“I’ve always been comfortable using the command line interface to get specific tasks done. […] I already knew that I could do pretty much anything from the command line if I was willing to sit down, read manual pages, and learn — or if I really had to. To prove it, recently I forced myself to use only the CLI for a week. I ended up learning a lot more than just a few command line arguments.” Read the article at NewsForge. Jeff also writes: “I used Lynx as my browser; I don’t really like Lynx, but what else is there?” May we suggest eLinks 0.10.x, Links and w3m? They all have way better rendering than Lynx!
A week in the BSD CLI
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Links with svga would have given him the graphics he wants. I use links all the time, it could easily replace mozilla in the CLI.
Links is one of my favorite browsers. I use it on my systems a lot just for speed sake
a little OT.. does links support javascript?
I like w3m better for the reason that it can navigate web pages using “vi” style comands.
Or you can also tab yourself to death if you like. That just jumps around way too much for me.
Links version 2 does, and it can do graphics with svgalib or x
> a little OT.. does links support javascript?
Yes, Links 2.x ( http://atrey.karlin.mff.cuni.cz/~clock/twibright/links ) supports Javascript. It doesn’t support everything, but it does.
cool. thanks.
Favorite quotes:
Using a GUI makes you think in a broader, more synthetic, and object-oriented way. (…) you think of your programs and data in terms of how they look on the screen.
In the command line you’re forced to think in a more analytical way; you have to picture where your data is, where you want it to go, and how you want to view or manipulate it.
I always found that working with the CLI was a spatial experience (where am I? where am I going? what can I do here?) whereas the GUI is more of a static experience, like being seated and playing with one’s toy. In a GUI, you manipulate the interface, in the CLI, the interface manipulates YOU!
Not to imply that CLIs are manly and teach you about the meaning of life, whereas GUIs are for babies who are staying safely at home
, but I find that the kinetic aspect is a fundamental difference between the two paradigms.
Im not sure if this applies to BSD, but I like the ability to use links in frame buffer mode. It doesnt need a X server but you get almost full graphical capabilies.
Actually, you can use the mouse in the FreeBSD console, if you start moused. I don’t know if it exists in openBSD, though.
It’s suprisingly useful being able to select and paste text with a mouse in a CLI.
Speaking of CLI, is it possible an OS with applications based on ncurse?
I use elinks in CLI. I believe it forked from Links. Has alot of neat features and stuff. It’s easy to use, supports frames, and stuff. It’s pretty good.
Make sure you use 0.10.x, the latest (maybe even CVS I think). I did some beta testing for it a few weeks ago and we cleaned up lots of rendering bugs for it with its author.
some progs i like:
links2 — has text and graphics modes, i’ve been using “-g -mode 1024x768x256”
naim — ok program, i will try centericq next
mutella — p2p program
screen — excellent for shell accounts, supports detaching and reattaching, cut-and-paste also
wvdial — easy way to configure dialup
elm — works better for me than pine
framebuffer — i use 800x600x256 (vga=0x303 appended to loadlin)
slrn — Usenet
gpm — cut and paste with mouse
for GUI minimalism:
Ion2 — been using this a couple of days, nice WM
Dillo — small footprint
xterm — this, rxvt, aterm, or gnome-terminal will work
i plan to test mutt or fetchmail next.
the basics of screen:
control(a) c — new “window”
control(a) S — split screen
control(a) tab — move to the lower/upper window on split screen
control(a) brackets — cut / paste
exit — kill window
screen -r pid — reattach
control(a) d — detach
dont forget raggle, text based rss reader ;-).
http://www.raggle.org
elinks (CVS) now has JavaScript support, too. Unfortunately, it’s currently all-or-nothing, so you’ll get those unwanted pop-ups if you turn it on. Which is kind of twisted — ad pop-ups on a text-mode browser?
I used to use Links 2.0 for certain tasks — I hated the svga mode (it looks horrible and doesn’t work properly with GNU screen), but even in text-only it supported a subset of JavaScript. With elinks-CVS, there’s no reason for me to have Links 2.0 around anymore.
