Re:Promising? Yes. Usable? not really
on
KDE 4.2.4 Released
·
· Score: 1
Hmm, interesting. I can't say I've tested all of those, but...
Ftp works fine for me. For sftp I either use the console or use fish:, which also works.
Didn't try editing over a standard ftp, but fish works.
Don't use Samba, sorry. Even my Windows boxes have NFS (it's an optional feature built into in the higher Windows editions, believe it or not).
Fish works for me.
None of my networks use EAP. WPA-PSK works fantastically, much better than in 3.5.x.
I just turn off the Klipper applet and use it like copy-paste on any other platform. The only time I would even attempt to use Klipper is if I needed to transfer a large number of things from one window to another, but separately. In that case, I would usually use a sticky note widget or similar.
To my knowledge we have at least general knowledge of every major technology we would need to travel between stars
No offense, but WHAT?? There are so many problems with interstellar travel... lets consider a few, from the perspectives of spacecraft and tech available today.
Spacecraft range: The biggest single problem. Space is big. Really, really, REALLY big. The fastest spacecraft we currently posess would take centuries to reach Alpha Centauri (the next-closest star, a mere 4 light-years or so away). In theory, an ion-drive spacecraft could get up to a noticeable fraction of c, given enough fuel, but we can't currently provide that (ion drives are amazingly efficient, but they still need two resources - reaction mass and power). Ramscoops aren't currently possible, if they'll even work (we're not sure they would). Our best reactors have a lifetime of decades, but those are Earth-based installations - current long-range spacecraft are powered by radioactive thermal decay, which can't produce enough power over a long enough time - we would need to use fission, at a minimum.
Solar sails might be usable, but they're currently pure sci-fi. Also, at least some of the things they theoretically could do - "tack" toward a star, for example - just don't work. Sailboats can tack for two reasons: air pressure sufficient to make airfoils work, and water viscosity being much greater than air viscosity. It's possible - I don't know the math - that an "airfoil" could work in solar wind, but I rather doubt it. However, the thrust vector of an airfoil is perpendicular to the wind direction. Sailboats have keels for two reasons - to keep them from flipping over, and to force them to move only forward and backward (meaning that the airfoil thrust need only be slightly forward of 90 degrees to get forward motion). Without something to "grip" like that, even a 100% efficient solar-foil could only orbit a star at constant distance.
Warp drive... now we're out of even the realm of things we can begin to experimentally prove. There is a theoretical mathematical model that, *if* our understanding of relativity is correct, permits moving faster than light (actually, moving at any arbitrary speed, given enough power). In short form, it consists of compressing space in front of the ship, and expanding it behind - the space in the middle, where the ship is, technically doesn't move at all. Now, the problems: We can't really compress space. In essense, we're talking gravity control here. In theory, with enough energy (or mass, they're related after all), it's possible. We dont' know how, though. We do not know how to expand space. Negative energy - not the same as anti-matter, but akin to anti-gravity - has been theorized and *maybe* observed, but we can't produce it at will, certainly not over any useful distance or magnitude. The theoretical power output required to be useful is, at minimum, far more than our entire race can produce. It might be more than the rest energy of the entire universe. You can't see where you're going. Anything that hits the "bow wave" gets compressed into a burst of gamma radiation. This includes photons. I trust I don't need to continue? Look it up if you want - it's a cool theiretical model - but I doubt we'll even know if it's possible within my lifetime.
Ok, how about spacecraft durability? At any kind of decent speed, a micrometeorite, even a spec of dust, could do incredible damage. Making the spacecraft more durable generally means adding mass, which decreases its acceleration and therefore max speed for a given amount of fuel. We don't currently have any kind of shield or navigational deflector that could block a rock of the size you use to skip stones - easily large enough to utterly destroy a spacecraft (if it is anything at all like what we can build today) travelling at the pitiful speed 0.01 c.
Finally, consider longevity. This is related to range, but worth its own discussion. Power, fuel, and durability have already been co
Avoiding software monoculture. On both OS X and Windows, Safari has been shown to have a substantial number of security flaws. Even if I liked its minimal configurability and general look and feel (I don't, but that's a personal thing) the security issues would lead me to avoid using it, much like IE6. (Hmm... is this some kind of rite of passage for a browser bundled with an OS? I hope Apple gets its security act in gear faster than Microsoft did - they're starting to become popular enough to be a worthwhile target.)
For the record, IE8 (even the first beta) used multiple processes to handle tabs. It doesn't have Chrome's nice "Task Manger" interface, but you can fairly easily tell which tab is slowing things down (by cycling through them, if nothing else) and can kill a slow or even completely hung tab without brining down the rest of the browser. I'm not sure which of Chrome and IE8 had this idea first (or if they got it from some common source) but it's a good one and could stand to be copied widely. Firefox in particular still completely explodes if any tab (or the UI, or the freaking download manager even) crashes - it's all a single process.
Yeah, the KDE developers (rightfully) caught a lot of crap for this. I can see their point of view leaving KDE4 as a pre-release until 4.2's release would have been a LONG development cycle - but it cost them (the devs) a lot of reputation. Personally, I don't really care about their rep, but then, I don't write KDE, I just use it. What really bothers me is that it cost KDE, as a desktop environment, a lot of its reputation too. KDE 4.0 / 4.1 made Vista look snappy and bug-free by comparison, and now KDE has to win back a lot of its old users. A shame, really.
Re:The same old question applies ...
on
KDE 4.2.4 Released
·
· Score: 2, Informative
I would say stability is a step up; KDE itself I've never had crash, but I get fewer application crashes in 4.2.x (4.2.3 on my current system, I'll upgrade shortly) than I did in 3.5.9 (didn't try.10). Features are a little harder, but I'd call that an improvement too. The desktop is nicer but... different. Tinker with it a little and I think you'll like it better, but it has changed. Most applicaitons have been ported across pretty straight, with the same features as before, but a few have had significant changes (Amarok), a few have only been developed on 4.x for a while now and have significantly newer versions there (though only incremental changes), and a few might not be fully ported yet.
Mostly trivial, but my favorite new 4.2 feature, small though it is: Konqueror now has an option to close a tab when you middle-click it (like EVER OTHER TABBED BROWSER IN EXISTENCE) which fixes one of what I found to be its biggest usablity quirks.
Re:I'll switch to KDE 4.x when Debian stable has i
on
KDE 4.2.4 Released
·
· Score: 2, Informative
By 4.2, nearly all KDE utilities and applications have been ported, and as of 4.2.3 nearly all the noticeable bugs were worked out (it worked better than 3.5.9, the last 3.x I used). Don't assume anything about 4.2 based on 4.1 or 4.0; both of those were released before they should have been,a dn should have been considered more like a tech/API preview (4.0) and early beta of a finished version (4.1). Frankly, they both sucked, and it has cost KDE a lot of reputation, but 4.2 is solid. It's what 4.x should have been from the beginning.
In other words, give 4.2 (or later) a try; they finally lived up to the promise of the earlier versions, plus the apps you're used to have all been ported now.
I've heard this a lot, but I can't say it matches my experiences. 3 years on a laptop with Win2000 and a very old ATi chip (7xxx series): no video-caused crashes. 2 years on a laptop with WinXP and an old-ish ATi chip (9xxx series): no video-caused crashes. 6 months on a Vista laptop with (now outdated) Mobility 200M: no video-related crashes, EVEN WITH BETA DRIVERS (for comparison, nVidia's Vista drivers weren't what I would call release quality until more than 6 months after Vista's release. ATi was there over three months before RTM). Granted, these aren't top-of-the-line desktop (or even the best of the laptop) lines, but honestly, where are these driver-caused crashes ATi supposedly has so much trouble with?
Re:KDE 4 looks promising
on
KDE 4.2.4 Released
·
· Score: 3, Informative
KDE 4.2 is perfectly usable. It's what 4 (in general) should ahve been from the start. Don't even bother looking at 4.1 or 4.0, and if you do, don't expect 4.2 to be the same. It's not. The older ones are broken, yes, but don't assume taht means that 4.x is *ALL* broken.
Regardless of what section it is filed under, it's also worth noting that KDE 4.x runs on Windows too. I'm not quite ready to suggest replacing WIndows Explorer entirely, but the apps and even the desktop work pretty well. That said, I would never have heard of KDE without trying Linux...
4.1 isn't even close to 4.2. You might as well compare a beta to a release version (think of it this way - 4.0 was the tech preview, incomplete and buggy but with APIs in place. 4.1 is the beta - many of the features but not all, and still buggy. 4.2 is release, with bugs fixed and features in place).
You'd think that talking about 4.1 in an article about 4.2.4 would be obviously absurd, but apparently not...
If you can run software on the device, then you don't need drivers, no. Mounting flash drives and the like with noexec would make this a lot more difficult. If they copy data off the flash drive to your system, then the evidence has been tampered with and should no longer be admissible in court. If they know enough to run mount (which may require your password) then they really don't need this "detector" anyhow.
If the computer is locked and doesn't *automatically* execute code on a USB, then they need your password anyhow. There are lots of simple ways to make this a LOT more difficult. On the flip side, FireWire (with its DMA) might work better here, assuming the DMA hasn't been disabled.
XP x64 drivers are hard to come by for a lot of hardware; most OEMs didn't bother with the OS, and while Vista can load most XP drivers, the reverse is not true (typically, the OS, not the driver, is backward-compatible).
Technically it's the same kernel as Server 2003, so if that works on your hardware (in x64) then XP x64 will too. That isn't something you can count on for a desktop machine, though.
You ask what more I could want. There's actually lots of answers to that, but probably the biggest and most relevant is automatic documentation and syntax assistance. When you're using an API that you've used many times before, this is of course unecessary, but often you need to work with APIs that are at least a little unfamiliar. In this case, even something as simple as a listing of a function's parameter types and names, and preferably its return type, can help a lot (moreso if it has the Windows API IN/OUT/OPTIONAL tags). Ideally, the IDE would also display the documentation for the function (doxygen-style or even just a single-line comment) - this is especially helpful with macros since they are much less self-documenting. Most IDEs have these features.
Automatic completion of variable names and such is nice, but often you need more than that. The core C/C++ libraries aren't so bad, but many third-party libraries are rich with overly-long class, type, function, or macro names. Remembering these can be a bitch, and most editors will only auto-complete names that you have already used in the "document" (code). By comparison, most IDEs will build up symbol tables from your libraries, etc. which makes it possible to find the name you want even if you've not used it before. Visual Studio's "Intellisense" will even determine, based on context and previous usage, which auto-complete options you're most likely looking for and will place them at the top.
Background syntax check/compile. Rather than writing a bunch of code, running make, waiting while it recompiles *everything* because somebody touched a high-level header, and eventually getting to your code only to have some 157 lines of mostly incomprehensible error referencing things that don't even appear to be on the line for which the line number is given, you can get a nice red underline exactly under the incorrect tokens, with an explanation of why it is incorrect and often even a simple way to fix the problem (yes, I meant to put "using SomeRandomNamespace;" at the top, thank you for the reminder). This is also an easy well to tell if you've fixed a large number of those 157 errors, or only one, or you fixed a bunch but you create 37 new ones... you get the idea. IDEs do this kind of check, and it is a lot more helpful than slogging through a bunch of build errors.
Long ago, MS had their own JVM, much like Apple does. Like Apple's, it was outdated and inferior. Nothing very good came of it, and it was A Good Thing when they dropped it.
However, even MS never told Sun *not* to develop Java for Windows. I'm not saying Sun is blameless here, but Apple's choice to take full responsibility for Java on the platform (followed by them dropping the ball) damages the platform and harms their users. Unlike Sun, they have a financial obligation to their users, and they aren't living up to it. I think bitching and moaning about that is pretty reasonable.
Admittedly, sudo is actually fairly easy to circumvent. With standard permissions, I can put a hidden script on your system that will *look* like sudo (and pass your input through to sudo, so from your end everything appears to work) but either change your PATH or set an alias (perhaps by changing.bashrc or.profile) so that my script runs instead when you type sudo (with a little more work, you could even subvert things like kdesu or the Gnome equivalent). That program - still operating only with user permissions (although it *could* now be root if it wanted) could send me (the attacker/malware author) your username, IP address, and password.
I suppose technically it is sort of social engineering, but it would still work quite well for gaining root access with only standard permissions, so long as waiting until the target user does something administrative is acceptable.
PulseAudio seems to be making some big strides in fixing sound issues, and I wish them well. There are definitely still problems, though. There's also the issue of feature lag; things like per-application volume controls worked three years ago while Vista was in beta. Back then, if I wanted sound to work at *all* with 3D hardware acceleration enabled (the proprietary driver was hosing ALSA, don't ask me how or why) I had to manually patch, recompile, and upgrade ALSA... and it would break on the next kernel update.
I agree that these should be integrated, even though I use them very sporadically even on Linux. There are some great apps that provide them for Windows, though - the better ones can't honestly be called "bloatware" either.
2. Mouse wheel works on item however, not item focused. You can also use the scrollwheel to cycle between desktops, tabs, windows, comboboxes and more.
User preference. The mouse driver controls this behavior, on either Windows or Linux. On Linux I haven't a clue how to change it, on Windows it's a setting on the Mouse control panel for most drivers (in my case, Synaptic touchpad - yet the setting applies to my MS Intellimouse too). Also, sometimes I *want* to be able to move the mouse button off the tiny combo box and still be able to scroll it using the wheel.
3. An 'always on top' item in the window menu.
Windows has no built-in equivalent, but I can't say I've ever found Always On Top to be desirable behavior anyhow. For the scenario you described, try dragging a window to the side of the screen in Win7 - it will automatically "half-maximize" to fill exactly half the screen, which makes it very easy to work with two programs simultaneously. Drag the window off the edge and it restores to its original size.
4. Middle click pasting.
FAR too application-dependent. Some things it works with, some things it doesn't do anything, some things it tends to break (Konqueror, until they added the option for middle-click to close a tab like it does in damn near every other tabbed app, let alone browser, out there). Besides, using the clipboard with an application that prefers the select/middle click approach is annoying. I'll stick with one unified approach, thanks.
5. Notifications that get out of the way.
You're comparing a 2009 OS to Windows XP (2001, or 2004 if you count SP2). In Vista, there are far fewer notification icons and they won't appear while in a full-screen app. Dialogs are also much better designed, both in terms of options presented and general layout. Win7 improves on both.
6. If virtualization is good enough for videogames on a Mac (it is), then it is good enough for videogames on a Linux.
The support lags behind Windows somewhat and it requires substantial amounts of RAM for the VM, but yes, there is progress here. I much prefer Wine's approach, however.
7. Dual booting.
Not sure what your point is; dual-boot is nothing new and certainly not a "feature" of Ubuntu. It is, however, a pain. As for security stuff, if you're connected to the 'net you really need a firewall, at least the built-in Windows one. If you know what you're doing and you don't run as Administrator all the time, you don't need AV anyhow.
8. Installing, uninstalling and updating applications.
Linux does package management very well. However, suppose your audio driver is acting up and you need to patch ALSA or something (this happens to me usually at least once a year / upgrade cycle): download ALSA tarball, extract it in a terminal, download a patch file, apply the patch, make sure you have kernel headers and working build toolchain, configure, compile, install. The average user would have trouble getting past step 2. I, personally, will happily deal with a lot of wizards to avoid that kind of thing. Finally, don't forget that you now have a non-standard software package that may break things when your package manager updates some related software (like a kernel update or similar). Also, in the last year or so MS has started using Microsoft Update to provide a place to download an dinstall new software as well.
9. Codecs.
Funny you should mention those. Half the codecs that Ventrilo likes to use still aren't available on Linux, and only *might* wor
I recommend using a version of Windows less than, oh, three years old. Complaining that your old XP media - even SP2 is old by the standards of the fast-moving tech industry - invalidates your claim of Linux readiness unless you were testing it on Ubuntu 4.10 or similar.
For what it's worth, my new Win7 RC install (clean of course, not upgrade) had working sound, wired/wireless/bluetooth networking, and webcam on first boot. One-touch keys and video were both partially functional; they (and other things) installed from Windows Update with no fuss aside from 1 reboot.
By comparison, my Linux box still won't suspend correctly, and the audio explodes if I plug in my USB headset. Webcam works less than 50% of the time and I can't even figure out why sometimes it does and sometimes it doesn't - it sill randomly stop working in the middle of a Skype video call. That distro *is* nearly a year old, though...
You forgot/ignored two, maybe three popular distros, so I'm really not sure your point holds *that* well.
OpenSuse: Top 3 (usually #2) on distrowatch for years now. Has a unified, helpful, and intuitive graphical configuration tool that I've yet to see an equivalent of on any other distro, although most of them try and some beat it at specific points. Commercial support and a major userbase means that it has an excellent supply of packages. My only major wish for it would be a more frequent update schedule - their updates are solid but slow coming.
Mandriva: You claim Mandrake "went bye-bye" but Mandriva, both as a distro and an organization, seems alive and well. They do the "lots of shiny" thing really well, including liveCDs with bundled proprietary video drivers for full hardware accelerated eye candy (F/OSS-only images are also available). They're also fairly good at ease-of-use - in fact, I've had more success converting colleagues with either Mandriva or openSuse than with Ubuntu. Top 10 on distrowatch and commercial support again mean that the software and user support are present.
Fedora: While certainly less user-friendliness-centric than the preceding two (or Ubuntu), for anybody who likes to tinker - even if they know nothing about Linux (I didn't, when I started), it's a great distro; a nice blend of power and control over the system, while not requiring that you know everything before you even start. It's also one of *the* big names in Linux, and posesses plenty of books and such about it - one of the advantages you cite for Ubuntu.
Essentially, that is actually what EVE does. All players are connected to the same server cluster, can talk to one another, etc. However, the varius star systems might be running on different physical machines, providing high performance within a given system (where it really matters if a given player can see another instanly, lest they not be friends) but allowing the effort of keeping track of the huge number of objects in a given system (asteroids, ships, NPCs, starbase components, missiles, fighters, deepspace complexes, and of course players) to be split by system. Low-occupancy systems may be combined on a single box, but high-activity regions have their own dedicated hardware. You can even tell CCP if you plan to have a big fight somewhere, and they'll put that node on a dedicated machine (although in practice people don't always send this warning).
Jumping between nodes is already part of the game (star gates, jump drives, etc; you can't simply fly from one system to another). There are a few things, like joining a fleet, that will prevent you from switchign systems for a few seconds. Every now and then an overloaded system will be closed to incoming traffic because it is overloaded, which will only affect people trying to make that jump.
Ironically enough, I've actually been known to do the opposite when beta-testing IE versions (7 and 8). With 7 it was rare, but by the time of 8 there were plenty of sites that would intentionally feed IE bad code (either in an attempt to be backward compatible to 5 or soemthing, or because they didn't like the browser). Using an IE plug-in, I would masquerade as Firefox or Safari to see how IE's standards mode handled the site. It was a strange sensation to see a site work/look *better* because I *pretended* to not be using IE.
Admittedly, such testing is a very minor portion of the market. There are probbly orders of magnitude more people spoofing themselves as IE than the reverse.
I don't know about OO.o (don't use it very often) but I would assume that the keyboard shortcuts are listed next to their commands, the way most GUI menus are.
What about on a Ribbon-style interface? Try hitting Alt sometime in Office 2007. A little label appears next to every icon and tab in the ribbon, showing the letter to press (either in sequence or chorded) to use that feature/display that tab. Where possible, the shortcuts are also the same as in previous versions of Office, and some effort went into making them intuitive too.
Double-click the tabs at the top of the ribbon, and the ribbon itself vanishes. Single click a tab to show it temporarily, like a menu. Double-click to get it to show permanently again.
This has existed since it was beta software called "Office 12." It's surprisingly poorly advertised; the behavior is logical but a little un-intuitive since most UIs don't do things like that, so a little user education would be smart.
