And even when you've done that, and somehow worked out what bits of the Registry you need, there's absolutely no guarantee that by putting that chunk of the registry on another machine or another user, that you won't just trash the whole thing.
I've done this before and I know for a fact that it's incorrect. A registry edit will very rarely bring down a system, even if a major disaster (such as bad clusters where the registry is saved to the disk) will.
By comparison, knowing on Linux that your configuration file is likely to be a text file somewhere in/etc or in your home directory and maybe having to read the man page to find out precisely the name and location of the file, is infinitely easier.
With respect, that's irrelevant. We're talking about the end user here, and he shouldn't have to rootle around in the hidden files in his home directory and in/etc to look for his files.
The problem is that you're looking at it from a developer's point of view, which we need to get out of the habit of doing. The user wants the configuration file to... well... configure itself, and leave him alone. If he does have to change it, it should be easy (i.e. point-and-click easy) to do so.
No thanks. Use a proprietary software if you want your freedom to crumble.
This is exactly what I mean by preaching by the FOSS community. The user doesn't give a shit about freedom. He wants it to work and get out of his way.
Yet nobody complains when their mac doesn't work with whatever device, or whatever software, they just buy new ones or go mute. Is it because we're free that we need to do more? =/
People do complain about Macs failing, but Macs fail less frequently, because they're optimised to run on Apple's own hardware. Therefore, the OS knows exactly what hardware it'll be running on and works perfectly with it.
The major thing putting people off Linux is its lack of hardware support, and shit documentation and fixes for it. As an example, I tried installing Linux Mint on a Toshiba X1200(?) notebook last week. Turns out the only way to get the Realtek wireless card working is to download a tarball, untar it, run./configure, sudo make and sudo make install before finally configuring the kernel to use the new extension. The end user shouldn't have to do that, and if the FOSS community weren't so shitty about supporting proprietary drivers, it might improve users' experiences. All that had to happen was for it to be bundled as a nice little.deb file which could be double-clicked on, gksudoed, and installed in a jiffy.
I also think its funny that you like the cut down Gnome and OSX interfaces and yet demand that "every feature a user could possibly want" is included. If that happened those interfaces would be unusably cluttered.
The fact is that both these DEs aren't as cut-down as you think - the switches and knobs are there when you need them, invisible when you don't. As an example, one can use OS X but never run fsck - if he/she wants to, it's a quick trip to/Applications/Utilities/Disk Utility.app, which can still be easily accessed through a GUI.
The myth of OS X's featurelessness and rigidity, it seems, comes from the old days of OS 9 when the OS would say you don't need to know, don't want to know, and would you kindly return to your seat and watch the movie, and OS X 10.0, which was, to be fair, one of the most buggy mainstream OSes ever. (It was practically a beta, still.)
Just one observation. I agree about the flakiness of some of the GUI front ends for text file configuration but the fact is there are countless web pages, man pages & sample configuration files around that will guide you through a lot of this stuff if you do a bit of reading first.
But you forget that the documentation is often written with sysadmins in mind, and there's almost no step-by-step stuff available. Ask the average user to "edit xorg.conf as root" and he'll be completely befuddled. Communities also need to improve, and while some are very helpful (Ubuntu's has been very helpful when I've had my own problems) others are very hostile and simply say "you **** need to learn how to use a ****ing computer".
Whatever happens, there'll be more people using Linux-based OSes in 3/4 years. However, there's a long way to go before it can properly overtake Windows, IMO: there are several major problems that stop GNU/Linux becoming the ideal consumer desktop OS.
Linux still uses the outmoded FHS at the front end - I know I'll get flamed for this, but look at the other options. Windows installs end-user applications in/Program Files, dependencies either in the application folders or in/Windows/system32, the documentation and resources within the application folders and the rest of the core system files in/Windows, and puts all the home folders in/Users. I like Mac OS X's system even more: end-user apps live in self-contained packages (sometimes with dependencies) in/Applications, the home folders live in/Users, system resources live in/System, and dependencies (if not contained in the packages) live in/Library (or ~/Library). The core toolchain lives in the old FHS system for backwards compatibility. I suggest adopting a similar approach to GoboLinux, but sticking non-'end-user' programs such as ls, sh, etc in/System/CoreToolchain or something along those lines. Libraries should live in/Library or something like that, users under/Home or/Users, and end-user applications (like OpenOffice.org) in/Applications or/Programs. On first run, these apps unpack and install their dependencies, and then sit in/Applications, ready to be launched at the click of a button. There could even be an option for System V zealots to retain the old FHS at install.
