other ways of working with third party vendors [was Re: Proposal: Revision of policy surrounding 3rd party and non-free software]

[Matthew Garrett]

On Wed, Jan 22, 2014 at 11:05:48AM -0500, Christian Schaller wrote:

> But to answer your question, we could for instance make a browser extension
> that searches the web for Fedora repositories, repositories which we have 
> worked with the third-party vendors to make sure they contain metadata we
> can have the extension nicely display and then present in the web browser
> maybe with enable/disable buttons like the extensions.gnome.org site.

That's not really what I was suggesting. Let's think of this in terms of 
channels - the software centre would be a consistent UI that provides a 
way of viewing multiple channels, and Fedora would ship with a 
predefined set of channels installed.

You want that set of channels to include a number of third-party vendors 
who distribute non-free software. There's a few practical problems here 
- how do we choose those vendors? What process do we have for ensuring 
that they aren't distributing malicious code? What if they provide a 
package that breaks software that we ship as part of Fedora? What if a 
vendor with a known history of shipping broken software requests 
inclusion and kicks up a PR storm if we refuse?

Most of those problems go away if we adopt a model where the only 
default channels are ones pointing at Fedora repositories, and then make 
it easy for third-party vendors to provide their own channels. We'd 
provide best practices information to vendors on how to advertise these 
channels on their site, with recommendations for standardised wording. 
The extra work for the end user is minimal, but we avoid advertising 
non-free software, it's obvious to the end user that they're obtaining 
code from someone unaffiliated with Fedora and we don't end up with the 
strange situation where we're happy to provide the user with non-free 
software from Adboe but not free software from RPMFusion.

Your argument seems to be that the additional effort to locate these 
channels would be sufficient to deter a significant number of users from 
running Fedora, and altering our foundations is justified as a result. 
I'd really appreciate non-anecdotal evidence to support that position.

-- 
Matthew Garrett | mjg59 at srcf.ucam.org