Slashdot Mirror
Hmm, interesting. I can't say I've tested all of those, but...
Ftp works fine for me. For sftp I either use the console or use fish:, which also works.
Didn't try editing over a standard ftp, but fish works.
Don't use Samba, sorry. Even my Windows boxes have NFS (it's an optional feature built into in the higher Windows editions, believe it or not).
Fish works for me.
None of my networks use EAP. WPA-PSK works fantastically, much better than in 3.5.x.
I just turn off the Klipper applet and use it like copy-paste on any other platform. The only time I would even attempt to use Klipper is if I needed to transfer a large number of things from one window to another, but separately. In that case, I would usually use a sticky note widget or similar.
To my knowledge we have at least general knowledge of every major technology we would need to travel between stars
No offense, but WHAT?? There are so many problems with interstellar travel... lets consider a few, from the perspectives of spacecraft and tech available today.
Spacecraft range: The biggest single problem. Space is big. Really, really, REALLY big. The fastest spacecraft we currently posess would take centuries to reach Alpha Centauri (the next-closest star, a mere 4 light-years or so away). In theory, an ion-drive spacecraft could get up to a noticeable fraction of c, given enough fuel, but we can't currently provide that (ion drives are amazingly efficient, but they still need two resources - reaction mass and power). Ramscoops aren't currently possible, if they'll even work (we're not sure they would). Our best reactors have a lifetime of decades, but those are Earth-based installations - current long-range spacecraft are powered by radioactive thermal decay, which can't produce enough power over a long enough time - we would need to use fission, at a minimum.
Solar sails might be usable, but they're currently pure sci-fi. Also, at least some of the things they theoretically could do - "tack" toward a star, for example - just don't work. Sailboats can tack for two reasons: air pressure sufficient to make airfoils work, and water viscosity being much greater than air viscosity. It's possible - I don't know the math - that an "airfoil" could work in solar wind, but I rather doubt it. However, the thrust vector of an airfoil is perpendicular to the wind direction. Sailboats have keels for two reasons - to keep them from flipping over, and to force them to move only forward and backward (meaning that the airfoil thrust need only be slightly forward of 90 degrees to get forward motion). Without something to "grip" like that, even a 100% efficient solar-foil could only orbit a star at constant distance.
Warp drive... now we're out of even the realm of things we can begin to experimentally prove. There is a theoretical mathematical model that, *if* our understanding of relativity is correct, permits moving faster than light (actually, moving at any arbitrary speed, given enough power). In short form, it consists of compressing space in front of the ship, and expanding it behind - the space in the middle, where the ship is, technically doesn't move at all. Now, the problems:
We can't really compress space. In essense, we're talking gravity control here. In theory, with enough energy (or mass, they're related after all), it's possible. We dont' know how, though.
We do not know how to expand space. Negative energy - not the same as anti-matter, but akin to anti-gravity - has been theorized and *maybe* observed, but we can't produce it at will, certainly not over any useful distance or magnitude.
The theoretical power output required to be useful is, at minimum, far more than our entire race can produce. It might be more than the rest energy of the entire universe.
You can't see where you're going. Anything that hits the "bow wave" gets compressed into a burst of gamma radiation. This includes photons.
I trust I don't need to continue? Look it up if you want - it's a cool theiretical model - but I doubt we'll even know if it's possible within my lifetime.
Ok, how about spacecraft durability? At any kind of decent speed, a micrometeorite, even a spec of dust, could do incredible damage. Making the spacecraft more durable generally means adding mass, which decreases its acceleration and therefore max speed for a given amount of fuel. We don't currently have any kind of shield or navigational deflector that could block a rock of the size you use to skip stones - easily large enough to utterly destroy a spacecraft (if it is anything at all like what we can build today) travelling at the pitiful speed 0.01 c.
Finally, consider longevity. This is related to range, but worth its own discussion. Power, fuel, and durability have already been co
Avoiding software monoculture. On both OS X and Windows, Safari has been shown to have a substantial number of security flaws. Even if I liked its minimal configurability and general look and feel (I don't, but that's a personal thing) the security issues would lead me to avoid using it, much like IE6. (Hmm... is this some kind of rite of passage for a browser bundled with an OS? I hope Apple gets its security act in gear faster than Microsoft did - they're starting to become popular enough to be a worthwhile target.)
For the record, IE8 (even the first beta) used multiple processes to handle tabs. It doesn't have Chrome's nice "Task Manger" interface, but you can fairly easily tell which tab is slowing things down (by cycling through them, if nothing else) and can kill a slow or even completely hung tab without brining down the rest of the browser. I'm not sure which of Chrome and IE8 had this idea first (or if they got it from some common source) but it's a good one and could stand to be copied widely. Firefox in particular still completely explodes if any tab (or the UI, or the freaking download manager even) crashes - it's all a single process.
