On Sun, Apr 07, 2002 at 11:56:59AM -0600, Joel Baker wrote: > On Sun, Apr 07, 2002 at 12:12:47PM -0500, Joe Wreschnig wrote: > > On Sun, 2002-04-07 at 06:14, Federico Di Gregorio wrote: > > > people, i just want to remember you that DFSG stands for debian free > > > SOFTWARE guidelines. documentation is *not* software > > Unfortunately this is becoming less true. CSS contains statements for > > content generation and counting variables. Is this a program? I'm not > > sure, but it's definitely not just a document anymore. XSLT can be > > included as "documentation" (and probably is in a lot of places, in or > > outside of Debian), and XSLT is Turing-complete. Where does the line get > > drawn? Is it possible to draw one? > > IMO, an FDL-licensed document with invariant sections is non-free. As a > > user of Debian, I'd like to know that they're not installed on my system > > if I'm only using packages from main. > As noted - that will mean most of the GNU stuff goes right out the window. > Perhaps Woody+1 will no longer be "Debian GNU/Linux"? > I've said it before, but once again: the world of "writing" (that is, the > various forms of documentation, RFCs - many of which are 'non-free' under > the DFSG, and similar things does *not* have the same baseline of what it > means to be 'free', because it comes from a vastly different world. One in > which 'open distribution of work' is the primary goal, and the basic means > of 'modifying' a work all preserve the origional document intact (that is, > annotation, commentary, and bibliographical reference). > The DFSG is an excellent place to start, but trying to apply it to things > which *are not software* is silly, and results in the sort of sillyness > which we're seeing now - will we see an Orphan message for GCC next? The issue is that the Debian Social Contract doesn't say "All software in Debian will remain 100% free", it says "Debian will remain 100% Free Software." Therefore, for something to be part of Debian, it must be Free Software, even if it's documentation. Now, this may be an oversight in the original phrasing, but this is the Social Contract that we've all agreed to uphold as Debian developers -- unless and until it's clarified to address the various issues that arise with other forms of data, we really don't have anything else we can point to when judging the license on documentation. > I know we don't like 'patches only' software, but we *do* allow it - and > the basic assumption of most documentation is that it lives in a world in > which various forms of 'patching' are the *normal* method. I'm all for us > saying 'please try to minimize invariant sections', possibly even 'these > types of sections cannot be invariant to qualify for the DFDG', but if we > want to apply a standard to which the rest of the world will never allow > itself to be held to, we're going to take RMS's place as the zealots whom > large numbers of people ignore. I'm intrigued by this idea, and think it does indeed have a lot of merit. Documentation, after all, is akin to source code in the sense that both are intended as human-readable content, not as obscure instructions to be delivered directly to a computer. If we allow an author to place restrictions on how we can modify some kinds of source code while still considering the code free, why should the same not be allowed for other types of source code, like documentation? Steve Langasek postmodern programmer
Attachment:
pgpuPAfrJ54JB.pgp
Description: PGP signature