That's not quite where I was going, but actually yes, it can cause nasty problems. If there were 10 browsers (heh there probably is, good thing Firebird seems to be the de-facto default) for Linux all specialized at doing their own thing well, would that be good or bad?
Good. Guess what, not every app is a swiss army knife. Most shouldn't be. Users are different. This is why you have 2000 computer types and, wait for it, 2000 different type of people buying them. People's wants and needs differ. If there are 20 or 30 browsers, maybe one, MAYBE, will be what the person wants to use.
But even with that said, you'll note that of the handful of browsers out there most use Gecko or KHTML. So they share backend code, which is good, but can do different things for the user, which is... Oh, ya, good.
I agree, in a way. The problem isn't where it's stored, but that Windows/MacOS/KDE/GNOME/et al make associating where a file goes difficult. Mainly because they each use at least three different ways to work with the filesystem at any time.
Additionally, some apps will default to different directories when saving, which causes confusion because, honestly, save boxes are just another "click OK and close" popup these days.
Keyboard shortcuts differ in non-obvious ways would be the first and more painful one. Input methods take different text (ctrl-backspace on Windows is a character, in GTK it's a action), Alt-click a window and it'll move whereas in Window it's just a alt-click. The problem is it acts like windows, people will sit down and have muscle memory FROM windows, and when it doesn't work become upset.
I've only had a Windows install for all of a week (just got a new laptop) so I'm sorry I can't do a long, involved list, but it's the small differences that will confuse a user.
It's not that it looks like Windows, it's that it tries to feel like Windows, but isn't. Your sitting at a system that attepmts to give you a "Free Windows," but it's missing things. Like a cloned fragrence, sure it's the same in most ways, but it's the small things that stick out and sour the whole experience. You have apps that are coded like Windows apps, look like Windows apps, but have small changes that don't become obvious until you're sitting there using them awhile.
My problem isn't that they look similar. Hell, most of my problem isn't that they feel the same, really. It's that they look the same, attempt to feel similar, and add nothing that would instantly make a person say "ah ha! This is something new, better, and different. I may have to learn something new." Instead they're things that are interchangable in a lab context, with little to make someone say "oh, I'm using something different" when they sit down, but have small changes in how they interact when you use them. It's confusing when Windows apps don't act like Windows apps, it's doubly so when it's another OS. Granted, it's probably something that someone can adapt to, but it's still a pet peeve.
And, yes. Please remove the close button on the window frames. They are small, hard to hit, and too easy to hit instead of mini/maximize. But I put a close button right next to the options button as the first two buttons on the toolbar of all my apps, so meh. Menus on the top of windows should go away because they're horribly hard to use. "Hey, I've a great idea. Let's take all our functions, and toss them in a small sliver of the window's space at the top. It'll be great." It was a horrible idea in Windows 3.1, it's a horrible idea now. IMHO, the Mac way is good, the ROX way (menu target is the entire window, right click or hit the menu key to pop it up) is better. Large click-target, always there, keeps people from making really confusing popup menus or having to hunt them down with a mouse. Of course, I also disagree with "status bars" and a number of other common UI elements in todays software.
Beginners and advanced users are the exception, not the norm. Make software that's easy to use for a intermedite, make it easy enough for the newbie to not throw up their hands in defeat, and open enought the expert has things to learn and use.
The car example is more apt then you let on. A car is very complex. Internally. The face given to the user is simple, but if you open it up and tinker there's a lot of complexity there. Most won't, but if your inclined to do so it's possible. If your skilled at cars, you can change the engine timings and add/remove parts for better performance, but it's still easy enough to use for a end user that they can sit down with it and it "just runs." If they locked things down so no one could change the parts of the car, the experts wouldn't be happy because they could no longer have anything to investigate or tinker with, and if they only stuck with the complexity (hey, just try and teach someone to drive with a manual clutch and see how enthused they are about driving;) ) then the new drivers would be instantly put off, likely never to even attempt to drive.
Better UIs aren't about limiting choices, but about making choices easier to make and use. UIs aren't about making complex things simple, but making complex things accessable. There's a big difference.
Yes. But no one used the things they clone.;) As such, it gives people leeway in what they do, because they don't have to stay "true" to some ancient and broken UI.
ROX is inspired by RISC OS, and while some people seem to try cloning RISC apps for it, I myself have never seen a RISC box in action but just code with a similar UI mindset. Well, at least the parts that make sense. (The haphazard toolbar layout they RISC OS uses makes my skin crawl) I've seen images of apps similar to mine on RISC, so maybe I get close, but hopefully I'm just taking the good and ignoring the bad.
