I just installed 10.04 on an HP mini (using it now). Is the server edition that far behind the curve or are you just being conservative?
BTW, I frikkin love Ubuntu. It's going to seep into the market quietly over the next couple/few years. I'm so glad to see Gates and Jobs having some of their market taken away, even if it's only a couple of percentage points.
If you use Ubuntu as a notebook/desktop, check out Guake if you're unfamiliar with it:
If you use Ubuntu as a notebook/desktop, check out Guake if you're unfamiliar with it:
http://guake.org
Meh. Guake doesn't look that special to me. I can't see anything there that Snow Leopard's terminal doesn't do. Not that SL's terminal is so special either.
But this reminds me of a complaint I have had for years about Linux: they use repulsive and nonsensical names that absolutely repel those who might switch to Linux. Guake sounds like the lizard people from Planet X. Or a gay version of Quake.
Linux will never go anywhere until it adopts common descriptive names for its basic utilities. Like "Notepad", "Terminal", etc. Grandma and Cousin Jed are never ever going to run a program with a name like Guake.
The Linux folk are nerds and think these obscure names are wonderful. Well, maybe for nerds. But I'm pretty nerdy and I don't like these names either. They're just flat-out annoying. This is a major reason for the failure of Linux on the desktop, along with the ongoing rivalry and issues surrounding the Gnome versus KDE desktops and libraries (hey, Linux guys, pick ONE!).
As far as the geeky names go, the problem is that Linux with it's vast open source development base, has a dozen different software packages available for any single application need, and simple names like "notepad" just aren't available.
As far as the geeky names go, the problem is that Linux with it's vast open source development base, has a dozen different software packages available for any single application need, and simple names like "notepad" just aren't available.
I think they could adopt a convention of designating a particular app version as the official standard and giving it a decent name. I also think that Linux will never succeed as long as there is a divide between the Gnome and KDE camps.
User interface design really is no longer any kind of rocket science: just pick one and relegate the other to legacy support. It doesn't even matter which one you pick; just pick one.
If you had a system like this for Notepads and Terminals, then a new geeky-named competitor (fork or new project) might over time become more popular, then every few years they could choose to go to the new Notepad program. Let's face it, every system has Notepad programs. Some proliferation of names, often for obscure features, isn't that desirable. And just how extensive or subtle can a Notepad or a Terminal really be? This isn't actually a major divide that moves the hearts of men, causes people to kill each other, determines which computer system a person chooses to use (in the case of Linux, which computer system they choose not to use largely because of this endless confusing proliferation of minor versions of almost identical software).
You need continuity of sorts to succeed. Then you have more books and magazine articles to tutor and promote the various packages, you have users who become knowledgeable about a standard utility and can help others.
As I've said, this is one of the biggest single defects in any plan for Linux to conquer the desktop. If you're satisfied with obscurity (or if you purposely desire it for various reasons) then the current Linux practices are the way to go.
Linux needs to finally become a platform, in the same sense we use those terms towards Windows and Mac OSes.
Linux needs a way to standardize their APIs and utilities but keep the open-source nature of the Linux system.
Linux is weaker as an OS because of the divided effort and duplicated effort of so many teams, especially Gnome vs. KDE. I know some people would fight to the death over the two but that is legacy stuff. Both are more than capable enough and the division in UI development harms Linux far far more than it can ever help it. I'm on Mac instead of Linux largely because I actually like having a standard set of utils that everyone else is using. I also like that OS X, unlike Linux, actually is a UNIX 2003™ certified OS. This is one reason why it was fairly easy to change the Pinguinite install script to produce a working P forum on the Mac. (I should remember to email that script to you just in case anyone else asks for it.)
Of course, the very idea of standardization offends many Linux users. So I expect total marketshare/mindshare of Linux to continue to decline because of this tendency in the Linux community.
To make my own prejudices clear, I think Linux should standardize on Ubuntu with KDE (Kubuntu). But Ubuntu with Gnome would work too.
the very idea of standardization offends many Linux users.
The idea of top-down enforcement of standardization is what offends, not standardization itself. Standardizing based on merit, utility and efficiency appears to be the goal.