Linux needs to stop preaching about free software and get back to work - in the end, users don't give a damn whether or not they can 'modify' this new driver they're installing to get compiz to work. As long as it's free as in beer and it works, they're happy. At most, an EULA dialog should be thrown up.
Users don't want to be rebuffed with the old 'that's the great thing about FOSS - if there's something you don't like, you fix it yourself' line - in the end, the user wants to press a big red "INSTALL" button, get a coffee while his system configures itself, and then wants it to get out of his way and let him work. This is why I believe KDE is unsuitable for consumer use as yet - as a power user's DE it's great, but it has a switch and a knob for everything. GNOME, on the other hand, is elegant, speedy, and gets out of your way and lets you get on with your work. Look at the other options: first setup and configuration of a new iMac, say, is a fifteen-minute job. You can literally pick one up on your way to work and have it working perfectly by the morning coffee break. In short, developers need to drop the "someone else'll sort that out later, I can't be arsed" attitude - they must be diligent and include every feature a user could possibly want and more besides.
In short, devs need to really get their fingers out and concentrate on creating a truly kick-ass operating system that'll work out of the box on practically any machine you throw it at. This is what led Apple out of its slump in the mid-90s - if the FOSS community can do it now, when the popularity of FOSS is booming, it will truly be a force for Monkey Boy to reckon with.
There was a small segment that looked like some fireworks had been composited over the Beijing landscape. It doesn't matter. It was only used because it would have been dangerous to film from a helicopter, and the display that actually took place was identical to the CGI one.
That's different. Windows is a PC operating system, designed to be a fully-functional general-purpose computer. The iPhone is a phone. Apps are only allowed on to the machine after being given the green light by Apple, so it's feasible that a system could be put in place.
I must agree. I didn't say I liked COBOL: I'm simply saying it's very simple, because it has a very English-like syntax. However, I find its typing style laborious and slow.
That is not to say that it's not an incredibly simple language to pick up, and practically anyone could learn to code a basic functional program in it within the hour. I do think that there are going to be people who like COBOL who would be quite happy to code in it, for that amount of money, despite the fact that for many developers it is a cumbersome dinosaur of a language.
There are many people I know who were programmers in yesteryear, and love COBOL because of its English-like syntax. So I do think that COBOL could make a comeback (albeit indirectly) because of its simplicity.
COBOL could easily make a comeback, because it's so damn easy to learn. I could train a cuttlefish to write a complex accounts program in COBOL. Heck, I bet even Mike Huckabee could probably manage to work out "Hello World".
Perhaps the reason it may come back is because of the complexity of modern languages. Most OOP variants are horrifically difficult to understand, and a return to the statement-based languages of yesteryear might actually be a good thing: let the coder get on with writing his code, and let the interface builder do all the fancy OO stuff.
True, vim has a rather good tutorial, even if I'd prefer it to be along the lines of the 'old-style' tutorials.
I believe some distributions of Windows XP had a 'step-by-step' learning system which had manuals, instructional videos, and interactive tutorials. You were actually given a mock-up of the window, and guided through it keystroke-by-keystroke.
Vi, once you've got used to it, has a wonderfully powerful interface and becomes quite intuitive when you learn new commands. The problem is that someone who's never touched a computer before is unlikely to think "right... perhaps if I try pressing 'colon'... ah!" This also means they'd not be able to get into the tutorial. It is intuitive to people who've used it before, but it's not intuitive if you've never seen it before.
On the other hand, if they see a mouse and that moving the mouse moves that little arrow on the screen, they might figure that clicking the mouse will 'press' what the arrow is on. (This is even more intuitive on one-button mice, but even two/three/four-button mice are speedy enough to pick up.)
Oh, vi is easy. Colon, Q, enter. You're out. If it's not been saved, use:wq! to save or:q! to discard.
