Took me 6 seconds to open system preferences, get to the display tab and change the colours (refresh rate is right below it but I have an iMac so it’s disabled). If you give an average user 10-15 seconds in order to find the display section and find the refresh pop up menu. I’d like to see you locate, open, edit and save a text file in 6 seconds. I’d like it even more to see an average user to it in under 15
6 seconds? Maybe on a DP PowerMac G5.
It takes me 10.32 seconds just to get to the refresh tab on my iBook G4 (can’t actually change it)--add another 5 for an average user. OTOH, it takes me 14.29 seconds to change both horizontal sync and vertical sync in xorg.conf the long way and save it. That’s a far cry from the GUI taking ‘a fration of the time’.
Theoretically I could have made it faster by authenticating beforehand (as it is on OS X). That would have saved a second or two. I could have also saved another 2-3 seconds simply by defining the line and column I wanted (I happen to know it in this instance), but that would have been unfair.
And this is also a fairly rare example. For any sort of complex configuration changes, a text file is *substantially* faster.
Even I’m used to doing that. I don’t see the point of having a GUI if you have to go edit text files. Sure, text files were okay in DOS, but does is much more simple than something like Linux.
We still use text config files because they’re easier to edit, easier to work with (backing up, switching between multiple configurations, etc), and easier to transfer between machines. They’re even faster, when you’re trying to do something even moderately complicated. Not to mention all the headless server boxes without a GUI. Using plain text config files also makes troubleshooting much easier.
If you have 52 settings, would you rather see a panel with 52 checkboxes or a text file with 52 possible entries, some of which are not there and thus set to the default?
I’d much rather have the text file, especially if it is well commented.
It takes me 10.32 seconds just to get to the refresh tab on my iBook G4 (can’t actually change it)--add another 5 for an average user. OTOH, it takes me 14.29 seconds to change both horizontal sync and vertical sync in xorg.conf the long way and save it. That’s a far cry from the GUI taking ‘a fration of the time’.
Actually it was on my iMac Core Duo. Just did it again on this Powerbook in 6 seconds. Pretty much any mac in the past year or 2 I could do it in under 10 second. Of course that’s your 14 seconds for you to change it compared to my 6 seconds. When you add the time for an average user (I said 10-15 so that’s about 5-10 seconds added) then you get around 15 vs 24 seconds. That’s 2/3 of the time. But of course an average user is going to spend more time working in the text file than simply clicking a mouse 3 times because it isn’t something they’re used to. BTW, does your 14 seconds include locating the file and opening it, starting from a blank desktop? Cause my 6 seconds includes opening system preferences from a blank screen.
I’d much rather have the text file, especially if it is well commented.
But the problem is a lot of Linux users think that and think that it’s superior to what the other OSs have. People who use the other OSs just wonder why Linux users are stuck in the past and prefer the hassle of messing with text files. The whole problem that Linux has is that it has had little in the way of good usability design, this being one prime example.
So to take this back to the original topic. Can you imagine sitting a novice down in front of a black screen with a single blinking cursor and tell them to configure the display? They might be there for sometime before they figure out what to type to get help. Then after learning what the commands are to navigate they would have to read and learn through experimentation how to accomplish the task. A GUI of any kind is a massive leap forward since it gives a user visual clues on where to go and what to do.
So to take this back to the original topic. Can you imagine sitting a novice down in front of a black screen with a single blinking cursor and tell them to configure the display? They might be there for sometime before they figure out what to type to get help. Then after learning what the commands are to navigate they would have to read and learn through experimentation how to accomplish the task. A GUI of any kind is a massive leap forward since it gives a user visual clues on where to go and what to do.
BTW, does your 14 seconds include locating the file and opening it, starting from a blank desktop? Cause my 6 seconds includes opening system preferences from a blank screen.
1: Launch Terminal
2: “sudo nano /etc/X11/xorg.conf”
3: <entered password>
4: Ctrl+W; “HorizSync”; Enter
5: Make corrections (just retyped the values I already had)
6: Move down to next line
7: Make corrections (just retyped the values I already had)
8: Ctrl+O; Enter
This isn’t actually changing the refresh rate on the monitor though--it’s how to fix the problem that prevents the GUI frontend from doing it. Not sure why you insist on ‘locating’ the file. It’s not like the file moves around.
But the problem is a lot of Linux users think that and think that it’s superior to what the other OSs have. People who use the other OSs just wonder why Linux users are stuck in the past and prefer the hassle of messing with text files.
It’s not a hassle to mess with config files--that’s one of the reasons why we keep using them, even when GUI frontends are available.
The whole problem that Linux has is that it has had little in the way of good usability design, this being one prime example.
