providing
> > >>your changes (bandwidth charges for network distribution, or media costs
> > >>for physical distribution)?
> > >
> > > Yes, since the distribution will happen accordying to 6a, which says you can
> > > charge for the cost of data transfer.
> >
> > >From the QPL:
> > > c. If the items are not available to the general public, and the
> > > initial developer of the Software requests a copy of the items,
> > > then you must supply one.
> >
> > Where in there does it say that the copy you supply to the initial
> > developer is covered by the terms of 6.a? 6.a only covers recipients
>
> Well, it is evident. The section 6 covers how you distribute these code
> linking with the library. IF you distribute such code, you have to cumply to
> all of a, b and c, is it not ? You don't see in the main header of 6 that you
> have to satisfy one or the other, or you could safely ignore 6c and the whole
> point would be moot.
All three subclauses have to be satisfied or judged to not apply. 6a
doesn't apply to source-only distribution ("all recipients of
machine-executable forms"). 6b applies to all distribution. 6c only
applies if the items are private *and* the initial developer asks for a copy
of the item.
In the instance of applying 6c, we recurse through the licence, go through 6
again, and *again* we don't apply 6a because the initial developer asks for
a copy of the source. Of course, the obvious loophole there is that the
initial developer asks for a copy of the binary instead, in which instance
6a is invoked, and all's good. But is charging for a binary instead?
Presumably it is, as otherwise the licence is non-commercial-only, and
non-free, but there's no exception for the initial developer on that point,
so I can charge the initial developer an unrealistic amount of money for my
binary.
> > who have a binary and want the source. In this case, if you are
> > distributing source (that is not available to the general public), then
> > the source is one of the "items" in question, and it must be provided
> > under 6.c, which does not indicate that you may charge for cost of
> > distribution.
>
> Notice that 6c speaks about "copy of the items". How do you interpret this.
In the absence of clarification, I'd imagine it'd mean "a copy of the
source", because the binary is of very limited use to the initial developer.
No binary means 6a doesn't apply.
> This has no meaning apart from the stuff described in the 6 header, that is :
>
> You may develop application programs, reusable components and other
> software items that link with the original or modified versions of the
> Software. These items, when distributed, are subject to the following
> requirements:
>
> These items clearly refer to "application programs, reusable components and
> other software items that link with the original or modified versions of the
> Software", and this clearly states that you have to cumply with all of the
> following, 6a to 6c.
Comply or show as non-applicable. In the same way that 6c doesn't apply to
every act of distribution, 6a doesn't apply in all situations of
distribution either.
> > >>[Do you want both of your email addresses CCed on these mails?]
> > >
> > > Not really, but i prefer more of them than none at all, as hiting D is easier
> > > than reading mail in lynx.
> >
> > No problem; I'll continue to CC luther@debian.org on all of my mails
> > related to the QPL discussions.
> >
> > (Are you using webmail through lynx?)
>
> I have no choice, since i was not originally CCed, i have to go to the web
> archive to read the discussion, get the url of emails i want to respon, launch
> lynx over ssh on the box which does mail processing, open the url, go to
> respond to or whatever link and send the mail.
Does copy-and-paste not exist on your system?
- Matt
Reply to: