Guide to Univac for Non-Programmers

If you are a beginner at using the Statistical Package for the Social Sciences (SPSSX) or the Document Processing System (DPS) on the Univac system at Memphis State University, you have problems. SPSSX and DPS are problems in themselves, but unfortunately this guide will tell you almost nothing about them. You'll have to get someone to teach you those packages or learn them yourself out of the manuals. Univac is another problem, and this guide is an attempt to offer real help in dealing with the Univac.

When I was just beginning to learn the Univac system, I found it necessary to refer constantly to my handwritten notes as models for procedures. I could never remember the steps in creating a new file in CTS, for example, so I made a sheet listing them. I couldn't even remember if the logging-off command was @FIN or @@FIN, so I had the logging-off procedure on a separate sheet. And so forth — for every procedure, an "idiot sheet." After a time it occurred to me that others might benefit from my notes, so I typed them up, and this guide is the result.

It is intended to fill the gap between the Computer Services handout entitled "Selected Commands for Handling Disk Files on a Time-Sharing Terminal" and the "MSUCS Computing Handbook for the Univac 1100/62." Both are excellent but deficient in their own ways. The handout covers the absolute essentials but does not have much detail; the handbook has enormous detail but is incomprehensible to the beginner. The present guide is intended to be "just right" — telling you what a beginner really needs to know, and in understandable detail. You won't know everything about the system, but if you know what's in these pages, you'll be able to use Univac with some confidence.

The first edition of this guide, issued in August 1982, was written at a time when SPSS was in version 8 and users commonly had Hazeltine asynchronous terminals. I have revised it to reflect the arrival of SPSSX and the Sperry synchronous terminals, as well as a few changes in the operating system commands.

A. Survivor

Lake Woebegone, Mississippi

March 1985

2. Getting Started

2.1. Getting a Computer Account

To use the computer, you must first have a computer account. If you are in a class that requires the use of the computer, you will use an instructional account created for that class. If you are a staff member, you will probably use an account assigned to your office or activity. If you are a graduate student or faculty member and wish to use the computer with a research project, you may apply for an account at User Services, Administration Building 134.

2.2. Key to Symbols and Conventions Used in This Text

The user's responses (that is, what you would key in on the keyboard) are in BOLDFACE ITALIC CAPITAL LETTERS. (You may enter your responses in either lower-case or capital letters.)

Instructions and comments are in lower-case letters.

Names of terminal keys will be enclosed in < >. For example, <XMIT> indicates the key marked XMIT. (The keys on different terminals used to send lines to Univac may be marked <CR>, <RETURN>, <NEW LINE>, or <XMIT>; for convenience, <XMIT> will be used in this text.)

2.3. Special Warning to Novices

If you are in the habit of using the alphabetic character upper-case <I> or lower-case <L> to represent the numeral "one" and the alphabetic upper-case <O> to represent the numeral "zero," change your ways immediately. Use the alphabetic keys to represent only letters of the alphabet, and use the numeric keys (on the top row of keys or on the "number pad") to represent only numerals. Do not interchange letters and numerals.

2.4. The Cathode-Ray Tube (CRT) Terminal

For the most part, the instructions that follow assume that you will be using a cathode-ray tube (CRT) terminal that is

"hard-wired," or connected directly to the computer (perhaps through a multiplexer). If your terminal is not "hard-wired," you will use a telephone to call up the computer and a modem (modulator-demodulator unit) to make a usable connection ("interface," in the jargon) with the computer. Instructions for doing this will probably be posted near your CRT. Once you're on the computer, procedures are identical.

Before you can enter anything, the terminal must of course be turned on. Sperry terminals are always "on," though you may have to press a key to get the screen to light up. The switch on most other CRTs is located on the lower right rear, difficult or impossible to be seen, but easily felt.

You will find that the CRT is similar to a typewriter, except that the typed characters will appear on the CRT screen instead of on paper. One peculiar feature is the cursor, a marker which may be rectangular or wedge-shaped, which will move across the screen to mark your position as you type. And instead of each character being read by the computer individually as it is typed, a whole line of characters is "fed" to the computer as a unit when you press the <XMIT> key.

