I asked if there was a way to do something specific. Nowhere did I say or imply that I refuse to learn things. On the contrary, I said I will learn what I need to.

I pointed out the advantages GUIs have, but also acknowledged the usefulness of CLI. Nowhere did I insist on using a GUI.

It seems hardly anyone here is bothering to read what I have written.

I also get the sense that anyone who dares to wonder if there's a better method for doing things is to be regarded as some sort of infidel and treated with contempt or hostility. Am I learning an operating system here, or a religion?

If the answer is "No, there is no quick way to elevate an app from its shortcut," couldn't that have been said without all the condescension?

But that is the choice of the sysadmin.

The sysadmins where I work have always chosen to install GUI software on all the Linux servers, even though most are rack mounted and none have full time monitors and keyboards (most don't even have connection to a KVM).

Compared to the size of everything else on a server hard drive, the optional GUI software is trivial. The extra swap space (probably under 100MB) to swap out all the open but idle GUI stuff is also trivial.

It is very easy to configure Centos so you can log into it as GUI from any Windows or Linux workstation with X support. We also configure them so that we can use a command line login (putty) to launch a desktop in a VNC server, so that one can then use a VNC client instead of an X server on the physical workstation to run GUI on the rack mounted server (The Windows X servers I've tried tend to get confused by the multi monitor setup on my Windows machines, so VNC client is more robust than X server).

The nature of some of the servers here is such that X or VNC GUI access by non sysadmin users is often useful. They aren't the type of servers where non sysadmin access is only through the strictly defined services of the server. But sysadmins also use GUI access and even on the servers whose non sysadmin use is strictly limited to the defined services.

It isn't about creating applications, it's about using the ones that are already there. If you have a GUI application that does exactly what you want, then there is nothing to stop you from running it with root privileges. How do you run your GUI iteratively, with different arguments on each iteration?
It isn't about feeling good about anything or about learning things without purpose. Learning isn't a bad thing, and to suggest one shouldn't need to just sounds foolish.
Okay, so who's mischaracterizing now? No one suggested a commandline tool to replace every GUI, although in most cases, it could be done. Since you want to be running the commands with root privileges, it is reasonable to conclude that you are doing system administration tasks, since that is what root privileges are intended for. It is hard to imagine a scenario where browsing online forums requires root privileges.
If there is no repetition, how many seconds are saved by finding the icon you want to click on, elevating to root privileges, and launching the GUI? If there are commonly used GUI tools that you need to run with elevated privileges, then tools like gksu exist, and can be used in conjunction with your desktop menuing system. Sysadmins tend to avoid those, because they just want to do something they know will be present on any system they touch.
It is a lot easier to make assumptions about what commandline tools are available to online posters, than what GUI tools. If one searches for solutions to problems, the likelihood of finding a solution based on a commandline or script is much greater, since it is more likely to be found and characterized by a search engine, and because there is significant likelihood that a commandline or scripted solution will actually work on your system.
In the arena of system administration, this is even more true. Distros are fairly uniform in what commandline utilities are present, since most include the standard binutils package. It isn't generally feasible to predict what GUI tools are available, since those are less uniformly distributed, and much more subject to configuration by distributors and end users. Similarly, as a sysadmin, not all hosts that you touch are necessarily configured the same. I touch different systems with KDE, Gnome, and other desktops of various versions and which are configured according to the requirements and whims of end users. What happens when I right-click something or shift-click something is often different on different hosts, and sometimes changes from time to time. When I type a bash commandline, I rarely encounter inconsistent or surprising behavior.
As a great deal of system administration tends to be done through remote logins to headless servers, using GUIs on those tends to just get in the way. If you are already running a GUI desktop on your local workstation, the confusion that would arise from trying to run a desktop or window manager on each remote host (and, yes, it can be done) would far outweigh the benefit of simply typing commands. Using remote desktops just adds more confusion: if I right-shift-control-click on the remote desktop, will the result be interpreted locally, or on the remote host?
While I sympathize that there is some culture shock associated with moving from one style of work to another, I think it behooves you to consider that at lot of people that have gone before you have found very efficient ways of working without GUIs. There have been enough bright motivated people using Linux/UNIX for long enough that if they could get more done using GUIs for system administration, it probably would be more prominent by now. When you've proven us wrong, we'll all be happy to agree when you say 'Not good at all.'.

--- rod.

I hope someone working on Linux desktop software recognizes that would be a good feature. Maybe it even exists somewhere and those answering in this thread just don't know about it.