That said, they're both very counter-intuitive. The user doesn't want to have to patch their software. They want it to just work, and in an intuitive and sensible way. I learnt to compute on a Dell workstation with Windows 3.11, MS-DOS 5 (or 6) and Works 2, and I am entirely self-taught. Those interfaces were more intuitive than some FOSS interfaces around today, which is why I was able to sniff my way around and orientate myself quite quickly.
Something that might solve a lot of problems is better documentation, with proper tutorials. The tutorials in Works and Windows 3 were fantastic in that they explained things very simply, and were interactive: i.e. you learned by doing.
I'm using ume-launcher (the Netbook Remix launcher) on my Eee PC 701 right now, and it really isn't bad at all. It's still quite buggy though:
Clutter has a few problems, I think
It's impossible to edit the menus (I think it reads off the Debian menus file)
Sometimes it works after resuming from a suspend to RAM, sometimes it doesn't
Sometimes it works after switching back to tty7 from a text terminal, sometimes it doesn't
Apart from that, it's very efficient, and either way it pwns Asus's default Eee launcher: it's prettier, less resource-intensive and more space-efficient.
I agree about PMs versus Windows, but not versus OS X. Most of the time that simply involves dragging something to the applications folder and starting it - dead easy. A cucumber could do it.
OS X also has a package manager - of sorts. It's similar to dpkg in the way it's constructed, in that it's only a package format that also provides logging in the form of package receipts, and installation and configuration by wizards and post-flight scripts. In both cases, though, it's nice and easy to uninstall an app - all one has to do is drag it to the Trash.
I suppose it's easier for OS X to do this for two reasons:
Most OS X apps are self-contained, with very few dependencies. If there are dependencies not catered for by the Carbon or Cocoa APIs, they're included in the code.
All end user applications (Firefox, Safari, iWork, iPhoto, iTunes, etc.) sit in/Applications instead of in/usr/share/bin (or any of the million and one other folders apt-get could dump them in).
But the key introduces a critical weakness to the system: it's small, so it's easy to lose. I still think that sitting the OS on internal storage would make things better, and provide some kind of insurance policy against losing the key.
On a personal note, IMHO Mandriva is maybe the best linux distro, by far, give-it a try.
I recall reading a comment here, not long ago, stating that dealing with mobile providers in the US was like 'choosing between shit sandwiches'. This, unfortunately, seems to reiterate that this is the case.
Over here in the UK, things are a lot better. If you don't like customer service, you change network. Simple as. I switched from Tesco Mobile (poor customer service on O2's otherwise excellent network) to 3 earlier this year, and the process was quite painless. Am I right in saying that having mobiles on contract is more common in the USA which is what makes dealing with providers such a nightmare?
In this case, it means that losing the key, the smallest component, would mean the computer becomes useless. In that case, it would be better to have the base OS on the computer's ROM, especially because it's easier to lose a USB key than a notebook.
Also, I disagree with your statement about Emtec being a pioneer in this respect. The ultra-sub-notebook/netbook market is generally split into a high-end and low-end - the low end is dominated by the Asus Eee PC, and the high end has been practically cornered by the MacBook Air. Both of these machines have been selling like hotcakes, and the others were all somewhat late into the market. Emtec is, therefore, going to be very late onto the market at this rate. It's like the iPod - it was an early example of an HDD music player which was one of the first which was a truly usable concept. Most HDD players before that were, TBH, slow and rubbish. It got in early, and it's still going strong to this day: the same, I think, will go for the Eee and the Air.
As open-source software and OSes go, it may be pioneering, but I think as a computer and production OS the concept is utterly flawed, and they've entered the market very, very late.
As I understand this, the root partition is stored on the machine's internal ROM. The user's home folder sits on the USB key, along with something that somehow links to/etc/passwd to provide authentication. The key is automatically mounted when inserted. Correct me if I'm wrong.
This doesn't sound particularly promising - it would be very easy to lose the key. I also fail to see why, when most ultra-sub-notebooks are bought by a person for their use, and their use only.