Except for a few unusual commands such as DEMAND (used for logging-on) and @@X (used to terminate execution), there must be a "prompt" character showing at the left of the screen before you can type and enter a line. Some commands may take a fairly long time to be executed (and the whole system sometimes operates slowly if there are many users), so you may have to wait a while for the prompt to appear. If you attempt to enter anything before the prompt appears, the computer will usually advise you: *WAIT-LAST INPUT IGNORED* If you are in Executive Control Language (ECL) mode, the prompt character will be >; if you are in Conversational Time Sharing (CTS) mode, it will be ->. DPS has three modes and three prompt characters. For INPUT it will be > (note that this is identical to the ECL prompt; sorry about that); for EDIT it will be *>; and for COMMAND it will be ?>. Each mode will ignore commands appropriate to another mode (though there are some few commands that will work in more than one mode). The instructions on the following pages are coded with the kind of prompt character that should appear before you use a particular command. Besides the prompt character (which you do not type), notice the spacing between characters. Univac is very demanding about this. If the model shows a space between characters, leave a space; if the model doesn't show a space, don't use one. Observe carefully, because the computer will not be able to guess what you meant to enter — it can only know what you actually did enter.

2.5. Logging-On

Type:

DEMAND <XMIT>

or while holding down <FCTN>, press the numeral <1>. If you get a message saying: TIP SCHEDULING ERROR 003 — RESEND LAST INPUT, press <SOE>. The "start of entry" symbol will appear. Now type DEMAND <XMIT> again. When the computer responds, it will instruct you:

ENTER USERID/PASSWORD CLEARANCE LEVEL

You should enter your user identification and your password (although requested, the clearance level is not required when you are using a CRT), as, for example:

>ISYSTEM/CRASH <XMIT>(Remember: you don't type the prompt character.)

The user ID is assigned by Computer Services and appears on any printouts, so it is pretty much a matter of public record. The password is supposed to be secret and allows only those who know it to use the account. It is initially assigned by Computer Services, but the user can change the password any time he desires if there is reason to believe that any unauthorized person has come to know the old password. (If you are working on an instructional account you will not be able to change the password. If you have an individual account, Computer Services will tell you how to change the password.)

The computer will respond with a block of information, including the run number, the current time, etc. If you are using a Sperry synchronous terminal such as the UTS-20, it is not necessary to enable the backspace character. But if you are using a dial-up line, you should type:

>@@TTY C,<BACKSPACE> <XMIT>

(Remember: Anything between the symbols < and > is a key on the keyboard. Press the <BACKSPACE> key — do not type the symbols or the word.)

The computer will respond with:

-@@COMPLETE

This sets the <BACKSPACE> key as the "delete character" for correcting errors as they are typed — you simply backspace to the error and retype it and any succeeding characters. (If you are a perfect typist, you do not need to set a delete character.) The <BACKSPACE> key is not automatically a delete character, as some persons mistakenly assume. If you do not set it by the procedure outlined above, it will appear to make corrections on the CRT screen but it will actually introduce even more (invisible) errors. If you forget to set the delete character immediately upon logging-on, you may do it at any time, in either ECL, CTS,or DPS.

Sometimes when you attempt to set a delete character, the computer will respond:

* @@ FORMAT ERROR *

Possibly you didn't type the command correctly. Examine the line closely, then re-enter it carefully. If you get the error message again, don't worry — you haven't made a mistake; someone else has. An earlier user probably did not log-off correctly and the delete character is already set, left over from that earlier user's run. If you still aren't convinced that the delete character is properly set, deliberately type an error (such as @CTX for @CTS) and correct it — if the corrected command works, obviously the delete character is set.

3. Files

3.1. Kinds of Files

Files are units of material stored in the computer. Univac uses three types of files, but only two of them are immediately relevant to your needs.

A data file (sometimes called SDF) has no subdivisions. It is a single unit and must be used as an entire unit if it is to be used at all.

A program file may have any number of subdivisions called elements. A program file ordinarily cannot be used as an entire unit; only one element at a time may be used. (Elements may themselves be of several different forms. SPSSX uses symbolic elements, DPS uses omnibus elements, and other processors may use relocatable or absolute elements.)

The terms "data" and "program" are technical and somewhat misleading because a data file may actually contain a program (instructions for the computer) and an element of a program file may actually contain data (material to be processed by the computer). The only meaningful difference for non-programmers is that a program file has subdivisions while a data file does not.

Until you gain some experience in handling files, you might want to use only data files, because initially they are easier to work with and understand than program files.

