You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
365 lines
14 KiB
365 lines
14 KiB
|
|
These are the results of a Konsole Usage Survey conducted in March 2007 to ask |
|
about how users work with the command line, how they use Konsole as it exists |
|
in KDE 3.x and to ask for opinions on ideas for future features and improvements. |
|
|
|
The survey received ~8000 replies, these are not full results, but a summary |
|
of the results for each question, including the "additional comments" text boxes. |
|
|
|
Authors: Robert Knight < robertknight@gmail.com > |
|
Updated: 29-03-2007 |
|
|
|
PLEASE NOTE: The counts given for responses to the text-box questions |
|
(ie. the 'any additional comments' questions) are rough |
|
and should not be treated as absolutely accurate. |
|
Raw data: |
|
|
|
1. Operating System |
|
|
|
The vast majority ( 97% ) use Linux, followed by BSD ( 2% ) and Solaris ( 0.3% ) |
|
|
|
2. Frequency of Command-line Usage ( compared with equivalent graphical tools ) |
|
|
|
The average respondant classified himself as a frequent |
|
user of the terminal ( average rating of 4.2 out of 5 for frequency of usage ) |
|
|
|
3. Additional comments about Q.2 |
|
Most replies were details of what the respondants used Konsole for and how the command-line |
|
|
|
Reasons for using the command-line include: |
|
|
|
- Automation (scripting,batch processing) XXXXXXXXXXXXX |
|
|
|
The 'building block' operation of the command-line using pipes was |
|
mentioned by a number of respondants. |
|
|
|
- Package Management XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX |
|
XX |
|
|
|
A number of users ( presumably Kubuntu users ) cited the poor |
|
quality of the Adept package manager as a reason for using the apt-get |
|
package management tool instead. |
|
|
|
- Working with remote computers (general) XXXXXXXXXXXXXXXXXXXXX |
|
(includes a number of web developers) |
|
- Working with remote computers (sysadmin) XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX |
|
X |
|
- Efficiency XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX |
|
XXXXX |
|
|
|
A number of users cited using Konsole ( or ALT+F2 ) to launch |
|
GUI programs because it is quicker than using the mouse to |
|
navigate the K-Menu to find the application icon and activate it. |
|
|
|
A number of users also cited the relative inefficiency of the mouse |
|
compared with the keyboard for experienced power-users. |
|
|
|
As well as input efficiency, some respondants mentioned that |
|
startup time of GUI applications under Linux/KDE meant that using the |
|
command-line was much more efficient. |
|
|
|
- Monitoring system XX |
|
- Reading mail XXX |
|
- Performing tasks as an administrator (eg. using su/sudo) XXXXXXXXXXXXX |
|
- Local system admin XXXXXXXXXXXXXXXXXXX |
|
- File operations XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX |
|
XX |
|
|
|
basic commands - cp,mv,cd,ls,rm etc. but also mc in many cases |
|
|
|
- Using version control tools (svn,git etc.) XXXXXX |
|
- Command-line multimedia XXXXXXXXXXXXX |
|
|
|
mplayer in most cases. ffmpeg automation often cited as well. |
|
|
|
- Text editing XXXXXXXXXXXXXXXXXXXXXXXXX |
|
|
|
vi in most cases. some mentioned emacs |
|
|
|
- Writing code XXXXXXXXXXXXXXXXXXXXXXXXXXX |
|
- Building software XXXXXXXXXXXXXXXXXXXXXXXX |
|
- Injury from mouse usage X |
|
- Enjoy S&M/Use Gentoo XXXXXXXXXXXXXXXXXXXXXXXXXXXXX |
|
- Enjoy hardcore S&M/Develop Gentoo X |
|
- Missing GUI tools XXXXXXXXXXXXXXX |
|
|
|
A relatively small number of respondants noted that usage of the command-line |
|
has decreased in recent years due to improvements in GUI tools. |
|
A number of respondants also described themselves as long-time Linux or UNIX |
|
users who preferred command-line tools because it is what they grew up with, |
|
and new-fangled GUI applications weren't available back in the day. |
|
|
|
A number of users also mentioned that while GUI applications were available, |
|
they trusted the equivalent command-line tools more. |
|
|
|
A small number of users cited the lower system requirements of command-line |
|
tools compared with the graphical alternatives. |
|
|
|
BSD users in particular cited the lack of satisfactory GUI tools for |
|
various system administration functions. |
|
|
|
- Verbose output from commands XXXXXXXXXX |
|
|
|
Includes debugging problems with GUI applications, particularly those |
|
testing or developing bleeding edge software. |
|
|
|
- Database admin XXX |
|
- Running legacy (ie. terminal-based) application X |
|
|
|
A significant number of respondants also indicated that |
|