Also, will the home folder on the key be accessible when plugging into another computer, say, a desktop running OS X, Windows or another Linux distro? If so, it would kind-of defeat the object. Emtec would be entering the market very late, so they can't expect this to take the market by storm. If it doesn't, it kind-of defeats the object of sticking everything on a USB key.
It's possible to boot some *nixes from a RAM image - like when resuming from hibernation/Suspend to Disk. I suggest booting the actual OS from the image and putting the home folder on a different partition in the hard drive.
The advantage of this is that it bypasses init (or upstart, sysvinit, launchd or whichever startup daemon you're using) and boots things into a state which works immediately.
It's not so much handling, although Fedora's handling of wpa-psk is atrocious, believe me.
The problem is driver compatibility - at the moment, it is pretty much hit and miss whether or not your wireless card will work. Ironically, often the only way to obtain the required driver will be from a multiverse repository - by connecting to the Internet.
Slashdot Mirror
And even when you've done that, and somehow worked out what bits of the Registry you need, there's absolutely no guarantee that by putting that chunk of the registry on another machine or another user, that you won't just trash the whole thing.
I've done this before and I know for a fact that it's incorrect. A registry edit will very rarely bring down a system, even if a major disaster (such as bad clusters where the registry is saved to the disk) will.
By comparison, knowing on Linux that your configuration file is likely to be a text file somewhere in /etc or in your home directory and maybe having to read the man page to find out precisely the name and location of the file, is infinitely easier.
With respect, that's irrelevant. We're talking about the end user here, and he shouldn't have to rootle around in the hidden files in his home directory and in /etc to look for his files.
The problem is that you're looking at it from a developer's point of view, which we need to get out of the habit of doing. The user wants the configuration file to... well... configure itself, and leave him alone. If he does have to change it, it should be easy (i.e. point-and-click easy) to do so.
No thanks. Use a proprietary software if you want your freedom to crumble.
This is exactly what I mean by preaching by the FOSS community. The user doesn't give a shit about freedom. He wants it to work and get out of his way.
Yet nobody complains when their mac doesn't work with whatever device, or whatever software, they just buy new ones or go mute. Is it because we're free that we need to do more? =/
People do complain about Macs failing, but Macs fail less frequently, because they're optimised to run on Apple's own hardware. Therefore, the OS knows exactly what hardware it'll be running on and works perfectly with it.
The major thing putting people off Linux is its lack of hardware support, and shit documentation and fixes for it. As an example, I tried installing Linux Mint on a Toshiba X1200(?) notebook last week. Turns out the only way to get the Realtek wireless card working is to download a tarball, untar it, run ./configure, sudo make and sudo make install before finally configuring the kernel to use the new extension. The end user shouldn't have to do that, and if the FOSS community weren't so shitty about supporting proprietary drivers, it might improve users' experiences. All that had to happen was for it to be bundled as a nice little .deb file which could be double-clicked on, gksudoed, and installed in a jiffy.
I also think its funny that you like the cut down Gnome and OSX interfaces and yet demand that "every feature a user could possibly want" is included. If that happened those interfaces would be unusably cluttered.
The fact is that both these DEs aren't as cut-down as you think - the switches and knobs are there when you need them, invisible when you don't. As an example, one can use OS X but never run fsck - if he/she wants to, it's a quick trip to /Applications/Utilities/Disk Utility.app, which can still be easily accessed through a GUI.
The myth of OS X's featurelessness and rigidity, it seems, comes from the old days of OS 9 when the OS would say you don't need to know, don't want to know, and would you kindly return to your seat and watch the movie, and OS X 10.0, which was, to be fair, one of the most buggy mainstream OSes ever. (It was practically a beta, still.)
Just one observation. I agree about the flakiness of some of the GUI front ends for text file configuration but the fact is there are countless web pages, man pages & sample configuration files around that will guide you through a lot of this stuff if you do a bit of reading first.
But you forget that the documentation is often written with sysadmins in mind, and there's almost no step-by-step stuff available. Ask the average user to "edit xorg.conf as root" and he'll be completely befuddled. Communities also need to improve, and while some are very helpful (Ubuntu's has been very helpful when I've had my own problems) others are very hostile and simply say "you **** need to learn how to use a ****ing computer".