3.2. Names of Files

All files and all elements of files must have names. Both may have very complicated structures, but your needs can be met with very simple structures.

A file name may be 1 to 12 characters in length and may consist of any combination of the letters <A> to <Z>, the digits <0> to <9>, and the special characters <-> (hyphen or minus) and <$> (dollar sign). Blanks or spaces within the name are not permitted. Although it is not strictly speaking a part of the file name, a period must be used after the file name as a way of letting the computer know that it is dealing with a file and not with an element (normally the computer will assume that any name not ended with a period is an element).

An element name has essentially the same rules except, of course, that it must not be ended with a period. (It is probably wise to use only letters and digits in your names. The dollar sign can be especially troublesome, because several processors and programs use it to identify internally-generated temporary files and will refuse to accept your proposed file name if you happen to choose a name that is "reserved.")

According to these rules:

BASIC. is a file

ROOT.NR2 is the element NR2 within the program file ROOT.

NR57 is an element within an unidentified program file

Your file names must be unique. Univac will not allow you to use two or more files with identical names because it would not be able to tell which one you meant when you called for it. Within a given program file there may be several elements with identical names if they are of different forms, because Univac invisibly "flags" each form as being absolute, omnibus, relocatable, or symbolic, and can therefore distinguish between the forms. Identical names may not be assigned to two elements of the same form within the same program file. (Elements of different program files may also have identical names because the complete identification of the element includes the unique file name as well. For example, FILE1. and FILE2. may each have an element named CF1 because the computer will be able to distinguish between FILE1.CF1 and FILE2.CF1)

3.3. Cataloging Files

All files must be cataloged before they can be stored in the computer. The most direct way to catalog a file is through the ECL command:

>@CAT FILENAME. <XMIT>

The computer will respond:

READY

Many Univac commands may have "options" to more fully define the command. The option, or options, will immediately follow the command, being separated only by a comma, not a space. In the absence of any options being given, there is usually a "default," which is a definition automatically given to the command. In the previous example, no option was given, and the default was to catalog the file as a private file, available only to your account. If you wish to allow access by users on other accounts, you may create a public file by using the P option:

>@CAT,P FILENAME. <XMIT>

The computer will respond as above.

It is a good idea to redate your file as soon as you have cataloged it. Univac will automatically delete a file after 7 days if it has not been redated within that time. Type:

>@REDATE FILENAME.,30 <XMIT>

(If you have an individual account, you may redate for up to 150 days.)

In addition to this method of cataloging in ECL, you may catalog files in CTS by three different means:

(1) Type:

->CREATE FILENAME. <XMIT>

The computer will ask you three questions which it will propose in succession. You may enter answers, or you may ignore any of them by keying <XMIT>. If you simply key <XMIT>, defaults are put into effect; for your purposes, the defaults are quite satisfactory.

IS THIS FILE TEMP, PUBLIC, OR PRIVATE? (private, if <XMIT>)

READ AND WRITE KEYS: (none, if <XMIT>)

DEVICE CHARACTERISTICS: (disk storage, if <XMIT>)

The computer will respond, as, for example:

*CRE,U FILENAME.

which indicates a private file created and cataloged as FILENAME.

It has no read or write keys and will have characteristics suitable for mass storage on disk.

(2) If you specify a filename in the ECL command to enter CTS, as, for example:

>@CTS,IN FILENAME. <XMIT>

the file will automatically be cataloged, though the computer will make no special comment to that effect.

(3) If you do not use the ->CREATE command or if you do not specify a filename with the >@CTS,IN command, when you enter the ->SAVE command, the computer will first ask you for a program name and then will ask you the three questions mentioned above and will catalog the file after you answer them.

4. Using Conversational Time Sharing (CTS)

4.1. How CTS Works

When you log-on, Univac is in the "executive" mode, expecting Executive Control Language (ECL) commands. In effect the computer is saying "OK — what do you want me to do?" As already noted, one thing you may do in ECL is to catalog files. You may also execute or run command files in ECL. But first you must create them.

You will create your files in the Conversational Time Sharing (CTS) mode of operation, so-called because in a sense you do "converse" with the computer, giving and altering commands and responding to its prompts and error messages, and you are actually sharing the computer with other users on different terminals.

