----- "Al @ Lab42" <lab42...@gmail.com> wrote: > I keep on considering the node's inheritance approach a viable, > working, flexible and even elegant solution to variables assignements > to nodes.
I agree they aren't inherently bad and afford a lot of flexibility but the way they work you need to be a fairly advanced/proficient puppet master - almost programmer really - to use them correctly and to understand all the ways they interact with your nodes. This is mostly why I avoid recommending them. They also have the same issue as any OO based solution, the consequences of a code diff isn't clear. You change a class/node and some others inherit from them the results aren't obvious by just looking at the code you have open at the time. Again not a problem that lets say experienced programmers face or those using actual programming languages since they have IDEs that help them navigate the maze of inheritance, things we don't have. There was a great mail from Linus recently about C vs C++ that discussed this thing at length, you can apply a lot of his arguments for C rather than C++ to using inheritance and OO structures. Obviously the results you achieve in your lab42 infrastructure speak for themselves and stand as a good example for inheritance, but explaining it to new people - esp ones who arrive with expectations from other languages - is pretty hard, esp over IRC or docs. -- You received this message because you are subscribed to the Google Groups "Puppet Users" group. To post to this group, send email to puppet-us...@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