That’s not a prime example of poor usability--there are many valid ‘usability reasons’ to keep using config files. I make a concious decision *not* to use the readily available frontends, because it’s easier to use the CLI tools and config files. I do the same thing on OS X, when I can.
So to take this back to the original topic. Can you imagine sitting a novice down in front of a black screen with a single blinking cursor and tell them to configure the display? They might be there for sometime before they figure out what to type to get help. Then after learning what the commands are to navigate they would have to read and learn through experimentation how to accomplish the task. A GUI of any kind is a massive leap forward since it gives a user visual clues on where to go and what to do.
It doesn’t provide much in the way of clues, and it’s massively inefficient for experienced users. How many times am I going to be changing the refresh rate on my monitor? That’s not going to happen very often--I should just need to set it once. Hell, most people are going to have it set properly OOTB.
Your example is also inaccurate, since any of the distributions they would be using default to a desktop. It may be stuck at 1024x768@60hz (with some configurations), but they’ll still have a GUI.
1: Launch Terminal
2: “sudo nano /etc/X11/xorg.conf”
3: <entered password>
4: Ctrl+W; “HorizSync”; Enter
5: Make corrections (just retyped the values I already had)
6: Move down to next line
7: Make corrections (just retyped the values I already had)
8: Ctrl+O; Enter
Now go to a random computer user and ask them whether they’d prefer to do that or click a mouse button three times.
Re: LetterK
I wonder how many “experienced users” are there? I guess I’m not by your standards. Yet I have been using computers for more than 20 years. I started using them well before the mass adoption of the GUI. Back then the CLI was seen as a “real” interface. The infant GUI on Apple was truly groundbreaking. Since that time both MS and Apple have spent millions on making improvements to the GUI. Why? Well because it makes their products easier to use and more attractive and accessible to a broader range of people.
1: Launch Terminal
2: “sudo nano /etc/X11/xorg.conf”
3: <entered password>
4: Ctrl+W; “HorizSync”; Enter
5: Make corrections (just retyped the values I already had)
6: Move down to next line
7: Make corrections (just retyped the values I already had)
8: Ctrl+O; Enter
Now go to a random computer user and ask them whether they’d prefer to do that or click a mouse button three times.
Which is an unfair comparison. I picked the method that I found fastest--not the only method, or even the easiest method. If I was explaining to a novice, I would have suggested that they use gedit or kate, for example. Avoid the CLI altogether.
I wonder how many “experienced users” are there? I guess I’m not by your standards. Yet I have been using computers for more than 20 years.
Opinions on configuration methods do not indicate experience.
I started using them well before the mass adoption of the GUI. Back then the CLI was seen as a “real” interface. The infant GUI on Apple was truly groundbreaking.
Had I been in a position to be using a computer at that time, I probably would have been more fond of X11 than Mac OS. It was, after all, more advanced--even if it did have steeper hardware requirements.
Since that time both MS and Apple have spent millions on making improvements to the GUI. Why? Well because it makes their products easier to use and more attractive and accessible to a broader range of people.
I disagree. Their products are easier to learn to use, but more difficult to actually use. GUIs are very difficult to use for many complex tasks. They don’t handle data piping or regular expressions very well, for example. I mean, how do you mimic something as simple as “ls -R ~ | grep “foo” > ~/foo-results-`date +%A`” in a GUI?
If you have 52 settings, would you rather see a panel with 52 checkboxes or a text file with 52 possible entries, some of which are not there and thus set to the default?
I’d much rather have the text file, especially if it is well commented.
But you don’t know what the 38 setting that are not set happen to be labeled as as there is no documentation.
I think that Linux people need to get with the times and at least develop frontends for all the text files that need editing. It takes me about 5-30 seconds to figure out what to do with a dialog I haven’t seen before, it takes me anywhere from 5 to 30 MINUTES to figure out what to do with a text file, expecially if there are settings that exist but are not set and thus not visible and must be learned from a manual.
At this point in time, the Mac OS does appear to be more hassle-free, and the hardware that comes with it is very solid.
I would tend to disagree with the “very solid” comment. My Mini works well but it definteily needs more ram and vram and neither are particularly easy to upgrade from a non-techie POV. the MBP and MB are both inline with something I’d except from a cheap windows notebook not one of above average price. The discoloration, heat, and tons of other issue don’t qualify as “very solid” IMO. You could include the lack of good Windows drivers, two button mouse, card readers, DL Burning etc. in that quaulification as well.
forgoing the whole process of installing antivirus and anti-spyware, and constantly checking for patches and updates, can be avoided. People say the virus/spyware argument is unfair because a properly maintained Windows installation won’t get them, but there is something involved in keeping a Windows installation “properly maintained.” You need to constantly check for Mac patches and updates as well. In the last month or two, I’ve heard of just as many exploits and holes in iTunes, QT, and the Mac OS as I have in Windows, WMP and such. Not only is there some *very* public exploit code floating around for Macs but people are getting infected too.