CTS creates a "work area" that may be compared to a scratch pad in ordinary writing. Just as the scratch pad is not the finished document, so the CTS work area is not a computer file, though it may be converted into one or it may be a copy of one. After you have made several drafts of a paper, you may be ready to label the latest draft as the "official" version of the document. Similarly, after you have used CTS to correct and rearrange, you may designate the latest version of the work area to be entered into computer storage as the "official" file (through the ->SAVE command). Again, if you have a previously-completed document, you might make a photocopy of it to work on revisions without actually marking up the "official" document. Similarly, DPS will make a copy of a file that may be altered in any way you desire without affecting the "official" file. You might decide to deliberately replace the old version with the new (in CTS, you would use a ->REPLACE command), or you might decide after all to retain the old version and discard the copy (in CTS, you need do nothing in particular to accomplish this).

Entry to CTS is usually made by the command:

>@CTS,IN <XMIT> or >@CTS,IN FILENAME. <XMIT>

The computer will respond with a few lines about CTS and then give the CTS prompt: ->

The options I and N instruct the computer to initialize the work area (erase any previous material and make the area completely empty to start with), and to skip the news items which may be inserted on the system by Computer Services.

The second form of entry into CTS has some advantages if you intend to call up in succession several elements from the same program file. It can also save some effort if you plan to create new files in CTS without having already cataloged them by the >@CAT command.

See the sections that follow for the procedures to be used in CTS for creating a new file, calling up an old file, changing and deleting lines in a file, and numbering and resequencing lines.

Many command files, including SPSSX, can be executed on the CRT screen in CTS as well as in ECL. But if you need to get out of CTS for any reason, you may exit by typing:

->XCTS <XMIT>or ->XCT <XMIT>

If you wish to return to CTS and find the work area exactly as it was when you left it earlier, so that you can resume working with it, you should return to CTS with the command:

->@CTS <XMIT>

If you wish to return to a completely empty work area, cleared of any previous materials, you may initialize the work area by using the I option, as in:

>@CTS,IN <XMIT>

If you are already in CTS, you may initialize the work area at any time by typing:

->NEW FILENAME. <XMIT>

4.2. Creating a New Data File (SDF) in CTS

>@CTS,IN FILENAME. <XMIT> >@CTS,IN <XMIT>

->NEW FILENAME. <XMIT> ->NEW FILENAME. <XMIT>

Then type:

->N <XMIT>

This will begin automatic numbering of lines, beginning with 100 and incrementing by 10, as: 100, 110, 120, 130, etc. You may specify a different sequence of numbering (see the section "Numbering and Resequencing Lines in CTS"). Note that a new kind of prompt character will appear as long as you remain in automatic line numbering. Now type the lines of your file:

100>TITLE 'WHATEVER YOU WANT TO HAVE PRINTED' <XMIT>

110>FILE HANDLE MYNAME / NAME = 'UNIVACFILE.' <XMIT>

120>GET FILE = MYNAME <XMIT>

yyy>FINISH <XMIT>

zzz>*MANUAL <XMIT> or zzz>*MAN <XMIT>

This stops the automatic line numbering and returns to manual control in CTS. When you wish to get out of CTS, remember to enter:

->SAVE <XMIT>

If you have not previously cataloged the filename or did not use the filename in the >@CTS,IN line, you will have to answer here the three questions mentioned in the section "Cataloging Files."

This puts your file into storage, from which it can be recalled any time later. If you forget to ->SAVE the file as soon as you have created it, there is a strong likelihood that you will "lose" it and have to retype it. (There is a way to recover it under favorable conditions, but the beginner is not likely to remember the right things to do in order to accomplish it.)

4.3. Creating a New Element of a Program File in CTS

>@CTS,IN FILENAME. <XMIT> >@CTS,IN <XMIT>

->NEW ELEMENTNAME <XMIT> ->NEW FILENAME.ELEMENTNAME <XMIT>

From this point onward to the ->SAVE command, the procedure is identical to that of "Creating a New Data File (SDF) in CTS":

->N <XMIT>

(contents of file)

zzz>*MANUAL <XMIT> or zzz>*MAN <XMIT>

->SAVE <XMIT>

If you have not previously cataloged the filename or did not use the filename in the >@CTS,IN line, the computer will reject your attempt to save the file by reporting that the filename is not cataloged. Don't despair. You can still save the element. Type:

->CREATE FILENAME. <XMIT>

and then answer the three questions mentioned in the section on "Cataloging Files." When the computer reports:

*CRE,U FILENAME.

you then type:

->SAVE <XMIT>

4.4. Calling up an Old File in CTS

The file must previously have been saved, or there will be nothing to call up, naturally.