Yeah, the KDE developers (rightfully) caught a lot of crap for this. I can see their point of view leaving KDE4 as a pre-release until 4.2's release would have been a LONG development cycle - but it cost them (the devs) a lot of reputation. Personally, I don't really care about their rep, but then, I don't write KDE, I just use it. What really bothers me is that it cost KDE, as a desktop environment, a lot of its reputation too. KDE 4.0 / 4.1 made Vista look snappy and bug-free by comparison, and now KDE has to win back a lot of its old users. A shame, really.
I would say stability is a step up; KDE itself I've never had crash, but I get fewer application crashes in 4.2.x (4.2.3 on my current system, I'll upgrade shortly) than I did in 3.5.9 (didn't try .10).
Features are a little harder, but I'd call that an improvement too. The desktop is nicer but... different. Tinker with it a little and I think you'll like it better, but it has changed. Most applicaitons have been ported across pretty straight, with the same features as before, but a few have had significant changes (Amarok), a few have only been developed on 4.x for a while now and have significantly newer versions there (though only incremental changes), and a few might not be fully ported yet.
Mostly trivial, but my favorite new 4.2 feature, small though it is: Konqueror now has an option to close a tab when you middle-click it (like EVER OTHER TABBED BROWSER IN EXISTENCE) which fixes one of what I found to be its biggest usablity quirks.
By 4.2, nearly all KDE utilities and applications have been ported, and as of 4.2.3 nearly all the noticeable bugs were worked out (it worked better than 3.5.9, the last 3.x I used). Don't assume anything about 4.2 based on 4.1 or 4.0; both of those were released before they should have been,a dn should have been considered more like a tech/API preview (4.0) and early beta of a finished version (4.1). Frankly, they both sucked, and it has cost KDE a lot of reputation, but 4.2 is solid. It's what 4.x should have been from the beginning.
In other words, give 4.2 (or later) a try; they finally lived up to the promise of the earlier versions, plus the apps you're used to have all been ported now.
It means that he's a Slashdot admin or employee or something of that nature. Note the text at the top of the page:
Posted by timothy on Wednesday June 03, @01:43PM
from the smoothing-a-smoothie dept.
CmdrTaco and all the rest have it too.
I've heard this a lot, but I can't say it matches my experiences.
3 years on a laptop with Win2000 and a very old ATi chip (7xxx series): no video-caused crashes.
2 years on a laptop with WinXP and an old-ish ATi chip (9xxx series): no video-caused crashes.
6 months on a Vista laptop with (now outdated) Mobility 200M: no video-related crashes, EVEN WITH BETA DRIVERS (for comparison, nVidia's Vista drivers weren't what I would call release quality until more than 6 months after Vista's release. ATi was there over three months before RTM).
Granted, these aren't top-of-the-line desktop (or even the best of the laptop) lines, but honestly, where are these driver-caused crashes ATi supposedly has so much trouble with?
KDE 4.2 is perfectly usable. It's what 4 (in general) should ahve been from the start. Don't even bother looking at 4.1 or 4.0, and if you do, don't expect 4.2 to be the same. It's not. The older ones are broken, yes, but don't assume taht means that 4.x is *ALL* broken.
Regardless of what section it is filed under, it's also worth noting that KDE 4.x runs on Windows too. I'm not quite ready to suggest replacing WIndows Explorer entirely, but the apps and even the desktop work pretty well. That said, I would never have heard of KDE without trying Linux...
4.1 isn't even close to 4.2. You might as well compare a beta to a release version (think of it this way - 4.0 was the tech preview, incomplete and buggy but with APIs in place. 4.1 is the beta - many of the features but not all, and still buggy. 4.2 is release, with bugs fixed and features in place).
You'd think that talking about 4.1 in an article about 4.2.4 would be obviously absurd, but apparently not...
If you can run software on the device, then you don't need drivers, no. Mounting flash drives and the like with noexec would make this a lot more difficult. If they copy data off the flash drive to your system, then the evidence has been tampered with and should no longer be admissible in court. If they know enough to run mount (which may require your password) then they really don't need this "detector" anyhow.
If the computer is locked and doesn't *automatically* execute code on a USB, then they need your password anyhow. There are lots of simple ways to make this a LOT more difficult. On the flip side, FireWire (with its DMA) might work better here, assuming the DMA hasn't been disabled.
XP x64 drivers are hard to come by for a lot of hardware; most OEMs didn't bother with the OS, and while Vista can load most XP drivers, the reverse is not true (typically, the OS, not the driver, is backward-compatible).
Technically it's the same kernel as Server 2003, so if that works on your hardware (in x64) then XP x64 will too. That isn't something you can count on for a desktop machine, though.
You ask what more I could want. There's actually lots of answers to that, but probably the biggest and most relevant is automatic documentation and syntax assistance. When you're using an API that you've used many times before, this is of course unecessary, but often you need to work with APIs that are at least a little unfamiliar. In this case, even something as simple as a listing of a function's parameter types and names, and preferably its return type, can help a lot (moreso if it has the Windows API IN/OUT/OPTIONAL tags). Ideally, the IDE would also display the documentation for the function (doxygen-style or even just a single-line comment) - this is especially helpful with macros since they are much less self-documenting. Most IDEs have these features.
