I was wondering, what about an ncurses-based tool that runs xrandr to tell you what monitors you have, and then lets you make a list of monitors, specifying the position and resolution of each.

For example:

When you press configure, it would let you choose the resolution from a list, then whether you want absolute positioning or Above, below, RightOf, or LeftOf another monitor.

If you chose absolute, it will let you enter x/y coordinates, otherwise it will let you choose the monitor that this one is relative to.

When you're done, it generates an xorg.conf like in one of my previous posts.

Just installed a nvidia card today on my crappy dell. Got tired of not having good 3d with onboard Intel. Too bad it only has pci slot and not pci express, but it is still way better than the onboard Intel.

Installed card, booted, ran smxi, started desktop. Took longer to install the card than it did to install the drivers.

Graphics: Card nVidia G96 [GeForce 9400 GT] X.Org 1.7.4 Res: 1280x1024@50.0hz
GLX Renderer GeForce 9400 GT/PCI/SSE2 GLX Version 3.2.0 NVIDIA 190.53 Direct Rendering Yes

I guess no one thinks it is enough of an upgrade over editing a file. I know I don't think it is.

One thing that is too often overlooked with text-based configurations is the tendency for documentation to propagate spontaneously through the web. If someone on a forum such as this posts a snippet of a config file for some purpose, it is completely unambiguous and concise as to its purpose or meaning. Search engines and other search mechanisms will find it, and many recipients can copy the solution directly into their own configurations. GUIs (and even text-based UIs), on the other hand, serve to obscure the underlying information, and provide little unambiguous way to describe how to effect a useful modification. I know, you can tell someone to click here, pull down this menu, select the third item from the left, and type 'horty-florty' into the text entry box. The audience that will ever find such instructions is going to be much smaller than the audience that will find 'insert these 10 lines into foo-bar.cfg' (followed by the 10 lines of text nicely formatted in [code] tags, of course).

This is orthogonal to the problem of trying to help Windows users with their problems. It takes a huge amount of effort to describe the solution in terms of GUI actions, when a simple patch file and a shell command to use it would take much less time and effort, and probably be more accurate.

A picture is worth a thousand words, but none of those words will make it into the index.

--- rod.

@theNbomr:

It seems to me that if there should be any GUI/TUI config tools, they should be intuitive enough not to require the user to read any instructions, tutorials, or documentation.

By this, do you mean "why doesn't someone else do it?"

I don't know, I think I *might* be able to do it, if there is some kind of "ncurses TUI toolkit" available.

In that case, why not go for it? Up to this point, this thread has been two things: 1) Linux sucks because I can't figure out how to do it and 2) if I were the king of the world it would be different. It would really be great if someone took a problem that affected them and actually made a solution instead of just bitching about it.

But, I question the utility of this with current xorg. Isn't it a fact that with the latest xorg you don't even need an xorg.conf file to get a default X screen running? I'm not sure what exactly is available to the other manufacturers, but with my Nividia card, literally all I did was run "gksu nvidia-settings" to setup two monitors the way I want them.

Never mind all that, though. If there is a problem, and if you are willing to spend time on it to understand it and fix it, that would be great!

It shouldn't be too hard to make a program that gathers data from the user and creates a formatted text file, but the problem is that I don't know how to set the resolutions and refresh rates, set absolute positions, make sure xrandr is running, and setting drivers.

I suppose that is a nice holy grail sort of objective, but if were possible for the tool to automatically know exactly what is the right thing to do at all times, then why would it need any interface at all? Are computers supposed to be different from other tools, where there is some expectation that the user will have an understanding of how to use the tool? If I buy a table saw, throw some lumber at it, and the lumber doesn't turn into furniture, I don't condemn the tool.

In the final analysis, I'd say it comes down to where you draw the line between the knowledge built into the tool, and the knowledge built into the user of the tool. Since it clearly takes a human a great deal of time and effort to learn the solution to many of the problems in the scope of this discussion, it is probably similarly difficult to build that knowledge into the tools. That might explain why it hasn't been done.

I don't think there is as much attention paid to installation-related tools as there is to more end-user tools for a couple of different reasons. The daily-use type of tools are used, well, daily and as such deserve more attention. Installer tools are generally used once or a small number of times, and then forgotten. Installation of systems is frequently done by people with greater levels of technical knowledge and aptitude. Even the Windows world recognizes this. For end-user applications, it is much easier for the product to stipulate what prerequisites must exist prior to installation, and for the installation to leverage what is known to be there, compared to OS installations, which by nature are expected to accommodate a vast and diverse variety of hardware, the possible combinations of which is a staggeringly large number. In order for a configuration tool to work in a high percentage of cases, it must accommodate this large number of possible hardware combinations. Moreover, the tool must evolve over time as new technologies enter the scene. In the Windows world, much of the work of supporting hardware configuration has been undertaken by the hardware vendors. Microsoft has the clout to force the hardware vendors to support this, but Linux does not. Many hardware vendors are beyond unhelpful.

--- rod.

I don't want to make a fully automatic config tool, because it would have to read the user's mind to know what layout, resolution, and refresh rate the user wants.

How does the user know what refresh rate (or any other parameter) to choose? Why not allow for all possible configuration possibilities?
--- rod.

I never realized it was in an under-developed backwards state.

Adam

1 members found this post helpful.

I had to walk a customer through resetting ie7 to factory defaults over the phone last night. It took 25 minutes to do:

open ie
click tools
internet options
advanced
reset
ok
ok
close ie
open ie


Even though she was looking at the screen and I was telling her what to click. There is no such thing as a, "GUI intuitive enough not to require the user to read any instructions, tutorials, or documentation."

In an under-developed, backwards state relative to what it could be and ought to be. Relative to programming everything with switches and levers on a 1950's computer, its in an advanced, forwards state. I find it suprising that such things as compiz and enlightment exist, but not a standard GUI for configuring screen resolution and refresh rates etc.

No one is saying Linux sucks. Its quite impressive except for certain things that are relatively lacking. Which makes you wonder why the gifted individuals who are working on other things haven't addressed these issues. Issues which aren't issues for computer people, but are issues for people with average computer aptitude. Linux is being presented as ready for the average person. No one including me is being forced to use Linux, any other OS, or to even use a computer at all. But if people around the world have gone to all the trouble to make a powerful, free OS that's almost average user ready, -then why not polish the rough edges?

I'm actually asking; what are the factors involved that this hasn't been done yet? People say Linux programming and the open source paradigm are superior so why hasn't someone focused this capacity on these few remaining anachronisms?

In the open source world, people work on what they want to work on. They scratch an itch and it turns into a project that others might find useful. Either that, or they are payed to work on something :-) Presumably no one with the skills has felt a need to scratch that particular itch and none of the commercial players involved have seen this as a pressing need. They must feel that the tools that come with gnome and kde, while perhaps not perfect and lacking some features (such as the ability to write an xorg.conf file) are good enough for selecting resolutions and refresh rates and placing monitors.

2 members found this post helpful.