Whatever happens, there'll be more people using Linux-based OSes in 3/4 years. However, there's a long way to go before it can properly overtake Windows, IMO: there are several major problems that stop GNU/Linux becoming the ideal consumer desktop OS.
In short, devs need to really get their fingers out and concentrate on creating a truly kick-ass operating system that'll work out of the box on practically any machine you throw it at. This is what led Apple out of its slump in the mid-90s - if the FOSS community can do it now, when the popularity of FOSS is booming, it will truly be a force for Monkey Boy to reckon with.
I always use an answer which is memorable, but completely irrelevant to the question. For example:
QUESTION: What was the name of your first school teacher?
ANSWER: The Handle from a Power Mac G3.
Disclaimer: this is just something I made up off the top of my head. It's not in use. If you try to use it to steal my identity, you're bonkers.
There was a small segment that looked like some fireworks had been composited over the Beijing landscape. It doesn't matter. It was only used because it would have been dangerous to film from a helicopter, and the display that actually took place was identical to the CGI one.
That's different. Windows is a PC operating system, designed to be a fully-functional general-purpose computer. The iPhone is a phone. Apps are only allowed on to the machine after being given the green light by Apple, so it's feasible that a system could be put in place.
They were working for a French computer security journal. Sort of like ZDNet, Linux Format, PC World etc, but with a heavier emphasis on security.
Linux, *nix and OS X can already authenticate against AD, with a little effort. OS X does it out of the box.
I'll reckon they'll go with a dirty-tricks FUD campaign. For example:
These threats will attempt to be carried out by Steve Ballmer disguised in a ski mask and a shirt saying "I AM NOT THAT DANCE MONKEY BOY OFF YOUTUBE."
I must agree. I didn't say I liked COBOL: I'm simply saying it's very simple, because it has a very English-like syntax. However, I find its typing style laborious and slow.
That is not to say that it's not an incredibly simple language to pick up, and practically anyone could learn to code a basic functional program in it within the hour. I do think that there are going to be people who like COBOL who would be quite happy to code in it, for that amount of money, despite the fact that for many developers it is a cumbersome dinosaur of a language.
There are many people I know who were programmers in yesteryear, and love COBOL because of its English-like syntax. So I do think that COBOL could make a comeback (albeit indirectly) because of its simplicity.
COBOL could easily make a comeback, because it's so damn easy to learn. I could train a cuttlefish to write a complex accounts program in COBOL. Heck, I bet even Mike Huckabee could probably manage to work out "Hello World".
Perhaps the reason it may come back is because of the complexity of modern languages. Most OOP variants are horrifically difficult to understand, and a return to the statement-based languages of yesteryear might actually be a good thing: let the coder get on with writing his code, and let the interface builder do all the fancy OO stuff.
True, vim has a rather good tutorial, even if I'd prefer it to be along the lines of the 'old-style' tutorials.
I believe some distributions of Windows XP had a 'step-by-step' learning system which had manuals, instructional videos, and interactive tutorials. You were actually given a mock-up of the window, and guided through it keystroke-by-keystroke.
Vi, once you've got used to it, has a wonderfully powerful interface and becomes quite intuitive when you learn new commands. The problem is that someone who's never touched a computer before is unlikely to think "right... perhaps if I try pressing 'colon'... ah!" This also means they'd not be able to get into the tutorial. It is intuitive to people who've used it before, but it's not intuitive if you've never seen it before.
On the other hand, if they see a mouse and that moving the mouse moves that little arrow on the screen, they might figure that clicking the mouse will 'press' what the arrow is on. (This is even more intuitive on one-button mice, but even two/three/four-button mice are speedy enough to pick up.)
Oh, vi is easy. Colon, Q, enter. You're out. If it's not been saved, use :wq! to save or :q! to discard.
That said, they're both very counter-intuitive. The user doesn't want to have to patch their software. They want it to just work, and in an intuitive and sensible way. I learnt to compute on a Dell workstation with Windows 3.11, MS-DOS 5 (or 6) and Works 2, and I am entirely self-taught. Those interfaces were more intuitive than some FOSS interfaces around today, which is why I was able to sniff my way around and orientate myself quite quickly.