Automatic completion of variable names and such is nice, but often you need more than that. The core C/C++ libraries aren't so bad, but many third-party libraries are rich with overly-long class, type, function, or macro names. Remembering these can be a bitch, and most editors will only auto-complete names that you have already used in the "document" (code). By comparison, most IDEs will build up symbol tables from your libraries, etc. which makes it possible to find the name you want even if you've not used it before. Visual Studio's "Intellisense" will even determine, based on context and previous usage, which auto-complete options you're most likely looking for and will place them at the top.
Background syntax check/compile. Rather than writing a bunch of code, running make, waiting while it recompiles *everything* because somebody touched a high-level header, and eventually getting to your code only to have some 157 lines of mostly incomprehensible error referencing things that don't even appear to be on the line for which the line number is given, you can get a nice red underline exactly under the incorrect tokens, with an explanation of why it is incorrect and often even a simple way to fix the problem (yes, I meant to put "using SomeRandomNamespace;" at the top, thank you for the reminder). This is also an easy well to tell if you've fixed a large number of those 157 errors, or only one, or you fixed a bunch but you create 37 new ones... you get the idea. IDEs do this kind of check, and it is a lot more helpful than slogging through a bunch of build errors.
Long ago, MS had their own JVM, much like Apple does. Like Apple's, it was outdated and inferior. Nothing very good came of it, and it was A Good Thing when they dropped it.
However, even MS never told Sun *not* to develop Java for Windows. I'm not saying Sun is blameless here, but Apple's choice to take full responsibility for Java on the platform (followed by them dropping the ball) damages the platform and harms their users. Unlike Sun, they have a financial obligation to their users, and they aren't living up to it. I think bitching and moaning about that is pretty reasonable.
Admittedly, sudo is actually fairly easy to circumvent. With standard permissions, I can put a hidden script on your system that will *look* like sudo (and pass your input through to sudo, so from your end everything appears to work) but either change your PATH or set an alias (perhaps by changing .bashrc or .profile) so that my script runs instead when you type sudo (with a little more work, you could even subvert things like kdesu or the Gnome equivalent). That program - still operating only with user permissions (although it *could* now be root if it wanted) could send me (the attacker/malware author) your username, IP address, and password.
I suppose technically it is sort of social engineering, but it would still work quite well for gaining root access with only standard permissions, so long as waiting until the target user does something administrative is acceptable.
PulseAudio seems to be making some big strides in fixing sound issues, and I wish them well. There are definitely still problems, though. There's also the issue of feature lag; things like per-application volume controls worked three years ago while Vista was in beta. Back then, if I wanted sound to work at *all* with 3D hardware acceleration enabled (the proprietary driver was hosing ALSA, don't ask me how or why) I had to manually patch, recompile, and upgrade ALSA... and it would break on the next kernel update.
1. Multiple desktops.
I agree that these should be integrated, even though I use them very sporadically even on Linux. There are some great apps that provide them for Windows, though - the better ones can't honestly be called "bloatware" either.
2. Mouse wheel works on item however, not item focused. You can also use the scrollwheel to cycle between desktops, tabs, windows, comboboxes and more.
User preference. The mouse driver controls this behavior, on either Windows or Linux. On Linux I haven't a clue how to change it, on Windows it's a setting on the Mouse control panel for most drivers (in my case, Synaptic touchpad - yet the setting applies to my MS Intellimouse too). Also, sometimes I *want* to be able to move the mouse button off the tiny combo box and still be able to scroll it using the wheel.
3. An 'always on top' item in the window menu.
Windows has no built-in equivalent, but I can't say I've ever found Always On Top to be desirable behavior anyhow. For the scenario you described, try dragging a window to the side of the screen in Win7 - it will automatically "half-maximize" to fill exactly half the screen, which makes it very easy to work with two programs simultaneously. Drag the window off the edge and it restores to its original size.
4. Middle click pasting.
FAR too application-dependent. Some things it works with, some things it doesn't do anything, some things it tends to break (Konqueror, until they added the option for middle-click to close a tab like it does in damn near every other tabbed app, let alone browser, out there). Besides, using the clipboard with an application that prefers the select/middle click approach is annoying. I'll stick with one unified approach, thanks.
5. Notifications that get out of the way.
You're comparing a 2009 OS to Windows XP (2001, or 2004 if you count SP2). In Vista, there are far fewer notification icons and they won't appear while in a full-screen app. Dialogs are also much better designed, both in terms of options presented and general layout. Win7 improves on both.
6. If virtualization is good enough for videogames on a Mac (it is), then it is good enough for videogames on a Linux.
The support lags behind Windows somewhat and it requires substantial amounts of RAM for the VM, but yes, there is progress here. I much prefer Wine's approach, however.
7. Dual booting.
Not sure what your point is; dual-boot is nothing new and certainly not a "feature" of Ubuntu. It is, however, a pain. As for security stuff, if you're connected to the 'net you really need a firewall, at least the built-in Windows one. If you know what you're doing and you don't run as Administrator all the time, you don't need AV anyhow.