they use Yakuake (a program |
|
which provides a drop-down terminal display at the press of a key |
|
in the style of the ID software game Quake), the most common |
|
reason given by far was the instant availability of a terminal at |
|
the press of a single keystroke ( F12 by default ). |
|
|
|
Reasons for not using the command-line |
|
- Lack of discoverability of functions vs. GUI applications |
|
- Learning curve is initially steep |
|
|
|
4. The most popular tasks in descending order: |
|
|
|
- Administration of local machines |
|
- Administration of remote machines |
|
- Building Software |
|
- Viewing manuals |
|
- Text editing |
|
- Debugging software |
|
- Writing code |
|
- IRC |
|
- Email |
|
|
|
5. The vast majority ( 90% ) of respondants used Bash as their |
|
primary shell. Zsh was used by 4% of respondants. |
|
Alternatives mentioned in the "other" section included ksh, |
|
Midnight Commander and Python. |
|
|
|
6. In decending order, the most popular interactive terminal programs |
|
were: |
|
|
|
- SSH |
|
- man / info |
|
- vi / vim |
|
- GNU Screen |
|
- Midnight Commander |
|
- gdb |
|
|
|
7. Asked for their preferred colour scheme for Konsole sessions, |
|
users primarily chose between one of 4 colour choices supplied |
|
with Konsole. |
|
|
|
- Black on White ( 25% ) |
|
- White on Black ( 28% ) |
|
- Linux console ( Silver on Black ) ( 22% ) |
|
- Light on transparent ( 18% ) |
|
|
|
8. Additional comments for Question 7. A large number of respondants |
|
answered this question. |
|
|
|
Popular alternative colour choices mentioned included: |
|
|
|
- Green on Black ( Matrix-like ) |
|
(Probably the most popular colour not mentioned on the original list) |
|
- Dark background, transparent |
|
- Dark text, transparent background ( Relatively popular ) |
|
- Black text on a light, but non-white background |
|
(To reduce glare from monitor) |
|
This light colour was most often Yellow. |
|
(Probably the second most popular choice) |
|
|
|
Grey background was sometimes used. |
|
|
|
- Light text on a dark background |
|
- Yellow text on a dark blue background |
|
|
|
- Several respondants mentioned the use of a random background colour |
|
- Many respondants indicated that true transparency would be highly |
|
desireable. |
|
- Many respondants also said that true transparency was not useful |
|
for them. |
|
- With transparency, the most appropriate colour scheme depeonds on |
|
the desktop background. |
|
- Several respondants said that a white background was too glaring. |
|
- Several respondants requested a custom background colour (often |
|
one with a redish hue) for root sessions. |
|
|
|
Problems mentioned: |
|
|
|
- Standard Konsole colours not working well with 'emerge' coloured |
|
output. |
|
- Default BLUE colour ( I think this refers to the Linux console colour |
|
scheme ) does not work well in Vim or other applications. |
|
- Existing colour scheme editor is hard to use. |
|
- Existing 'fake transparency' has performance issues. |
|
|
|
9. The majority of respondants do not change their colour scheme |
|
depending on the computer they are logged into, but a significant |
|
minority ( 22% ) did. |
|
|
|
10. The majority of respondants do not change their colour scheme |
|
depending on the program running in the terminal, with 12% |
|
claiming that they did. |
|
|
|
11. The majority ( 77.5% ) of respondants used a colour scheme which |
|
came with Konsole. 22% used a custom colour scheme. |
|
|
|
12. The most favoured graphical enhancements to Konsole were, in |
|
descending order: |
|
|
|
- Real transparency |
|
- Konsole as part of the deskop background |
|
|
|
With the other choices several points behind on average: |
|
|
|
- 24bit colour |
|
- Background images |
|
- Shadowed text |
|
- Fake transparency |
|
|
|
13. The default session type which simply runs the system shell |
|
( typically bash ) was the most frequently used session by far. |
|
|
|
The "Root Shell" and "Linux Console" sessions were the next |
|
most frequently used types, followed distantly by Screen. |
|
|
|
14. The majority of respondants (84%) had not created their own custom |
|
session types, although 16% ( 283 ) had. |
|
|
|
15. Popular uses for custom session types: |
|
|
|
- SSH: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX |
|
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX |
|
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX |
|
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX |
|
XXXXXXXXXXXXXXXXXXXXXXXX |
|
|
|
( Often used custom colours to distinguish between machines ) |
|
