Having been living in Linux for a while (and to a lesser extent, Mac OS X), I’ve had to learn a bit about the ways of the command line. As someone who’s only other experience with a command line interface was with DOS (with which I was about as adept as one could be), I was skeptical of the value of the command line interface, and reluctant to learn.
That said, I have come to learn that while a command line interface makes it difficult to discover what options are available to you, it does come with some inherent advantages. While I’m certainly not willing to give up my graphical user interface for a text-based console any time soon, I have learned to keep a terminal window handy while using both Gnome/Linux and Mac OS X for a few particularly handy tools.
While these tips are probably absurdly obvious to many people, I suspect there are also plenty of people like myself, web developers in particular, that were raised in Window 3.1-through-95 or Mac OS Classic that never ventured near a command line (except maybe to play Commander Keen now and then).
- WGET
-
Usage: wget [-options] [location]Example: wget http://domain.com/filename.ext
Occasionally I’ll come across a page with an embedded media file, like a QuickTime movie, that I would like to download. QuickTime doesn’t let me save things locally, so I view the HTML source of the page and sniff out the URL of the movie. However, with some file types, depending on how you have your browser and media applications setup, if you paste the direct URL of a media file into the browser, it will load the file in the media player — again, sometimes without any “Save” functionality. If I’m not too annoyed and frustrated to give up trying to download that butt-finger-sniffing-monkey video by now, I would actually have to open a text editor, create a quick HTML file that included a text link with the URL of the item I want to download. Then, I would save the HTML file, open it in a browser, right-click on the link, and choose “Save As…”. Good lord!
With WGET, as long as you have the URL of an item, you can download it. Just type wget followed by the URL.
For bonus points, WGET can resume downloads and supports any kind of downloading option you might need: gzip encoding, SSL, http-authentification.
- WHOIS
-
Usage: whois [domain]Example: whois www.actsofvolition.com
Who is indeed. When you’re trying to figure out a human name or address behind some evil scheming website (or just checking DNS info for a domain), the WHOIS domain database has your info. I used to use, and suspect many people rely on, sites that offer a web-based interface to the WHOIS database (Register.com, Userland, etc.). It usually faster to to open up a terminal and type whois domain.com. No anoying banner ads, and the results are in a simple text format — which you can even save into a file using…
- Piping | Rules
-
The real power of the command line starts to become apparent when you realize the power of piping. Piping is a simple means of passing output from one program as input into another. For example, to list the contents of a directory, I would type ls. I can search the results of that directory by passing the list of contents into a search tool, grep. ls | grep spaceman will return any files in the directory containing “spaceman” in their name. Similarly, I can pass the results of a WHOIS query into my text editor (GEdit) by typeing whois actsofvolition.com | gedit, or save the results directly into a file using the redirection feature “>” (whois actsofvolition.com > aovinfo.txt).
This saved me a load of time last week when I wanted to create a chart displaying the sizes of the various components and packages included in the Fedora Linux distribution. First, I needed a list of the file sizes of packages, so I ran ls *.rpm -l > filesizes.txt on the install CDs. This created a text file with the packages names and their file size which I could then open in a spreadsheet application, sort by size, and graph to my heart’s content.
All of this said, I still love a good icon.
If you’re using Firebird/fox for browsing, you can also do the following:
1) Open up the Page Info
2) Select the Media tab
3) Select the media/graphic/flash file from the list
4) Click the “Save As” button
Very handy.
I always use “curl” for what you do with “wget”. I live on the command line though.
I am sure that you will love grep too. This is one of the tools that every geek misses in the real life.
It’s magic lies in the ability to find any text in any text file so if you want to know per example from where in the source this output comes you can always ask:
grep -rn [word/pattern] /usr/local/src/app/
and it will search all the source tree and tell you every occurence on every row in every file
And this is just the begining 🙂
But what if the command line was anywhere, and the commands were yours.
Looking at some of the illustrations above makes my head hurt! How about just having a command like “find” that finds stuff.
I keep wondering when you will use ActiveWords to name this stuff instead of using some obscure character string!
Buzz
I’m not a user of ActiveWords, but it strikes me that it has one of the same shortcomings that a GUI shell has: no way to connect small programs together to perform larger tasks. Granted that might not be it’s intent, but the outcome is that it’s a commandline-like interface that can accomplish some GUI equivalent tasks with fewer mouse clicks. An ideal shell (text or graphical) should do alot more than just provide a mechanism for launching of programs quickly. Being able to connect programs together in ways not envisioned by their authors is an extremely powerful concept. It allows for existing programs to be used in new ways and encourages the creation of smaller programs that do one thing well rather than large monothlic applications that each live in their own universe and lead to integration nightmares.
My second point is that inuition and obscurity are entirely in the eye of the user. It all depends on the individual’s past experience. If you knew nothing about Office, then typing “excel” into ActiveWords is equally as obscure as typing “grep” into Bash. If you had no past experience with computers, then you would have no idea what outcome to expect from either command.
In Internet Explorer there is a way to search text using search prefixes in the address bar. For example I would type “? flowers” and it would access the site “http://www.google.com/search?q=flowers”.
There is a way to add more search prefixes by editing the registry or by using TweakUI. I’ve, however, used this feature instead to launch javascript functions. For example, instead of using wget on a URL, I’d type “get http://www.yahoo.com/file.wmv” and it would run the script “javascript:void(document.body.innerHTML=”<a href='”+unescape(“%s”)+”‘>click here</a>”);”. Then I would just right-click on the link and save-as to download the file.
Cheers.
If you want to be cool like Steven and use these GNU tools from your Windows command prompt you can download win32 versions here. Although you totally feel like a poseur typing ‘grep’ or ‘find’ (used to find things, just not text in a file) from a C prompt, just tell yourself it’s only until you make the transition.
You can also install cygwin and type ‘grep’ and ‘find’ from a bash prompt instead of a C:> prompt.
Other wonderful tools…
cat /tmp/foo.txt | sed -e ‘s/find/replace/’ > /tmp/new.txt
and you should also try dig, sort, uniq -c, awk, and others.
Some much CLI, so little time.
I don’t like sed as much as I like using find and perl together. xargs allows to use the names of files in the next command :
find -type f -name “*.ext” | xargs perl -pi -e “s/string_to_find/replacement_string/gi;”