The basic problem with Run Command (Alt+F2) is that it can not take "sudo" commands; sudo requires a followup query to input a password. About 95% of the commands I use in terminal require a password, such as adding ppa's or installing software from a script. Thus, the alt shortcut is 95% of the time useless.

In my search for a solution, no lens appears to exist. I found two mockups of what this would look like. The one at this link is the nicer of the two:
http://askubuntu.com/questions/259234/where-do-i-find-a-terminal-lens-for-unity
In an answer to his query, "Teester" said this: "There is currently nothing that does exactly what it pictured above since, at the moment, a lens cannot define a content area like the one pictured in order to display (and update) the output of a command. "


The other mockup was given as a solution to another brainstorm idea (http://brainstorm.ubuntu.com/idea/28565/). That idea was resolved, however did not address the issue I am bringing forth. Anyways, here is his mockup:
http://people.ubuntu.com/~komputes/term_within_dash.png
--------------------------------
edit:
Disregard the "Alt or " in the title. I learned something new today, mainly, that the Alt button is program specific.

Multiple pointer devices, e.g. one touchpad and one mouse, might require different sensitivity settings.

If mice report their model to the computer (they do, right?), reasonable default settings could be kept in a database and mabye adjusted according to how the user has adjusted the settings of other devices. Any changes the user makes should of course be remembered until next time the same mouse gets plugged in.

With or without a database, a window could pop up the first time a particular model is plugged in, allowing the user to adjust the settings.

== From merges ==
When can this be needed?

(paste from a duplicate idea)
Problem: I'm using a trackpad that I sometimes attach a USB mouse to, so I switch between using the trackpad and using the usb mouse, but when the movement and acceleration for the trackpad is sufficiently fast to move around the screen comfortably, it magnifies the usb mouse movement drastically, requiring careful cramped movement of the usb mouse to navigate because it's using the same movement speed and acceleration settings as the trackpad.

When a log file becomes old or when it reaches its maximum size, it is quite always compressed into a new file in order to rotate it with a new empty log file.

The "nocompress" and "compress" options of logrotate configure 2 different behaviours of logrotate.

If the log file isn't compressed,
logrotate will just rename it to something like 'log_file_name.log.1'.

If the log file is compressed,
logrotate will compress this file and rewrite the content into a new file with a name like 'log_file_name.log.1.gz'.

By default, /etc/logrotate.conf doesn't activate log compression. However the below packages are activating log compression in their default logrotate configuration file:
- apport
- apt
- aptitude
- checkbox
- consolekit
- cups
- dpkg
- jockey-common
- pm-utils
- ppp
- rsyslog
- unattended-upgrades
- wpa_action
- wpa_supplicant

I think that log files compression lower the system performance on desktop computers which have now enough disk space for storing old logs. I think that files compression is a good tradeoff in case the compressed files have a longer lifetime than the log files (i. e. man pages, documentation, fonts, Debian package files, ...)

[....]

Synaptics provides Touchpads and Sticks.
They are usually present in notebooks and connected via PS/2.
With the Synaptics driver loaded they provide additional features such as horizontal and vertical scroll areas, sensitivity settings, hand palm detection etc.
The driver provides a feature that enables twofinger scrolling just like with Apple machines.
If not loaded it's just a dumb trackpad.

The problem is that the X.org driver only loads when the touchpad is connected via PS/2. USB is not usable with this driver, it fails to load and the devices are unuseable.
There exists a kernel module that enables this driver to use USB devices. It is provided by Jan Steinhoff here: http://www.jan-steinhoff.de/linux/synaptics-usb.html

In 8.10 it compiles but does not work. It does conflict with usbhid module. Switching the driver causes crashes.

IBM/Lenovo provides an external keyboard with touchpad and styck that are connected via USB.

It would be highly cool it that keyboard and any other USB connected Synaptics device would work with the full featureset.

Without resorting to a completely separate distribution, a "tablet PC" meta-package would collect applications, utilities, kernel modules, and all tablet/stylus/digital-ink related packages into a identifiable block as part of each ubuntu release. User and developers working in the tablet PC world, could follow the status of this meta-package instead of searching and following the various "tablet pc" related parts as individual packages.

More than one meta-package might become necessary if the set of manufacturer specific packages becomes large enough relative to the size of the packages common to all tablets.
For example, the HAL/ACPI parts needed to detect and respond to a switch from laptop to tablet or reverse is pretty common to all manufacturers. In contrast, each manufacturer has a unique set of tablet-mode programmable keys in addition to other programmable keys that are common across manufacturers.