( Some of these turned on 'send to all sessions' - very useful |
|
for performing the same command across many machines ) |
|
( Users with many SSH session types complained of the menu |
|
growing too large. |
|
They would also appreciate the ability to bookmark all open |
|
SSH sessions ) |
|
|
|
- Other remote connection: XXXXXXXXXXXXXXXXXXXXXX |
|
|
|
( Including databases, telnet ) |
|
|
|
- Interactive Interpreter (Python,Ruby etc.): XXXXXXXXXXXXXXXX |
|
XXXXXXXXXXXXX |
|
|
|
ipython was a popular choice, also irb ( ruby ). |
|
|
|
- Custom shell: XXXXX |
|
- Custom Screen-based: XXX |
|
- Tabs opened in specific directories: XXXXXX |
|
- Custom environment (eg. a 'KDE 4 development environment'): |
|
X |
|
- Customise the default: XXXXXXXXXXXXXXXXXXX |
|
|
|
- Often to change colours or font size |
|
- Or to make the default a login shell ( eg. bash --login ) |
|
|
|
- Disable various UI features (scrollbars,menu bar etc.): XXXXX |
|
XXXXX |
|
X |
|
- Custom root shell: XXXXXXXXXX |
|
|
|
A number of respondants did not understand the concept of a |
|
'Session' in Konsole. |
|
Others found it easier to use scripts to set up different |
|
environments. |
|
Some respondants actually discovered this feature as a result |
|
of answering the survey question. |
|
|
|
Invoking a session without a keyboard shortcut requires use of |
|
the mouse. This is inefficient. A number of respondants suggested |
|
that sessions should be activated automatically when certain programs |
|
are started etc. |
|
|
|
Some people need to interface with machines or applications which |
|
work better with different terminal, keyboard or encoding settings. |
|
|
|
A notable number of respondants complained that the user interface |
|
for creating new sessions or modifying the default session type |
|
was hard to understand. |
|
|
|
16. Around half of respondants ( 42% ) did use GNU Screen, a slightly larger |
|
proportion ( 57% ) did not. |
|
|
|
17. The most favoured candidate for better integration with Konsole was |
|
the SSH program ( 43% ), followed by Vim ( 16% ) and GNU Screen ( 11% ). |
|
Others mentioned include Konqueror,mc,nano and emacs |
|
|
|
18. The rankings did not give clear answers about the ease of performing |
|
various tasks in Konsole, but the additional comments ( Q #19 ) gave |
|
many useful responses. |
|
|
|
19. [ TODO ] |
|
|
|
20. Typical number of Konsole windows open: |
|
|
|
1 ( 40% ) |
|
2 ( 29% ) |
|
3-4 ( 19% ) |
|
More ( 11% ) |
|
|
|
21. Around half of the respondants used small Konsole windows occupying |
|
a quarter of the screen or less. A significant number ( one third ) |
|
typically used maximized Konsole windows. A further 10% used |
|
vertically maximized Konsoles occupying half of the screen, and a further |
|
7% used horizontally maximised Konsoles. |
|
|
|
22. 40% of the respondants typically had 2-3 tabs open at once, |
|
26% typically had 4-5 tabs open, 10% typically had 6-7 tabs open. |
|
Only 7% had more. 15% claimed to only have one tab opened typically. |
|
|
|
23. The preferred default naming for tabs were, in descending order: |
|
|
|
- Foreground program ( 24% ) |
|
- Computer logged into ( 20% ) |
|
- Current directory ( 17% ) |
|
- Session number ( 12% ) |
|
- Window title set by session program ( 12% ) |
|
|
|
Most user suggestions were a mix of the above, depending on context. |
|
Popular user suggestions: |
|
|
|
- computer:dir ( Might be hard to implement ) |
|
- mix of dir and program |
|
- according to bookmark name |
|
- computer:program ( Again, might be hard to implement ) |
|
- customiseable per session type |
|
- session type name |
|
- user@computer |
|
|
|
24. The majority of respondants ( 68% ) did not typically rename the tabs |
|
manually after they were created, although a significant minority ( 25% ) did. |
|
|
|
6% did not know how to rename tabs. |
|
|
|
25. The majority of respondants ( 63% ) did not change the tab font |
|
colour manually. A significant minority ( 12% ) did occassionally. |
|
A significant percentage |
|
( 22% ) did not know how to change the tab colour. |
|
|
|
This could be attributed to poor wording of the question, which asked about |
|
"tab colour", when it meant "tab font colour". |
|
|
|
26. The most popular methods for choosing tab icons were: |
|
|
|
- According to foreground program ( 48% ) |
|
- According to session type ( 35% ) |
|
- Random coloured icon for each new tab ( 7% ) |
|
|
|
Popular user suggestions: |
|
|
|
- According to machine logged into |
|
- Disable tab icons |
|
(Probably the most popular actually) |
|
- Don't care |
|
|
|
|
|
27. [ TODO ] |
|
|
|
28. [ TODO ]
|
|
|