On Tue, Jan 13, 2015 at 9:15 AM, James B. Byrne <byrnejb at harte-lyne.ca> wrote: > > Is one to infer from that remark that the E in RHEL has no meaning > whatsoever? And that it should be ignored? Or perhaps redefined to > whatever is convenient for the moment and the POV of the definer? In > which case is it anything more than noise? In any case, the point of > the defining the word was to show that Enterprise != Large, nothing > more. The place where it matters is for companies large enough that they have written their own applications and need a stable OS and library set to run on. Every interface change and install/operating procedure change is going to cost development and training time that would generally be better spent improving your own application. If you just run the applications included in the distribution, it doesn't matter so much since even if the internal interfaces change, they stay consistent within the packages the distribution ships together - that is, someone else has already been forced to deal with the breakage. > In any case, it seems to me that the rather recent innovation of > software collections indicates that perhaps I am not alone in that > observation. And the need for docker as an even more extreme defense against OS/lib instability really points out the problem. > As it happens a most useful, to me at least, piece of information was > revealed in the course of this thread. That was the existence of a > server based stream for Fedora. That's an interesting turn of events, but is this just a separation of packages or is there really a group in Fedora that actually maintains large server farms and has an interest in keeping their applications working? -- Les Mikesell lesmikesell at gmail.com