Converting to the Unix T shell at Princeton University

This page gives instructions pertaining to

Getting the basic commands
Changing your shell
Changing shell environment settings
Adding aliases
Adding a couple of scripts

plus, some tips that will help you

Watch your quota !!


If you made it here then you must have at least some desire to get to know your unix account a little bit better. The tips that I give below are designed to remove some of the frustrations of working with of the mainframe, and help you manage your account a little better. The main philosophy here is that we should be spending our time in graduate school worrying about the content of our papers and the relevance of our statistical output, and not in the finer aspects of computer science. I have a longer rant about using wysiwyg (what you see is what you get) text programs (like Micro$oft Word), but that is further down the road. For now, all I want to do is help you see unix as a way to get you out of this graduate program sooner.

When I first arrived on Princeton's campus, I knew next to nothing about main-frame computing. A slightly obsessive personality disorder brought me to my current state, but in the end I realized that it only takes a few little tricks to make unix much less intimidating. So, off we go.

telnet to arizona.princeton.edu

Go to this Getting Started with Unix, a guide by Princeton's Data and Statistical Services, and print it out. The commands that you will use most are in the headings Files and Directories. Note that there's only 5 or 6 in each category -- not so much to learn after all. Keep this list handy as you go through the rest of the steps below.

Note that I will follow standard computerese manual conventions below: terminal input and output will be in courier type , the Unix prompt will be given by %, variables like username, filename and filename.sfx will appear in italics.

I also assume that you have a working knowledge of either the Windows NT H drive, or of ftp (file transfer protocol) programs. If you have not already done so, I suggest taking the time to read Julian and Kieran's The Computer and You handout, which is an excellent resource.

Every time that you login to your account, the mainframe looks in two hidden files for environment variables. Think of these as your account preferences. Your account came with standard versions of .cshrc (or "C shell resources") and .login files. They're referred to as hidden files because their names begin with a period. The standard ls command won't show them in your directory, but ls -a (or "list all") will. Unix will read the .cshrc before the .login file, which (again) you won't really need to remember after you read this section.

If you are running the T shell, unix will first look for a .tcshrc file, and will read the .cshrc file only if it can't find it. Let's leave the .cshrc file alone, and create a new .tcshrc file. This will help define some of your new T shell toys.

Save this file in the top level of your H drive account (assuming, as apparently I must these days, that you're working on a Windows NT machine and not a Macintosh), or somewhere on the hard drive where you can find it later and ftp it.

Now go into the top level of your unix account and type

% ls

to make sure it got there. If it did, change its name (or, in unix parlance, "move" it) with the following:

% mv tcshrc.txt .tcshrc

Now we're going to take something out of your .login file. Using your favorite text editor, go in there and remove the line that begins set prompt . The same line appears in the .tcshrc / .cshrc files, and it doesn't belong in .login . Trust me.

Once you've made this change, type

% source .tcshrc

You only have to do this for this session -- normally the file will be read when you login. You should have received the following error message

/u/username/.aliases: No such file or directory.

but we'll take care of that in a moment.

You should see an immediate change in your prompt, hopefully one that you'll agree is for the better. Change to one of your lower directories, and you'll see how useful it can be.

Besides having access to your previous commands, you now have the amazing abilities to

• mispell a command and be prompted for the correct spelling
• use the tab key to complete the spelling of directory and file names
• use the backspace key without getting garbage on the screen
• and so much more!

Well, maybe not a lot more, but you're now a lot better off than you were 15 minutes ago. For more information on how the T shell differs from the C shell, you can take a look at the first two pages of its manual. (You can also consult this manual anytime by typing

% man tcsh

but note that the manual is fifty-some pages long.) It's possible that you won't have all of this available to you right away. At this point, you might as well logout and login again. That should set everything straight.

DO NOT DELETE your old .cshrc file !!! If you have ever made any changes to this file, or if any or your geek friends have helped you in the past, you may now need transfer some of these commands to your new .tschrc file. Let me know, and I'll help you go through the .cshrc file to see what's in there that you need.

You can further increase your unix pleasure to pain ratio if you put this file into the top directory of your account. Now change its name with

% mv aliases.txt .aliases

An alias is simply a command definition. If you now type

% source .aliases

(or)

% sou[tab] .ali[tab]

(taking advantage of your new skills) you will read them into your account. Again, you don't have to do this for every session.

Now that you have your aliases, you might want to know more about what they are and what they do. Aliases define commands, and can do this in one of two ways. First, they can create a new command name that unix wouldn't normally recognize. For example, in the aliases.txt file, I create a new command called " m " which invokes the e-mail program pine:

alias m /usr/princeton/bin/pine.396 -l -z

(The -l flag tells pine to automatically expand the folder collection, and the -z flag allows me to "suspend" the program, which I'll explain some other time.) Note that this is the old version of pine, which I prefer. Princeton changed to the newer version a year ago or so. If you like the new version, you could use this alias instead:

alias m /usr/princeton/bin/pine -l -z

You can change the aliases anytime you want by editing the .aliases file. To do so, you could use the " p " alias, which now stands for " pico ", and type

% p .ali[tab]

As you might have figured out, one of the goals here is to type less, but get more done.