Part of the reason for me getting this new Mac was to further my programming skills. I’ve successfully “hacked” this Mini several times. It’s beyond trival to make software that doesn’t need admin previlges to run or install… or to create sociall enginnering attacks (such as phishing) and have them work flawlessly. And the “firewall off by default” thing is ridiculous considering that my Mini shipped with both Bluetooth and Wifi on and broadcasting out of the box (actually the BT might not have been broadcasting just discoverable but I didn’t test it). The damn thing jumped right on my neighbors wireless network, without notifying me at all, and before I could even download the 300+ mb of updates it said it needed.
FYI… “properly maintained” XP install IMO means SP2+Windows Defender. And given the fact that XP will automatically install those patches it is actually easier to keep it updated than it is my Mac.
If you have 52 settings, would you rather see a panel with 52 checkboxes or a text file with 52 possible entries, some of which are not there and thus set to the default?
I’d much rather have the text file, especially if it is well commented.
But you don’t know what the 38 setting that are not set happen to be labeled as as there is no documentation.
There’s documentation for it somewhere. If not in the config file itself, then in the documentation, or online somewhere. Of course, your complaint is mostly about poor documentation in general, which is just as much a problem in GUI frontends as it is in config files. I’ve seen some *very* poorly worded options in GUI frontends before. I mean, if you were a novice and you saw a blank string for “Dist_hardw”, would you know what the hell you were supposed to put there?
GUI frontends do not magically solve the issue of poor documentation.
Which is an unfair comparison. I picked the method that I found fastest--not the only method, or even the easiest method. If I was explaining to a novice, I would have suggested that they use gedit or kate, for example. Avoid the CLI altogether.
Using a graphical program over a text program doesn’t make this process any easier or faster. In order to know how to edit a text configuration file, you have to know what essentially amounts to a scripting language: the specific commands and syntax that need to be entered into a file to be understandable to the software. Even if a file is well-commented, which you say makes everything easy, a user would have to sit there and read the comments just to figure out what to type in. I assume you already know what needs to be entered in many of these files, but the typical user will not know these things, and they won’t remember just from doing it once or twice. Their take is “why do I need to read instructions just to change my display settings?” which is completely reasonable. This is why the vast majority of people aren’t interested in CLIs or editing text configuration files: in this day and age, requiring users to spend a significant amount of time just learning how to handle files and configure things is considered unacceptable. You said you can’t understand why people think doing these things is difficult or unrefined, and we’ve told you why in great detail.
Opinions on configuration methods do not indicate experience.
Indeed. We are both experienced users. You think editing text configuration files is the best way to do things, and I think it’s a headache.
I disagree. Their products are easier to learn to use, but more difficult to actually use. GUIs are very difficult to use for many complex tasks.
But are those tasks that would be regularly used by a large number of people? Sure, I can change the extension on 1000 files instantly with a command line, but whether or not that’s useful is a different question. Granted, there are some things that command lines can do faster that are very useful, but only a few people will be interested in doing those things. I think most of us can agree that having a command shell available for those few people is a good idea, but you seem to be saying that it’s better for everyone, which I can’t find evidence for.
Keep this in mind, too: learning something once doesn’t guarantee you’ll remember it. A command line puts the responsibility on the user to remember a lot more than a GUI does. A CLI is faster if you already know how to do everything and can remember it all, and you’re a reasonably fast typist, and most people don’t fall into those categories.
They don’t handle data piping or regular expressions very well, for example. I mean, how do you mimic something as simple as “ls -R ~ | grep “foo” > ~/foo-results-`date +%A`” in a GUI?
Calling that command simple isn’t going to make your case look any stronger. What it does may be simple, but how simple is it to remember what everything in that command means, in which order to enter everything, etc.? I know you will say it’s a no-brainer, but most of us don’t work that way. Speaking of which, can you explain what that command does?
And my point about editing text files is that it often requires entries not too different than that (from my point of view). It’s not just a matter of selecting things; you actually have to know how to put the entries in.
There’s documentation for it somewhere. If not in the config file itself, then in the documentation, or online somewhere. Of course, your complaint is mostly about poor documentation in general, which is just as much a problem in GUI frontends as it is in config files.
The difference is that for a typical user, editing a text file requires some serious reading to find out what’s going on, but there’s at least a possibility of figuring out a GUI.
I’ve seen some *very* poorly worded options in GUI frontends before. I mean, if you were a novice and you saw a blank string for “Dist_hardw”, would you know what the hell you were supposed to put there?
I take it that came from a Linux GUI frontend?
GUI frontends do not magically solve the issue of poor documentation.