XFCE is slowly becoming something completely different from it's CDE-clone days. Which doesn't matter as, again, no one used CDE.
Though, honestly, I've myself wanted to start my own DE which owed lineage to nothing, but I don't have the time or coders to do so because, quite frankly, every DE out there is a clone of a broke UI that failed to another broken UI or a clone of the broken Windows UI... Ah well. Point was simply that Windows is the defacto ui now, and cloning it will always lead to confusion, as it will never clone it closely enough to not let the user down at some point.
I couldn't agree with you more. So, why not support ROX-Desktop, GNUStep and XFCE?
We need to have something completely different as the base desktop to allow coders and interface designers the freedom to make things totally unique and different. You hand people more or less Windows interface and, well, like you said, you get a broken Windows clone.
Choice is out there. We just need people with coding and UI experience to push them farther, and people to get the word out.
Re:Nice, but they've got it all wrong...
on
Linux Desktop Guide
·
· Score: 1
Users do not stay beginners. Most users are intermediate.
This may not sound important, but it is. Most people looking to use a GNU/Linux system are people who are experts and want to gain new skills, or intermediates who want some feature GNU/Linux systems provide. Beginners aren't the audience because, as you said, they're not going to be exposed to it.
Nore would it matter if they were. Beginners wouldn't want a guide like this. They want to turn it on and use it immediately. This, IMHO, is why there's such a backlash against the Windows UI. It's a horrible interface to begin with, but it's also the one a beginner is expected to learn. It's insane. Hell, it's over a decade now and Windows still doesn't make a bloody text box or window scroll consistantly across the toolkits. (Some require clicking in the box, some require just focus on the window, some will scroll thing X instead of thing Y, it's a mess)
But I'm getting off point. Which is to say that a beginner is not the audience for this. A GNU/Linux beginner, perhaps, but a beginning computer user? That'd be silly.
I'm no big RedHat fan, but there's no reason to expend a large sum for RedHat with proprietary software. Flash? Java? MP3s? All exist in either 3rd-party repos, or are free downloads to install.
So basically, this argument is "SuSE is more pragmatic because it doesn't even give the user a choice of if they wish to use all-free software or not?"
While I'll give you the parent's comparison isn't fair, I have to say compiling has not been fixed. Last time I tried Fedora, I gave up when it wouldn't compile FluxBox and some other random things. (I remeber flux because it meant I had to use MetaCity, and that is something I can not forgive.;) )
It does still have weird "quirks" with compiling. I'm sure there are ways around it, but damned if I want to spend my time learning how to special-compile with RedHat when Slackware, Debian, Gentoo, FreeBSD, NetBSD, OpenBSD, SuSE, and innumerable other Free *N*X distros compile things I want without giving me trouble. (though, in all fairness, if it had plain-jane theme packages for GTK instead of only BlueCurve, I might actually consider it for a desktop as pretty much everything I run these days'd compile without a hitch, AFAIK)
It's not like running "rm/etc/passwd" It's like running "cat/etc/passwd",/etc being unreadable, and it instead of failing running dd if=/dev/random of=${ETC_PARTITION}
It takes one function to check if/usr is writeable. Two, maybe three, to see if everything the RPM needs to install has access. This is what a installation utility should and must do. If it can't write, a message should pop up. If it's a problem, add a option to the apt configuration, or a command line to "force" installation, in which case the problem would be on the user's hands. Simply put, there's no excuse for a app to destroy it's database if something fails - otherwise, why even bother with it? Just unpack tbz2s under root. It'd provide the same level of utility.
Eating it's database on a simple error that should not be destructive is not something a user's desktop (Fedora) or a "Enterprise-level OS" (RedHat) should do.
Nothing four or five levels in is readable on a printout. Nothing. At least with space indentation, it's readable on the computer screen, and if need be could be tossed to a preprocessor to turn the leading tabs to spaces before printing to make it fit on the page easier. Not to mention that, quite frankly, if the coder in another language is even half conserned about readablity of the code, they'd be doing the same in a non-whitespace-sensitive language.
Back when I was taking marketing classes (teh horror), one teacher was talking about a bad experience he had. As he left this store, he found a manager and told them why he was leaving, but only because he usually liked the store. His rationalization being that if you like a store and want it to suceed, you tell them what they're doing wrong to fix it. If you don't, you just leave and leave them to their own devices.
Quite frankly, these days I say leave bestbuy to their own. Maybe they'll be able to use one of those PSP when they, themselves, are "broke."