8. Installing, uninstalling and updating applications.
Linux does package management very well. However, suppose your audio driver is acting up and you need to patch ALSA or something (this happens to me usually at least once a year / upgrade cycle): download ALSA tarball, extract it in a terminal, download a patch file, apply the patch, make sure you have kernel headers and working build toolchain, configure, compile, install. The average user would have trouble getting past step 2. I, personally, will happily deal with a lot of wizards to avoid that kind of thing. Finally, don't forget that you now have a non-standard software package that may break things when your package manager updates some related software (like a kernel update or similar). Also, in the last year or so MS has started using Microsoft Update to provide a place to download an dinstall new software as well.
9. Codecs.
Funny you should mention those. Half the codecs that Ventrilo likes to use still aren't available on Linux, and only *might* wor
I recommend using a version of Windows less than, oh, three years old. Complaining that your old XP media - even SP2 is old by the standards of the fast-moving tech industry - invalidates your claim of Linux readiness unless you were testing it on Ubuntu 4.10 or similar.
For what it's worth, my new Win7 RC install (clean of course, not upgrade) had working sound, wired/wireless/bluetooth networking, and webcam on first boot. One-touch keys and video were both partially functional; they (and other things) installed from Windows Update with no fuss aside from 1 reboot.
By comparison, my Linux box still won't suspend correctly, and the audio explodes if I plug in my USB headset. Webcam works less than 50% of the time and I can't even figure out why sometimes it does and sometimes it doesn't - it sill randomly stop working in the middle of a Skype video call. That distro *is* nearly a year old, though...
You forgot/ignored two, maybe three popular distros, so I'm really not sure your point holds *that* well.
OpenSuse: Top 3 (usually #2) on distrowatch for years now. Has a unified, helpful, and intuitive graphical configuration tool that I've yet to see an equivalent of on any other distro, although most of them try and some beat it at specific points. Commercial support and a major userbase means that it has an excellent supply of packages. My only major wish for it would be a more frequent update schedule - their updates are solid but slow coming.
Mandriva: You claim Mandrake "went bye-bye" but Mandriva, both as a distro and an organization, seems alive and well. They do the "lots of shiny" thing really well, including liveCDs with bundled proprietary video drivers for full hardware accelerated eye candy (F/OSS-only images are also available). They're also fairly good at ease-of-use - in fact, I've had more success converting colleagues with either Mandriva or openSuse than with Ubuntu. Top 10 on distrowatch and commercial support again mean that the software and user support are present.
Fedora: While certainly less user-friendliness-centric than the preceding two (or Ubuntu), for anybody who likes to tinker - even if they know nothing about Linux (I didn't, when I started), it's a great distro; a nice blend of power and control over the system, while not requiring that you know everything before you even start. It's also one of *the* big names in Linux, and posesses plenty of books and such about it - one of the advantages you cite for Ubuntu.
Essentially, that is actually what EVE does. All players are connected to the same server cluster, can talk to one another, etc. However, the varius star systems might be running on different physical machines, providing high performance within a given system (where it really matters if a given player can see another instanly, lest they not be friends) but allowing the effort of keeping track of the huge number of objects in a given system (asteroids, ships, NPCs, starbase components, missiles, fighters, deepspace complexes, and of course players) to be split by system. Low-occupancy systems may be combined on a single box, but high-activity regions have their own dedicated hardware. You can even tell CCP if you plan to have a big fight somewhere, and they'll put that node on a dedicated machine (although in practice people don't always send this warning).
Jumping between nodes is already part of the game (star gates, jump drives, etc; you can't simply fly from one system to another). There are a few things, like joining a fleet, that will prevent you from switchign systems for a few seconds. Every now and then an overloaded system will be closed to incoming traffic because it is overloaded, which will only affect people trying to make that jump.
Ironically enough, I've actually been known to do the opposite when beta-testing IE versions (7 and 8). With 7 it was rare, but by the time of 8 there were plenty of sites that would intentionally feed IE bad code (either in an attempt to be backward compatible to 5 or soemthing, or because they didn't like the browser). Using an IE plug-in, I would masquerade as Firefox or Safari to see how IE's standards mode handled the site. It was a strange sensation to see a site work/look *better* because I *pretended* to not be using IE.
Admittedly, such testing is a very minor portion of the market. There are probbly orders of magnitude more people spoofing themselves as IE than the reverse.
I don't know about OO.o (don't use it very often) but I would assume that the keyboard shortcuts are listed next to their commands, the way most GUI menus are.
What about on a Ribbon-style interface? Try hitting Alt sometime in Office 2007. A little label appears next to every icon and tab in the ribbon, showing the letter to press (either in sequence or chorded) to use that feature/display that tab. Where possible, the shortcuts are also the same as in previous versions of Office, and some effort went into making them intuitive too.
Double-click the tabs at the top of the ribbon, and the ribbon itself vanishes. Single click a tab to show it temporarily, like a menu. Double-click to get it to show permanently again.
This has existed since it was beta software called "Office 12." It's surprisingly poorly advertised; the behavior is logical but a little un-intuitive since most UIs don't do things like that, so a little user education would be smart.