It might make sense to separate low level (hal, udev, acpi, module, etc) packages from user-level (xournal, cellwrite, etc) packages. The low level packages could then be further collected where manufacturer (of the laptop) specific parts are needed. Where the manufacturer (of a component) has unique features for a tablet, it might remain a common package so long as other laptops are able to use that same component (eg-miniPCI video cards).

A "Tablet PC" is a family of laptop/notebook systems with features that enable use that some might call an electronic legal pad. These features include (but are not limited to)
display in both landscape [laptop mode] and portrait [tablet mode], input using a traditional keyboard [laptop mode] or a virtual keyboard [tablet mode], pointing using a touchpad or eraser-mouse [laptop mode] or a stylus [tablet mode], extensive use of programmed and programmable feature keys,
hardware detection of laptop or tablet mechanical orientation, and so on.

Once upon a time, there was a group trying to create a tablet pc specific ubuntu variant -- (?)TABUNTU(?). It seems to this author that that effort is either stalled or lost in the dust as other aspects of ubuntu evolution changes so rapidly (eg - xorg vs. mice and pointers and displays, etc) as to render tablet pc work impossible.

If you have a plain input box for text (1 line), in HTML terms

and you want to select certain part of entered text, unless you keep your mouse exactly over the text box, it's impossible.

Example:
The box contains "here is my very-long-text to be selected".
If I want to select "very-long-text", I have to start at "v" and keep my mouse until "t" over the box to actually select.

When accessing a hard disk/flash memory etc where there is a lot of IO activity, there should be more affect to the responsiveness of my mouse (touch pad in this case) or keyboard. Currently when I copy / move photos from my flash card (via my card reader on my laptop) to my hard disk, my mouse pointer becomes vary jumpy and unresponsive and it is almost impossible to reliably click on anything. This should not happen. An effort should be made to reduce and ultimately remove this unresponsiveness.

There should be some way of making the system tell you what you just did and what happened as a response.

Sometimes, I hit the wrong keyboard shortcut by accident and do something that I didn't intend to, but might be useful in the future. In that situation, I'd like to know what I did so I can do it again in the future when it's useful.

This especially happens when I'm new to an application; so for users new to linux, meaning they're new to _all_ the applications, it might speed up their process of discovering what the applications can do and how.

[Edit: this is *not* meant to replace undo. Sometimes, often, you just want to undo, but in some cases you also want this.]

It would be good if there was a program that hung back and collected X events. On request, it'd display them back to the user in a human-friendly form; something in the style of this:

- You scrolled up five times
- You moved the mouse from (x,y') to (x',y') [click to see motion path]
- You clicked the left button
- Window "Firefox" was minimized
- You pressed Ctrl-Alt-Right
- You switched to desktop "2"
- You pressed Alt-F6
- You switched to window "dialog" belonging to "application"

My mind hasn't set in stone how the UI should be; maybe as a panel app with the label "What just happened?" Also, the granularity of the events is open to discussion. The typical scenario requires to look at a few of the most recent events, 95% of the time at most 20 (making up numbers is fun).

I think it'll take a lot of time to implement in *every* application, but maybe something can be done at the toolkit level, such that at least all _GNOME_ applications can report more detail like "you opened menu 'file'", "you pressed button 'button0'" You could of course do similar things with other toolkits (QT, Motif, Xt, ...).

[....]

Idea ;

Advances

In the event that a GUI system becomes available for accessing the CLI (command line interface), the following additional advancements could be made;

1) Commands could be arranged in pipes by using a graphical manipulator similar to the patchboxes and GUIs used for linking software audio devices.
The command is given its own "black box" (click on it to see internal properties). It has its input and output. Connect the boxes together with lines to form your command pipe.

2) Drag and drop method of concatenating commands, taking output from one process to the input of another.

3) "black box" could be any thing from a GUI created script, through to a fully automated process done by a large program.

An example of an automated program in a pipe;
(use of Graphics editor and FTP programs)

Inside one box;
Open picture in editor,
detect extremities (head feet arms) of subjects,
remove tilt of camera,
crop to extremities of subject,
color adjustment,
resize to common width,
save as jpeg optimised to 20 to 30kB,
close editor.


[....]

Have you ever changed your languages to Chinese accidentally (assuming you do not know Chinese)?

The problem of Ubuntu, and Linux in general, is input method support.
Most people around the world use the same keyboard layout but not input method. When an average Joe comes up in front of a computer, what will he do? Isn't it typing document or surfing the internet? What can he do if he can't input his language. So why does he stay? They even can not read and thumb-up for this post.

Localization is a critical things, it makes things usable.
Language translation works very well thank to Launchpad contribution, but the same does not apply to SCIM developers.
If you take a look at any of Linux forum, you'll known how painful to get SCIM work. And in most case, SCIM works only partially (buggy) or not at all.
All we want is improving support of input method. Do something, do anything... to make it better.