Re:It is more a case of terminals
on
Semper WiFi
·
· Score: 1
You can't use wifi for battlefield communications. Hell, it's barely secure enough to read email and surf the web reliably assuming no one is actively monitoring it with the intent to intercept/decode it. Even MAC based auth. isn't good enough, as spoofing the MAC is "easy" if you have broke the WEP key.
Until something better then WEP is made available, don't expect this to be anything but a moral booster. ANd that doesn't even account for the problems of signal range.
What if I went to Microsoft and said I wanted to license their technology so I could make MS Office compatible with Linux, and Microsoft said NO and I still went ahead and did it by doing exactly what Real did? I probably would have gotten the shit sued out of me.
You mean like OpenOffice.org does? Granted, they didn't ask permission, but it doesn't change anything in this example.
this should make it clearer. The collection can be anything, say a array. It's been awhile since I touched Ruby in any real manner, but I believe for_each would also work as well. Basically, you're looping through the entire array/list/what have you. I believe you can even for it with a string, going through each character.
Anyway, in the front of the block you have a |iteration|. This is defining what you want each iteration to define the variables of. Say you had a matrix of:
[ [1, 2, 3],
[4, 5, 6]]
And you want to go through it all. You'd have a: matrix.each() { |x, y, z| do_stuff(x, y, z) } or matrix.each() { |iter_array| do_stuff(iter_array) }
Agreed, the PyQT binds are great. But they need documentation. I was playing with them using the C++ API docs (about all that exist that I could find, unless windriver has them hidden somewhere...) and figuring out what does and does not map from the class declarations alone was a big pain.
Personally, I always though he was a zombie or vampire. So form of undead, at least.
I mean, have you ever see video of him? I've been around computers for over 15 years, and I've never seen someone that pale. Dark rings around his eyes in every photo. A aversion to all things holy. All the signs are there.
AOL is dialup computer network that provides internet psuedo-access through a proxy. It's not a true ISP. The majority of pluses a software firewall offers would be redundant due to the fact that your behind AOL's proxy.
Full OpenGL support, keybind support so you don't have to change your finger layout from all the new FPS;) , improved deathmatch, support for overlapping-area WAD files, tons of nifty stuff. Just enough changes to stay fresh while keeping the old skewl game in place.
Slashdot Mirror
Psuedocode:
main()
load_needed()
show()
start_icon_load_thread()
event_loop()
How does that affect your start time? Not at all. How does it affect the app's percieved speed? Dramatically.
Good. Guess what, not every app is a swiss army knife. Most shouldn't be. Users are different. This is why you have 2000 computer types and, wait for it, 2000 different type of people buying them. People's wants and needs differ. If there are 20 or 30 browsers, maybe one, MAYBE, will be what the person wants to use.
But even with that said, you'll note that of the handful of browsers out there most use Gecko or KHTML. So they share backend code, which is good, but can do different things for the user, which is... Oh, ya, good.
I think we should have a feature-complete opensource IE in 2011. That's when the last blaster patent expires.
I agree, in a way. The problem isn't where it's stored, but that Windows/MacOS/KDE/GNOME/et al make associating where a file goes difficult. Mainly because they each use at least three different ways to work with the filesystem at any time.
Additionally, some apps will default to different directories when saving, which causes confusion because, honestly, save boxes are just another "click OK and close" popup these days.
I think you're missing a ELF header.
No, but the code to decode it (ie. the .doc format's codec) does.
Keyboard shortcuts differ in non-obvious ways would be the first and more painful one. Input methods take different text (ctrl-backspace on Windows is a character, in GTK it's a action), Alt-click a window and it'll move whereas in Window it's just a alt-click. The problem is it acts like windows, people will sit down and have muscle memory FROM windows, and when it doesn't work become upset.
I've only had a Windows install for all of a week (just got a new laptop) so I'm sorry I can't do a long, involved list, but it's the small differences that will confuse a user.
It's not that it looks like Windows, it's that it tries to feel like Windows, but isn't. Your sitting at a system that attepmts to give you a "Free Windows," but it's missing things. Like a cloned fragrence, sure it's the same in most ways, but it's the small things that stick out and sour the whole experience. You have apps that are coded like Windows apps, look like Windows apps, but have small changes that don't become obvious until you're sitting there using them awhile.
My problem isn't that they look similar. Hell, most of my problem isn't that they feel the same, really. It's that they look the same, attempt to feel similar, and add nothing that would instantly make a person say "ah ha! This is something new, better, and different. I may have to learn something new." Instead they're things that are interchangable in a lab context, with little to make someone say "oh, I'm using something different" when they sit down, but have small changes in how they interact when you use them. It's confusing when Windows apps don't act like Windows apps, it's doubly so when it's another OS. Granted, it's probably something that someone can adapt to, but it's still a pet peeve.