Something that might solve a lot of problems is better documentation, with proper tutorials. The tutorials in Works and Windows 3 were fantastic in that they explained things very simply, and were interactive: i.e. you learned by doing.
I assume he was forced to respond to them, because otherwise he wouldn't have been paid.
He's a troll. Send him to the killfile, ban list, or whatever your BBS software's equivalent is.
I'm using ume-launcher (the Netbook Remix launcher) on my Eee PC 701 right now, and it really isn't bad at all. It's still quite buggy though:
Apart from that, it's very efficient, and either way it pwns Asus's default Eee launcher: it's prettier, less resource-intensive and more space-efficient.
I agree about PMs versus Windows, but not versus OS X. Most of the time that simply involves dragging something to the applications folder and starting it - dead easy. A cucumber could do it.
OS X also has a package manager - of sorts. It's similar to dpkg in the way it's constructed, in that it's only a package format that also provides logging in the form of package receipts, and installation and configuration by wizards and post-flight scripts. In both cases, though, it's nice and easy to uninstall an app - all one has to do is drag it to the Trash.
I suppose it's easier for OS X to do this for two reasons:
But the key introduces a critical weakness to the system: it's small, so it's easy to lose. I still think that sitting the OS on internal storage would make things better, and provide some kind of insurance policy against losing the key.
On a personal note, IMHO Mandriva is maybe the best linux distro, by far, give-it a try.
Where do you work at Mandravia?... [/sarcasm]
I recall reading a comment here, not long ago, stating that dealing with mobile providers in the US was like 'choosing between shit sandwiches'. This, unfortunately, seems to reiterate that this is the case.
Over here in the UK, things are a lot better. If you don't like customer service, you change network. Simple as. I switched from Tesco Mobile (poor customer service on O2's otherwise excellent network) to 3 earlier this year, and the process was quite painless. Am I right in saying that having mobiles on contract is more common in the USA which is what makes dealing with providers such a nightmare?
In this case, it means that losing the key, the smallest component, would mean the computer becomes useless. In that case, it would be better to have the base OS on the computer's ROM, especially because it's easier to lose a USB key than a notebook.
Also, I disagree with your statement about Emtec being a pioneer in this respect. The ultra-sub-notebook/netbook market is generally split into a high-end and low-end - the low end is dominated by the Asus Eee PC, and the high end has been practically cornered by the MacBook Air. Both of these machines have been selling like hotcakes, and the others were all somewhat late into the market. Emtec is, therefore, going to be very late onto the market at this rate. It's like the iPod - it was an early example of an HDD music player which was one of the first which was a truly usable concept. Most HDD players before that were, TBH, slow and rubbish. It got in early, and it's still going strong to this day: the same, I think, will go for the Eee and the Air.
As open-source software and OSes go, it may be pioneering, but I think as a computer and production OS the concept is utterly flawed, and they've entered the market very, very late.
As I understand this, the root partition is stored on the machine's internal ROM. The user's home folder sits on the USB key, along with something that somehow links to /etc/passwd to provide authentication. The key is automatically mounted when inserted. Correct me if I'm wrong.
This doesn't sound particularly promising - it would be very easy to lose the key. I also fail to see why, when most ultra-sub-notebooks are bought by a person for their use, and their use only.
Also, will the home folder on the key be accessible when plugging into another computer, say, a desktop running OS X, Windows or another Linux distro? If so, it would kind-of defeat the object. Emtec would be entering the market very late, so they can't expect this to take the market by storm. If it doesn't, it kind-of defeats the object of sticking everything on a USB key.
It's possible to boot some *nixes from a RAM image - like when resuming from hibernation/Suspend to Disk. I suggest booting the actual OS from the image and putting the home folder on a different partition in the hard drive.
The advantage of this is that it bypasses init (or upstart, sysvinit, launchd or whichever startup daemon you're using) and boots things into a state which works immediately.
It's not so much handling, although Fedora's handling of wpa-psk is atrocious, believe me.
The problem is driver compatibility - at the moment, it is pretty much hit and miss whether or not your wireless card will work. Ironically, often the only way to obtain the required driver will be from a multiverse repository - by connecting to the Internet.