On Monday 2002 December 02 12:32, Jonas Smedegaard wrote: > On Mon, 2 Dec 2002, Xavian-Anderson Macpherson wrote: > > On Monday 2002 December 02 11:11, Peter Palfrader wrote: > > > Also be aware that one cannot build _the_ binary image which works in > > > all possible scenarios. There are reasons to build things differently. > > > > Is this only due to architectures? I am primarily concerned about > > changes made to software running on the same architectural-platform. I > > realize that there is and must be a tree which conforms to these > > differences. But I just want to know how much duplication can be > > avoided on every branch of the tree? How much is necessity, and how > > much is 'different for the sake of being different'? > > "different for the sake of being different" is called a fork I would call them themes. But I don't have the authority .to call them anything! > > Our job as package maintainers are to maintain the *packaging* of > software, not to make it stand out from the packaging done by other > distributions. > > Debian packages tend to be more true to the original source than those of > other large distributions. In the past there has been problems with Redhat > releasing a version of glibc that was not authorised by the upstream > developers of the library, and caused lots of trouble because software > built against the Redhat glibc was binary incompatible with official > releases - both older and newer - of glibc. This includes closed-source > applications like Netscape Communocator, Adobe Acrobat and RealPlayer that > cannot be recompiled against a proper glibc.
This is why I said no one should have the right to do this. There needs to be a rigid air-traffic control system, just as there is on any major airport. If you want to change course, you have to get permission first, not after you have already crashed! > > - Jonas