On Sat, Jan 9, 2016 at 11:47 PM, Josh Triplett <j...@joshtriplett.org> wrote: > Not true at all. A future change to build a more fine-grained version > of non-free could happen just as easily with or without this change.
I don't agree. If there is now lots of effort put into adding another suite, people will probably not be all to happy doing this soon again. Also the name non-free-firmware would kinda conflict with the idea of non-open. Most packages that would go to non-free-firmware are actually also non-open, so we'd end up having this, should non-open ever be added later: non-free-firmware non-open where the former should actually rather non-open/* or non-open-* so it would require another rename or split+moving packages. I can already read how people reject doing this again. > First, and more importantly, because we can very easily define exactly > what "firmware" means to support classifying packages, whereas we'd > spend a long time bikeshedding other fine-grained distinctions among > proprietary software and determining exactly which classifications make > sense to have sections for. The most critical classification people > have asked for separates firmware from other proprietary software; let's > get that implemented rather than putting it off in favor of more > discussion. I don't see which discussion we'd need?! There already is a definition on what's non-free (i.e. DFSG). For non-open, the definition is quite clear: all or some of the sources are no available. > Second, as a point of terminology (and a critical one, *not* a > bikeshed-painting distinction): "free" and "open" refer to the same set > of licenses, and the phrasing you've used would produce massive > confusion. I'm open for other names, I even mentioned "closed-source", even though I'd considered "non-open" to align nicely with "non-free". And in practice I think, that the confusion may not be *that* massive. People being new to Debian have no clue about what main, contrib, non-free mean and they'll have to read it up. Especially contrib is not directly guessable. non-free would be just another line at the list of the other three that needs to be described. Actually I had the impression that many FLOSS activists differentiate very well between free and open, which is why it's the Free Software Foundation, and not the Open Software Foundation, and which is why some projects chose intentionally Free/Libre in their name (and I don't just talk about LibreOffice, where the reason may have been also that OpenOffice was already "taken"). > (The terms differ in what they imply about the views of the > person using them, and in the exact definitions used, but in practice > every license that qualifies as "Open Source" qualifies as "Free > Software" and vice versa.) "Proprietary software with source available" > does not qualify as either Open Source or Free Software, and we should > not call it either. non-open or closed-source wasn't about proprietary or not. unrar is a proprietary format, but the code is open, so it would go to non-free Similar, one could think about software, that is closed but processes something that is not considered proprietary. > I welcome the work currently in progress to establish an archive for > non-free firmware. Well sure it's nice, but we could at the same time have gotten something better that makes more people happy, and therefore it's kinda confusing that this isn't done. Perhaps I've stumbled over one of these cases where Debian is typically moving quite slowly or not at all. > And rather than spend a pile of time arguing over > more extensive classifications, I'd rather work on making even more of > non-free obsolete. But you probably must admit, that this isn't much more than a nice long term goal and nothing that one can expect to ever happen. Sincerely, Philippe