The second way you can use an alias is to change the default behaviour of the command, keeping its original name. For example, when I use the program dvips , which transforms .dvi files into .ps files, I want to always use the flags -o and -N0 . Since I can barely remember the exact syntax of these flags, and can't think of any instance when I wouldn't want to use them, I just redefined the command with

alias dvips /usr/princeton/bin/dvips -o -N0

If I ever wanted the original, unaliased command, I could type

% unalias dvips

to restore it to default settings. Take a look at the aliases.txt once again to see what's in there. If you see an alias for a command that you don't know, don't worry about it. You can delete it if you want, but I would suggest that you just leave it unless you have a good reason to dump it. (One good reason might be that you want to use the alias "j" for something else, for example.) The printer commands in the last section are specific not only to Princeton, but (in general) also to the Department of Sociology, and will probably be either the most useful or the most worthless commands.

---

(Note to those outside of Princeton:

these aliases are based on specific locations of the files within Princeton's unix system.
To find them on your own system, type the following commands

% unalias command
% which command

and replace as needed in the .aliases file.)

---

Now there's just one last thing to do, which will help you later with file and directory management. I want you to add two little scripts to your account, which will allow you to monitor the size of your directories. (Unlike Windows and Macintosh Operating Systems, you won't be able to automatically see how large your directory is when you ls your files.)

First, we make a new hidden directory in the top level of your account (also known as your "root"), and name it .bin . With your new prompt, you should know that you're in the top level if the only thing you see in your path is " ~ ". If you're not there, then get there with

% cd

Now make the new directory called ".bin" with

% mkdir .bin

Now save / ftp the following two files, filesize.txt and dirsize.txt in this new directory. Now change their names (sorry -- you have to do this because of Windows, not Unix ! ) with the following

% cd ~/.bin
% mv filesize.txt filesize
% mv dirsize.txt dirsize
% cd

Your .aliases file already contains aliases to these two scripts. If you were to now type

% dirsize

you would create a few named dirsize.txt in your account. This contains the size, in kilobytes, of every directory you have. (And the filesize script will also tell you the size of every file within those folders.) Anything that is over 1000 is over 1 megabyte. When you run into quota problems, and someday you will, these two scripts will help you figure out what to remove or compress. Note that these are simply snapshot text files -- they won't update themselves as you remove or add things. Running the filesize and dirsize commands again will overwrite the text files, and you can remove them any time that you want.

When your account was set up by CIT, you were given a "quota" of 10 megabytes. This is enough for your to save some messages in your e-mail account, and to create a few Word documents, but once you start doing some data analysis you'll find that you can quickly exceed your alloted memory.

If you exceed your quota, you have seven days to remove some files. If you do not do this, you will no longer receive any e-mail, and CIT will remove some files for you. This is a very bad situation to be in.

If you exceed your quota, you will receive a message about this every time that you log in. Unfortunately, most people ignore this message. In a panic, they often then begin removing files in a mad dash once they figure out that they only have an hour until their account is about to be altered. Here's some steps you can take to avoid these perils:

1. Put the following line at the bottom of your .login file:
   
   quota -v
   
   This will show you your quota everytime that you log in. You can also type this command anytime during your session to see what you have.
   
   
2. Run the dirsize command (see above) at least once a week to monitor the size of your directories.
   
   
3. Delete messages in your e-mail account that include unnecessary attachments. If you want to hold onto them, then you should consider saving the attachment as a separate file in your Unix account, and then " compress " them (see below). You should especially look in any " sent-mail " folders that you might have. If you have sent someone an attachment of a file that already lives in your Unix account, then you are now saving a duplicate copy of it in your mail directories -- delete those messages, too.
   
   
4. "compress " any large files, especially data sets, that you are not currently using. You can do this with
   
   % compress filename.sfx
   
   This will replace filename.sfx with filename.sfx.Z . You can always " uncompress " this file when you need to use it again, with
   
   % uncompress filename.sfx.Z
   
   If you have a lot of files in a particular directory that you are not using, then you can type
   
   % compress *
   
   The asterisk ( * ) is a "wildcard" character -- it refers to any string that might appear in a filename for the particular directory you're working in. You could, for example, uncompress all .Z files with
   
   % uncompress *.Z
   
   
5. If you have done all of the above, but you are still over (or near) your quota, then you should get it increased. The department gets charged for this, so don't go overboard, but you should certainly get enough to do your work. Cindy will fill out the paperwork for you, and specify a dollar amount. (She usually writes in $1,000. This is how much the department is willing to pay for your Unix needs. And you thought they didn't love you.) Take this form to CIT, at 87 Prospect, and tell them how many megabytes you want for your quota. Again, you don't have to jump to 200 megabytes -- with Cindy's form on file, you can always increase your quota later when your needs increase. And they will.
   
   
6. If you ignored items 1 through 5, and files in your account were indiscriminately deleted by our friends in CIT, do not panic. First, you should clean up what does remain, by deleting and compressing. Then you should send an e-mail to restores@princeton.edu and ask that the files in question be brought back to life. CIT backs up every file on every Unix account every night, and saves them on tape for one year. If you should happen to delete an important file or directory from your account, you can specify the filename / directory name and the day that it last appeared in the version you want, and within a day it will reappear.