Oh, and naim’s realy nice — especially when coupled with screen.
I felt the same way in the beginning, but after a while i stopped using file managers and used a cli instead
By the way, Links 2 supports graphics via framebuffer too.
My ideal OS would be freebsd(or even an ubuntu-like Linux, really) with a fully ncurses(or even turbovision, think turboC / turbopascal) interface. All the basic auto-discovery plug-and-play style stuff, ofcourse, too.
But an ncurses native system that supported images(like many of the above links screenshots do) with a mouse and a native screen used by default on all programs where I can attach and detach my entire desktop at will.
A terminal/cli would just be an ncurses window(like xterm) and a nice tui/ncurses interface to ports would be great. Imagine a ncurses-version of ubuntu’s gui package manager.
That is my dream system, really.
The point of the exercise was not singular. To begin with, I wanted to get into OpenBSD a little more. Secondly, I wanted to know the CLI better because it’s possible that someday I’ll be stranded there and will need to know how to do things like retrieve email, get on IRC with a usable interface, and work on articles. Thirdly, my notebook system’s screen only supports 800×600 in X, and that’s kind of annoying. It also uses more power in X, and I want to be able to extend the battery life while keeping the software as simple as possible. If I had to replace the hard drive, I could put OpenBSD and all of those programs back onto the computer in less than a half hour. Not so with most GNU/Linux distros.
-Jem
I fail to see the value of this exercise. This is not an “exercise”. There are actually some very good reasons to be able to work in CLI mode:
1. Believe it or not, many kinds of tasks can be done much faster from a command-line mode. For example, any kind of task that has you hopping all over your filesystem is much more quickly done in the command line than by clicking up and down, in and out of different folders in a file manager. Even though I use KDE on my desktop, when I have to quickly read/edit many files, I usually open a terminal.
2. Wonderful tools such as >, >>, |, <, grep, etc make it very (extremely) fast to manipulate text in and out of many different files on your system. There is currently no equivalent in a desktop environment.
3. Being familiar with all the CLI tools means you are completely at home administrating a Unix system halfway around the world, just as if it were right beside you. If you rely on GUI tools to administrate a system, you are in for a lot of frustration with remote management, even for the slickest and most expensive of these tools.
4. Servers should never run a GUI anyway, as they provide a significant unnecessary drag on a system.
I am sure there are some greasy,long haired and bearded Unix diehards out there that refuse to evolve, but why force yourself into a lesser effecient means of working? Its the equivalent of a back hoe operator writing an article about digging a 6 ft deep utility trench with a shovel just for fun.
This is just laughable. As mentioned above, the CLI provides a rich syntactical way of dealing with your system, which the GUI does not. With a GUI, its like trying to communicate without verbs and adverbs. Anyway, it’s not a question of evolving; there are simply some things which work better without a graphical metaphor to get in the way, while other things don’t. I would hate to try and draw a picture using words, but I would also hate to apply for a loan using only pictures.
Most of the responses above are written by people who
know nothing about OpenBSD and what software is
available for it. (No better than the so-called
author of the quoted article 😉
Do a little research before wasting bandwidth.
Well, sure, but the point is – you can do all those things in a graphical environment with a terminal window, as you (and I) do. Unless you have extremely poor hardware it’s hard to see why it’s better to use nothing but the command line environment than to use both.
Or better , whats wrong with me ?
What year are you living in ?
Is it 2004(2005) ? Or we are still in the mid 60’s ?
Text interface is good if you have a small PDA, portable or anyother kind of small device.
Nowdays even the worst factory display manager has a nice GUI.
When i see you beeing so content about this i imagine the feeling beeing just like saying :
These wheels from the wheelchair are spinning so in circle, it is all i need to “WALK” ?!?!?!?
Exactly ….
well, if you have a server and you need all the hdd space you can possibly muster out of it, pretty guis quickly become irrelevent if you can do all you need to do via cli. an extremely minimalist install can save you a good amount of space.