Remember when we had a bazillion proprietary network protocols to contend with while TCP/IP was chugging along silently in the background? Then BOOM! TCP/IP is now the standard (even with its high overhead) that even IBM and Microsoft had to defer to. I think that Linux (and other *nix variants) are going through a similar process. I'm willing to go through the BS of the occasional frustration dealing with irregularities in the OS if it leads to freedom on the desktop and the server.
Standardizing based on merit, utility and efficiency appears to be the goal.
If that's true then the Linux elite have failed miserably with their apartheid repository system. One repository for the "pure" free stuff, and at least one more for the good stuff that has something they don't like about the license. Gotta segregate the pure stuff from the tainted, doncha know.
In most distros this ends up as a real pain in the ass for the end user, who has to enable repos, and find the missing magic codecs, etc. to make stuff actually work.
END THE HATE, integrate the repos!
I love Linux and am on openSUSE now, but the purist "free" fanatics tick me off!
#36. To: bluegrass, Pinguinite, SonOfLiberty (#22)
Hell, Linux and FreeBSD already own the internet.
Mac is FreeBSD. And it is the only UNIX 2003™ certified UNIX distro.
The argument can be made that Apple is the last standing UNIX platform. And it is the only BSD that is actually a certified UNIX.
Darwin was the Mac version descended from FreeBSD 4.4, a major and stable version of FreeBSD. They combined this with the Mach microkernel to make a UNIX system, then built the Mac desktop functionality on top of all of it.
The Mac APIs are elegant and straightforward. All the 32-bit stuff is almost gone now though they have some optional 32-bit compatibility remaining. They have eliminated the old Quicktime and are building a new 64-bit Quicktime from scratch around H.264.
I would prefer that Apple had retained more of the flavor and utilities from FreeBSD. For their own convenience, they replaced some of the standard FreeBSD config files and utilities with their own. To me, this has weakened the Mac offering overall. To me, this is unnecessarily proprietary and reduces interoperability. I am also disappointed they didn't fully implement ZFS in Leopard yet; I felt that was promised and they failed to deliver it. Not entirely but we were led to believe that it would be a consumer filesystem by now. Apple dropped the ball there despite Sun's generosity in making it available to them. It's just time for Apple's HFS to die off entirely.
Apple offers some geeky freaky stuff for developers, stuff you just don't find on other platforms. OpenCL to allow you to harness all your CPUs and GPUs at once (including multiple machines on a network), this is a budding Apple standard. Stuff like Quartz Composer, a graphics effects engine that is so cool you just have to play with it; it's like building electronics but you produce graphics (which can then be embedded in websites, in Dashboard widgets, in screensavers, etc.). And it is easy, something you don't have to be a programmer to do (though it helps). Offering the Instruments profiler and other Xcode tools has made Xcode a real competitor to Visual Studio in productivity terms. VS is still ahead in its sheer size and capabilities but so much of that is crufty legacy features they have to support for their old crappy stuff that they foolishly sold years back. Apple cuts that old crap off entirely and orphans it so you don't get tied to bad ideas someone had 5-10 years ago that never worked out but which during their heyday acquired a certain number of users. Microsoft is badly hobbled by having done so much of this over the years.
Apple is using OpenCL to harness multicore CPUs and to allow transparent effortless use of GPU power in a system that readily and transparently adapts to whatever system you run it on without recompiling. They offer live kernel debugging using a dialect of D. They have GCC in XCode but have moved to LLVM and will complete the move to Clang, all very advanced compiler technologies. They offer Grand Central Dispatch to control tasks and threads, a very sensible approach to the proliferation of hardware we all have but which is often completely unused by the system because there are so few rational APIs to make use of them, largely because parallelism is such a difficult topic.
I don't see Linux positioned in any way to make use of these new technologies. Technically, there is very little rocket science in Linux these days. It is essentially a moribund platform, dying a slow death. Torvalds is not a computer scientist and his tight control over the Linux kernel may have contributed stability but it has stagnated, even in comparison to Windows. And it isn't even in the same ballpark with OS X.