And, yes. Please remove the close button on the window frames. They are small, hard to hit, and too easy to hit instead of mini/maximize. But I put a close button right next to the options button as the first two buttons on the toolbar of all my apps, so meh. Menus on the top of windows should go away because they're horribly hard to use. "Hey, I've a great idea. Let's take all our functions, and toss them in a small sliver of the window's space at the top. It'll be great." It was a horrible idea in Windows 3.1, it's a horrible idea now. IMHO, the Mac way is good, the ROX way (menu target is the entire window, right click or hit the menu key to pop it up) is better. Large click-target, always there, keeps people from making really confusing popup menus or having to hunt them down with a mouse. Of course, I also disagree with "status bars" and a number of other common UI elements in todays software.
An easy way to put this is:
;) ) then the new drivers would be instantly put off, likely never to even attempt to drive.
Beginners and advanced users are the exception, not the norm. Make software that's easy to use for a intermedite, make it easy enough for the newbie to not throw up their hands in defeat, and open enought the expert has things to learn and use.
The car example is more apt then you let on. A car is very complex. Internally. The face given to the user is simple, but if you open it up and tinker there's a lot of complexity there. Most won't, but if your inclined to do so it's possible. If your skilled at cars, you can change the engine timings and add/remove parts for better performance, but it's still easy enough to use for a end user that they can sit down with it and it "just runs." If they locked things down so no one could change the parts of the car, the experts wouldn't be happy because they could no longer have anything to investigate or tinker with, and if they only stuck with the complexity (hey, just try and teach someone to drive with a manual clutch and see how enthused they are about driving
Better UIs aren't about limiting choices, but about making choices easier to make and use. UIs aren't about making complex things simple, but making complex things accessable. There's a big difference.
Yes. But no one used the things they clone. ;) As such, it gives people leeway in what they do, because they don't have to stay "true" to some ancient and broken UI.
ROX is inspired by RISC OS, and while some people seem to try cloning RISC apps for it, I myself have never seen a RISC box in action but just code with a similar UI mindset. Well, at least the parts that make sense. (The haphazard toolbar layout they RISC OS uses makes my skin crawl) I've seen images of apps similar to mine on RISC, so maybe I get close, but hopefully I'm just taking the good and ignoring the bad.
XFCE is slowly becoming something completely different from it's CDE-clone days. Which doesn't matter as, again, no one used CDE.
Though, honestly, I've myself wanted to start my own DE which owed lineage to nothing, but I don't have the time or coders to do so because, quite frankly, every DE out there is a clone of a broke UI that failed to another broken UI or a clone of the broken Windows UI... Ah well. Point was simply that Windows is the defacto ui now, and cloning it will always lead to confusion, as it will never clone it closely enough to not let the user down at some point.
I couldn't agree with you more. So, why not support ROX-Desktop, GNUStep and XFCE?
We need to have something completely different as the base desktop to allow coders and interface designers the freedom to make things totally unique and different. You hand people more or less Windows interface and, well, like you said, you get a broken Windows clone.
Choice is out there. We just need people with coding and UI experience to push them farther, and people to get the word out.
Users do not stay beginners. Most users are intermediate.
This may not sound important, but it is. Most people looking to use a GNU/Linux system are people who are experts and want to gain new skills, or intermediates who want some feature GNU/Linux systems provide. Beginners aren't the audience because, as you said, they're not going to be exposed to it.
Nore would it matter if they were. Beginners wouldn't want a guide like this. They want to turn it on and use it immediately. This, IMHO, is why there's such a backlash against the Windows UI. It's a horrible interface to begin with, but it's also the one a beginner is expected to learn. It's insane. Hell, it's over a decade now and Windows still doesn't make a bloody text box or window scroll consistantly across the toolkits. (Some require clicking in the box, some require just focus on the window, some will scroll thing X instead of thing Y, it's a mess)
But I'm getting off point. Which is to say that a beginner is not the audience for this. A GNU/Linux beginner, perhaps, but a beginning computer user? That'd be silly.
I'm no big RedHat fan, but there's no reason to expend a large sum for RedHat with proprietary software. Flash? Java? MP3s? All exist in either 3rd-party repos, or are free downloads to install.
So basically, this argument is "SuSE is more pragmatic because it doesn't even give the user a choice of if they wish to use all-free software or not?"
I call FUD.