The CLI is very well developed, and programs play together almost sexually; it’s just that amazingly good.
But it’s simple, there is input, process, output. CLI is very easy to design for (this discludes curses, that’s text based interactive interfacing).
The GUI is a big complicated mess, with so many options and so little time or means to stick a button here and a button there.
I see GUI continually improve, but it’s a harder road to design for. Realistically though, in many tasks, the GUI increases productivity a lot. It’s very helpful to be able to grab your mouse, scroll down 500 lines with three clicks, make two small changes, hit f9, click the error and have it open that file and put you on that line….etc etc.
A little off subject. But one thing that is ever present to me after switching from a Windows world to a *nix world. Windows is very memory oriented, saving is something to help you when your computer crashes. *nix is very file oriented, you save files often just to open them with another program. I imagine this has to do with unix history, and machines with comparitively no memory.
> Is it 2004(2005) ? Or we are still in the mid 60’s?
2004, but GUIs lack some features. You can’t redirect the input and output, you can’t “connect” it with pipes or use logical operations. Therefore for every new feature you will need a new program or a new version (or a plugin if the program is modular and the feature is feasible with a module. GUIs aren’t flexible. Almost all GUIs aren’t small (KISS) or you can only use your mouse (3/5 keys against 104 keys on your keyboard).
Yes, we have the year 2004, but we don’t have usefull concepts for GUIs. That’s the problem.
@iongion
I bet you never tried to use small memory, headless or remote machines.
Just as what rycamor said, trying to use GUI on aforesaid situations is a pain in ass. Try remote-x some machine across the alantic and you will see. WTF would one bitch with GUI across a low latency link while a simple ssh just works?
There are just too many scenerio that GUI is too troublesome. “Desktop” type of application runing on local net is better GUIed, but it simply can’t replace CLI on remote management.
@ iongion:
If CLI is so 60’s, why would Microsoft invest in CLI improvements?
The command line interface of windows 2000 is better then DOS (tab completion and the like) and the CLI of Longhorn will include Monad, the new Windows Shell. I’m currently beta-testing it and while it doensn’t feel as complete and natural as bash, it is object-oriented which creates very interesting possibilities.
I actually started to play with Linux because of it’s nice command line interface. It’s often easier to work wih than the countless windows where you can get lost easily.
One year ago when I was in the 2nd year at aerospace engeneering faculty I concluded a study about brain reaction in stress situations and the points that camed out where the facts that:
All has to do with the brain in the following ways(do’s and dont’s):
1) do not show the brain a verry fragmented structure with no details.
2) under high stress conditions, do no let the brain pick the structures from a sea of details with no meaningless design.
3) think at big scale, present the brain with noticeable big structures, think in an object oriented manner, it stimulates the decision time and the rate of beeing right.
CLI has his own deserved place, but what makes me think at it in a bad way is that I see colleagues with big machines(2Gb RAM, 4x250Gb HDD) that keep on having a CLI as their way to do stuff, even if they just jumped from another OS, somehow it gives them a geeky feeling or something, is just like: if you are using CLI then you are for sure a bigbrainer.
GUI is didactic, is teachable, it can be spoken about.
(Try to communicate about doing something in CLI with a non CLI expert, you will see that even though you tell him exactly the steps it must follow, he would either ask you for a pen or he will forget beacuse of too many structures, too many divisions, interdependend with just a few that really matter but obturated by some sintactic rules)
system administration should be thoughtful, not something that is desinged to be handled with stress.
if you have no experience with a GUI, you can’t speak about it either.
CLI commands can be typed, emailed, spoken of… GUI screens need much more explanation, like “click there and there”, most users have difficulty with that.
having a big machine is nice, and CLI tools are nice because they are powerful. the new microsoft command line interface will let you do things you cannot do in your GUI. CLI is made for programmers because you have much more control over what, when and how things happen.
well, for weeks on end i used to use the command line exclusively.
screen with vim, irssi and mplayer is all anyone really needs
recently ive taken the time to use ratpoison for a while, if you like screen youll like ratpoison. though change C-t to C-s or something, C-t is a rather long reach.
of vague relevence, has anyone heard of the commandline brotherhood? i saw a reference in bb once and have been interested ever since.
i kind of want to join, or create if it doesnt exist (anymore?)
anyone else interested?