In Windows that feature is a poor, but necessary, solution to a very severe problem. In Linux, you certainly cannot argue that it is necessary. The other ways around the issues are much easier in Linux. And there aren't as many situations in Linux in which you even need something like that.

So I understand the natural resistance to making a Linux copy of what is essentially a work around for flaws of Windows. But beyond being such a work around, it is also a useful feature worth copying.

BTW, in my own use of Windows 7, this work around has proven totally inadequate. The intent is that one could have a user with admin rights run with such rights in a normally disabled mode (almost like being logged in without admin rights) and only enable the effective admin rights when you really need them. I'm sure many people use Windows 7 in a mode where that model works. But some of us still use computers for more complicated activities in which the fundamental Windows assumption still exists that a power user always has admin rights. Too many things running on my Windows 7 system simply don't work unless I have constant admin rights. The stop and/ask approach would be unbearably tedious even if it worked. But when I tried to use it, it didn't even work. Say yes to everything that asked for admin rights and the activities still fail with security violations. So I use Windows 7 the same way I use XP, with constant admin rights and constant risk that any malware I touch will destroy the whole system. Linux is safer, but for various reasons, Linux is not a practical choice for the two computer on my desk at work.

Meanwhile, for the main question of this thread, I still have no better answer than multiple logical desktops. Having used that, I will claim it isn't a terrible solution. Probably I would have preferred what the OP was asking for, but I don't have it. I have also use gksu and similar tools. In certain limited situations, those are the best answer. But I don't think that addresses the basic convenience issues.

Half the time we see a beginner "how to" question in this forum, the more constructive answer is "here is why you should not do that" rather than "here is how to do it". That is true even if we know how to do it.

Obviously, I disagree with the majority of experts responding in this particular thread over whether this is one of those situations in which "why you shouldn't" is a better answer.

But please remember that even those answering with "why you shouldn't" are trying to be constructive (and in this case there were important things to read in those posts even if you disagree with the main message).

One more detail:

I'm still using Centos 5, so maybe by 6.3 they fixed the GUI package manager.

I am enormously biased in favor of GUI. I really wanted to like the GUI package manager in Centos. But it is crap.

A package manager is one of the places I think a GUI is just fundamentally better than command line. Synaptic (GUI) in Debian based distributions (again not exactly current) was so much better than apt (command line).

But you need to temper bias with reality. Yum is a great command line package manager in Centos.

If you try GUI package management in Centos and find it is still terrible, don't jump to the conclusion that package management is terrible in Centos. Try yum instead (try yum anyway, even if the GUI isn't terrible, I bet yum is still better).

No, Windows can be administrated without a gui. It's just that a gui is convenient and efficient enough that there is often no point to using the cli, or scripting changes.

It's about efficiency. And I was asking about using an existing application.

What was suggested was that I give up using a GUI that was otherwise meeting my needs.

That really wasn't the point of the analogy.

A lot, if I need to run the same app multiple times, on an as-needed basis, without necessarily implementing the same settings every time.

Which works, but is less efficient than what I'm accustomed to. And that's why I asked. I wanted to know if there was a way to perform a task as efficiently on Linux as I'm used to with Windows.

It doesn't seem that way at all. As I search for how to do something, I have to constantly make sure the series of commands are valid for my distro, my version. Following commands for different distros or versions will result in immediate failure, or successfully making changes that shouldn't have been made. E.g. even from the CLI what works for Fedora stands a good chance of being inapplicable to CentOS. Do the commands work the same? Sure. But that's of little help when what you're modifying with the commands is different.

The GUI doesn't have to be on the target machine at all.

What is there to be confused about that couldn't happen with CLI? You pay attention to what machine you're connected to, whether you're using a gui or not.

How do you know if a key combo is going to do something to your local machine, or the remote machine via cli? I know what the rdp passes through and what it doesn't.

The GUI is prominent in system administration. It's used extensively in a very common OS called Windows Server, designed and used by a lot of bright and motivated people. Also, the linux server OS I downloaded has gui apps for doing the things I want... Gui apps that I'm sure didn't exist until some bright and motivated person said, "Hey, you know what would be better than doing this from command line? Being able to do it from the gui." And it doesn't seem to exist right now, but it's probably a matter of time until a bright and motivated linux dev says, "You know what would be better than gksu? Shift+right-click." It hasn't been done yet, but that isn't evidence that it shouldn't be done.

Anyways. I've my answer to my question now.

Thanks.