On Thu, Dec 14, 2006 at 02:48:14PM -0800, Danek Duvall wrote: > But while the /usr/xpg[46] directories are a decent jumping-off point for > thinking about /usr/gnu, I'd be hesitant in extending the analogy too far. > The difference, I think, is in perception. /usr/xpg[46] are very small and > precisely populated, and are likely to get even less populated as what > functionality can be folded into the main utilities is (now that it's been > made explicit that they can be and should). No one could possibly mistake > them for being complete environments. I think that most people see them > as being odd little warts that you can or should use if you need special > behavior, and that most people don't. > > /usr/gnu, on the other hand, provides (for) a well-known set of > functionality which, in a full implementation, provides a near-complete > unix-like environment. (I agree that some method of combining /usr and > /usr/gnu is necessary for a full Solaris experience, but that's well beyond > the scope of this project.)
I'm in no way convinced that the analogy is a poor one. Much of the functionality of the GNU variants can (and probably should) be incorporated into the base programs as well. It's unfortunate that licensing precludes direct incorporation of the code, but it's by no means impossible to implement it independently. There will remain incompatible options and behaviours, which ought to be the rationale for /usr/gnu. > I'll also note that none of /usr/ucb, /usr/xpg[46], and /usr/ccs have their > own manual, and that for at least /usr/xpg[46], the variants are mentioned > in the mainline manpage. For the record, I think we ought to do the same for /usr/gnu functionality. What is the *technical* rationale for doing otherwise (assuming, of course, that the present course is the correct one for /usr/xpg4 and friends)? -- Keith M Wesolowski "Sir, we're surrounded!" FishWorks "Excellent; we can attack in any direction!"