I pretty much did the same thing back in 2002 when I was learning to use NetBSD. What I didn’t care for was the column/row dimensions of the NetBSD console, as NetBSD did not (at least at that time) have anything like the Linux framebuffer on x86, so I ended up running X without a window manager to get big xterms. I used .xinitrc and the -geometry switch to set up all xclients.
You can see my Dec 2002 screenshot of this at:
http://aknaton.phpwebhosting.com/SS/2002-12-24–.jpg
I used Links at a web browser, Mutt for mail, and Vorbis Tools for audio. It did 90% of what I needed. Where I ran into problems was image viewing and that some websites are not coded with text browsers in mind. (Links offers a GUI mode, by the way.)
… For different folks. Some people like X, some prefer the console. I tend to prefer the console and uses it as much as possible. This is not to say that I don’t use X, I do, but only when I need to do something that can best be accomplished by using an X-based application.
As I learn more and more about the console and console-based tools, I am finding that I don’t need to fire-up X as much as I used to. For example, instead of starting X to run OpenOffice.org to jam out a newsletter or article, I can use ViM and LaTeX. LaTeX is a very powerful tool, once you get to know it.
Besides, I spend all day sitting in front of a Windows XP machine, and it is kind of relaxing to come home to my Debian Linux computer running just a console session. 🙂
I haven’t used LaTeX before but I think that I will looking into it.
One nice thing about doing most of your work at the command line; your interface is consistant. For example, when I had the NetBSD machine running, I would log in from work via ssh (Putty) and it would be just like being at home.
I used to do a lot of development on client’s servers scattered throughout the U.S. via ssh. I found that I could work quite well even over a dial-up connection, normally running 46k or less. I use FreeBSD, Debian and Slackware at home, and I find that learning to use the console pays off when switching from one *nix to the other.
I started using LaTeX because I am the secretary and newsletter editor of our Amateur Radio club, and got tired of trying to format the newsletter in Word (or OpenOffice.org on the Linux box). LaTeX has a bit of a learning curve, but for simple stuff, you can be up and running in an afternoon.
Most of the responses above are written by people who know nothing about OpenBSD and what software is
available for it.
RTFA and GAFC. most everything mentioned in this threads compiles on OpenBSD, whether it’s in ports or not. if you lack the finesse to compile and/or port, that’s not my problem.
i personally only use OpenBSD for one shell account. in fact, i’m not terribly impressed with it for several reasons. because of the shell, i am quite familiar with CLI and text mode apps on OpenBSD.
(No better than the so-called
author of the quoted article 😉
at least he did something instead of trolling.
Do a little research before wasting bandwidth.
please grow a brain before posting.
I suspect most of the people who think the CLI is a relic of the past have never used it and don’t know the power of it. Worse are those who can’t concieve of a system without one. Most of the websites they read are probably hosted on machines that don’t have a GUI.
I no longer use X for.. three years now. I do everything on console and honestly, I now believe that using a Unix-alike with anything but the console is a sin to daemons
I dont need X anymore. And not having pictures to disturb me makes me go for information directly itself. It’s like the difference between programming using vi and vim : plenty people use vim because it adds colors to syntax. Of course, it looks cool at first but by using colors you don’t learn to read code by itself, just the text. So I came back from vim to vi and now I can, just glancing over C code, read it. I don’t need the colorization anymore to spot something missing.
The only motive for using X would be for web browsing. I can’t think using it for anything else.
If you don’t like syntax-highlightning just turn it off; set syntax=off.
Vim has lots of other advantages over Vi, like multi level undo, multi windows and buffers, command line editing, filename completion, yada-yada-yada.