While I'll give you the parent's comparison isn't fair, I have to say compiling has not been fixed. Last time I tried Fedora, I gave up when it wouldn't compile FluxBox and some other random things. (I remeber flux because it meant I had to use MetaCity, and that is something I can not forgive. ;) )
It does still have weird "quirks" with compiling. I'm sure there are ways around it, but damned if I want to spend my time learning how to special-compile with RedHat when Slackware, Debian, Gentoo, FreeBSD, NetBSD, OpenBSD, SuSE, and innumerable other Free *N*X distros compile things I want without giving me trouble. (though, in all fairness, if it had plain-jane theme packages for GTK instead of only BlueCurve, I might actually consider it for a desktop as pretty much everything I run these days'd compile without a hitch, AFAIK)
It's not like running "rm /etc/passwd" /etc/passwd", /etc being unreadable, and it instead of failing running dd if=/dev/random of=${ETC_PARTITION}
/usr is writeable. Two, maybe three, to see if everything the RPM needs to install has access. This is what a installation utility should and must do. If it can't write, a message should pop up. If it's a problem, add a option to the apt configuration, or a command line to "force" installation, in which case the problem would be on the user's hands. Simply put, there's no excuse for a app to destroy it's database if something fails - otherwise, why even bother with it? Just unpack tbz2s under root. It'd provide the same level of utility.
It's like running "cat
It takes one function to check if
Eating it's database on a simple error that should not be destructive is not something a user's desktop (Fedora) or a "Enterprise-level OS" (RedHat) should do.
Nothing four or five levels in is readable on a printout. Nothing. At least with space indentation, it's readable on the computer screen, and if need be could be tossed to a preprocessor to turn the leading tabs to spaces before printing to make it fit on the page easier. Not to mention that, quite frankly, if the coder in another language is even half conserned about readablity of the code, they'd be doing the same in a non-whitespace-sensitive language.
Maybe, maybe not...
Back when I was taking marketing classes (teh horror), one teacher was talking about a bad experience he had. As he left this store, he found a manager and told them why he was leaving, but only because he usually liked the store. His rationalization being that if you like a store and want it to suceed, you tell them what they're doing wrong to fix it. If you don't, you just leave and leave them to their own devices.
Quite frankly, these days I say leave bestbuy to their own. Maybe they'll be able to use one of those PSP when they, themselves, are "broke."
You can't use wifi for battlefield communications. Hell, it's barely secure enough to read email and surf the web reliably assuming no one is actively monitoring it with the intent to intercept/decode it. Even MAC based auth. isn't good enough, as spoofing the MAC is "easy" if you have broke the WEP key.
Until something better then WEP is made available, don't expect this to be anything but a moral booster. ANd that doesn't even account for the problems of signal range.
You mean like OpenOffice.org does? Granted, they didn't ask permission, but it doesn't change anything in this example.
More disturbing, they have editors? And you thought tech support was a bad job.
Customer: Brainfuck?!?! Why brainfuck?!?!
Coder: Oh... Job security, mostly.
I kid, I kid. Still, you have to respect anyone who codes in something that makes Perl look readable.
just incase anyone's wondering and doesn't want to look it up:
collection.each { |iteration| do_something(iteration) }
this should make it clearer. The collection can be anything, say a array. It's been awhile since I touched Ruby in any real manner, but I believe for_each would also work as well. Basically, you're looping through the entire array/list/what have you. I believe you can even for it with a string, going through each character.
Anyway, in the front of the block you have a |iteration|. This is defining what you want each iteration to define the variables of. Say you had a matrix of:
[ [1, 2, 3],
[4, 5, 6]]
And you want to go through it all. You'd have a:
matrix.each() { |x, y, z| do_stuff(x, y, z) }
or
matrix.each() { |iter_array| do_stuff(iter_array) }
It really nice.
Agreed, the PyQT binds are great. But they need documentation. I was playing with them using the C++ API docs (about all that exist that I could find, unless windriver has them hidden somewhere...) and figuring out what does and does not map from the class declarations alone was a big pain.
Personally, I always though he was a zombie or vampire. So form of undead, at least.
I mean, have you ever see video of him? I've been around computers for over 15 years, and I've never seen someone that pale. Dark rings around his eyes in every photo. A aversion to all things holy. All the signs are there.
AOL is dialup computer network that provides internet psuedo-access through a proxy. It's not a true ISP. The majority of pluses a software firewall offers would be redundant due to the fact that your behind AOL's proxy.
Feh, prboom. Legacy Doom is the only way to play:
;) , improved deathmatch, support for overlapping-area WAD files, tons of nifty stuff. Just enough changes to stay fresh while keeping the old skewl game in place.
legacy.newdoom.com
Full OpenGL support, keybind support so you don't have to change your finger layout from all the new FPS