> Raul> I just don't think that specifically enhancing the package structure > Raul> with extra kludge to specially support non-free packages is the right > Raul> way to go. > > Look below to see why it is not a kludge. > > Raul> I think that the right way to go is to put the references themselves > Raul> into non-free. The data structure itself is less complicated, the > Raul> complexity is represented in the data. > > Raul> And, by the way, you've still not explained why you think the approach > Raul> I'm advocating is grotesque and that the approach you're advocating > Raul> is elegant.
On Fri, Dec 03, 1999 at 11:03:19AM -0600, Manoj Srivastava wrote: > Because the suggests and dependencyis and other relationships > are based on how the packages work: they are, in some ways, a natural > consequence of the what the package is, and does. > > Free and non-free are a consequence of the *licence*, which > has little to do with how the package works technically. Sure, and now you seem to be advocating some new package headers which are a consequence of the license. > Now, the relationships will be independent of the licence, > just depending on what the packages are (elegant, in my opinion), and > I tell my tools what packages I do not want installed (not imposing > my vierws on other people, nor using licencing issues to distort > relationships. No problem there. > This is configuring how my package management system behaves > on my machine. Again, elegance. It shows me what I want to see, as it > should. Still no problem. But all this reasoning applies for the case of a free package with a non-free micro-package which enhances the free package and which suggests various non-free elements. Except that it doesn't require any license dependent package headers. The "Enhances:" header is already going to be useful for things like gimp plugins and perl libraries. -- Raul
