Wally Balljacker wrote:Well, Ubuntu and MEPIS both use Debian testing/unstable for their codebase,
I am aware of that, but if you look at it from an "overview point", it's still different. For example, don't Debian Sarge and Ubuntu have different versions of glibc? When I look at "Debian" I look at the current stable pool, because it's their current version (Or what you could call "release version").
Wally Balljacker wrote:but if MEPIS switched to Ubuntu, that would mean MEPIS packages would no longer be compatible with standard Debian. That's my biggest gripe with Ubuntu, is that their packages won't work on Debian systems.
Yeah, I am of the same thought most of the way, but one thing I think Ubuntu really has going for it over Debian is that it's so much more recent, for example, if hal for some odd reason had some new dependency on some fundementals, and it provided good new fuctionality for desktop usage, Ubuntu would be able to include it (I am aware that Debian is more for servers, and Ubuntu is targeting desktops) and the users would benefit from it. One thing I think we need to get in place is the ability to plug in a USB wireless adapter and instantly have hal tell some GNOME or KDE protocols that there's a new WiFi adapter in the system, and when it's taken out again, hal or something else needs to tell GNOME or KDE that the WiFi adapter is gone from the system (I'm not getting my hands dirty down in the kernel layer every day, so it could be that what I'm describing would already be in use today, but in another form), so you have more of this plug'n'play functionality that Windows users take for granted these days, and in all fairness is useful. I can certainly see the point Ian Murdock makes about Ubuntu forking more and more, because you don't want to have RPM hell, but I think that if Linux going to win the desktop someday, we can't stick with older versions of software, if newer versions dramatically improve the user experience, which I think we will see with KDE 4.0, because then Microsoft and Apple will always be ahead of us, for example if GNOME couldn't offer Beagle when Vista is released, we can never "advertise" with a better desktop experience than Microsoft does. Let's face it, if you actually sit down and REALLY REALLY tweak and harden a Windows machine, it actually can be pretty safe to work with for a dad or a mom, if they don't install KaZaA that is (Teenagers do this, unless they're savy like me of course, so parents' machines + non-savy teenagers in the house is not a good combination), and at that point, people don't care if Linux or OS X has better security, what they care about at that point is the overall experience of using the system. If a person gets addicted to the Spotlight features in OS X, I doubt he'll even try GNOME or KDE without an equally good desktop search program, the recent versions of Beagle are getting pretty good and the SUSE guys has made a KDE frontend for it and the Beagle guys have now seperated the GUI and daemon so it could get desktop independant some day. If you were to use Debian as the foundation for a desktop Linux system, and you're gonna be based on the latest stable Debian, you will be quite a step behind Microsoft and Apple, because they don't have this issue of "derivative compatibility" (Which in my mind, and the free software/open source thinking, is a bad thing) so they can rush ahead and add all sorts of bells and whistles that turns avid users heavily on to the product. If Vista is going to ship in 2006, we will probably still have Debian Sarge (Given the Debian history I think Etch will slip into 2007) which has KDE 3.3.2 (I'm using KDE as the example, because most new Windows users probably favor it over GNOME, based on my experience), which will be heavily outdated by then, and Windows Vista be FAR ahead of it. If Linux is gonna beat Vista, we need at least KDE 3.5, and better yet KDE 4.0 with the XGL stuff, because Vista is going to offer "good" looks, and eye candy, and that's something the user will notice right away, and that will catch a lot of people's interest, and if we're stuck with KDE 3.3.2 we will fail for sure, because we're stuck with the current stuff that Debian has to offer (Unless you can satisfy the dependencies another way, which I can't imagine).
Let me point out again, I can certainly understand Ian Murdock's thinking about Ubuntu forking, and that compatibility is a good thing so we don't get "Debian hell" the same way you can run into "RPM hell" (Which I havn't encountered with SUSE yet, because the community package repositories are not as big in numbers, and the two that are dominant are good at looking at what each other is doing), and that you can build a package on Debian and that it would work on say Linspire or something else. But in this time where we in the free software/open source world really need to show the millions of Windows users the good stuff that we have to offer, we just can't let outdated software be the barrier, because then our own strenghts becomes our weakness in my mind.
As I said before, people won't care about how secure GNU/Linux or *BSD are if Windows Vista's desktop experience is better, it's just a plain fact.