A data file may be called up by:

>@CTS,IN <XMIT>

->OLD FILENAME. <XMIT>

An element in a program file may be called up by:

>@CTS,IN <XMIT> >@CTS,IN FILENAME. <XMIT>

->OLD FILENAME.ELEMENTNAME <XMIT> ->OLD ELEMENTNAME <XMIT>

After calling up the file you will probably want to list or print all or a part of it on the screen so you can examine it:

->PRINT ALL <XMIT> or ->P A <XMIT>

This will list the entire file, and if it is not more than 23 lines, all of it will be visible on the screen at the same time. If you have a long file, you will probably want to print it in sections, as:

->P 100,300 <XMIT>

This prints lines 100 through 300. (Note the use of the comma — do not use a hyphen here no matter how logical it seems.) Then for another screen full of lines:

->P 310,500 <XMIT>etc.

4.5. Changing and Deleting Lines in CTS

In CTS, any changes you make are made immediately in the work area, which you should remember is only a copy of the file. To change the file itself, which is still in storage, you must enter ->REPLACE or ->REP after a change or series of changes. (Failure to enter ->REPLACE ranks along with failure to enter ->SAVE as perhaps the most common and most frustrating errors made by beginners.)

There are numerous possibilities for changing and deleting. Here are a few common examples. (Be sure you are not in automatic line numbering when you attempt to make changes — if you are, not only will these commands have no effect on the lines referenced, but the commands will themselves be entered as lines in your file. Note that CHANGE may be abbreviated C; DELETE may be abbreviated D.)

->C /I/IS/ 110 <XMIT> Change I to IS in line 110; this will

change only the first I found in that line

->C /T/TT/ 150 R <XMIT>Change T to TT in line 150; this will

change all occurrences of T in that

line to TT (R means Repeat)

->C /,/ / A <XMIT> Change the comma to a blank space

everywhere it occurs in every line

(A means All lines)

->D 100 <XMIT> Delete line 100

->D 100,400 <XMIT> Delete lines 100 through 400

->D 100+ <XMIT> Delete line 100 and all lines

following 100

->D 100- <XMIT> Delete line 100 and all lines

preceding 100

->D A <XMIT> Delete all lines

->MOVE 140 105 <XMIT> Move text from line 140 and place it on

new line 105; eliminate line 140

(In the examples above, the slash mark </> was used as a delimiter to set off the changes. Any character, for example, a single quotation mark <'>, may be used as a delimiter provided it is not in the string of characters to be changed.)

One way to change a line is to retype it completely, including the line number.

New lines may be inserted between existing lines, provided the existing lines are not successive integers, such as 10, 11, 12, because Univac does not allow decimalized lines such as 10.5 or 11.5. Type the new line number and the text for that line. The computer will automatically put any corrected or new lines in their proper sequence when the file is listed or run.

The procedures listed above will work with any terminal. Sperry synchronous terminals have special editing keys which allow you to do full-screen editing, using the cursor-movement keys, INSERT IN LINE, and DELETE IN LINE. It is difficult to describe in writing how to use them — ask the attendant in the terminal area, or go to the Consulting Desk in Administration 134 for help.

4.6. Numbering and Resequencing Lines in CTS

When you want to enter new lines, you may enter line numbers yourself. However, it is much easier to let the CTS system do the numbering for you. Here are some examples of what may be done:

->N <XMIT> Begin automatic line numbering; if no specifications are given, the default is

that numbering will begin at 100 and

increase in increments of 10:

100, 110, 120, 130, etc.

->N 100,5 <XMIT> Begin automatic line numbering at 100

and increase in increments of 5:

100, 105, 110, 115, etc.

->N 1,1 <XMIT> Begin automatic line numbering at 1 and

increase in increments of 1:

1, 2, 3, 4, 5, etc.

->*MANUAL <XMIT> Stop automatic line numbering; resume

manual control in CTS

->*MAN <XMIT>

->N ! <XMIT> This is used after the ->*MANUAL command

has been used. If you later wish to add

more lines to the file, this command

will resume automatic line numbering

from the highest line number entered

earlier and will increase in the same

increment as specified earlier

->RESEQUENCE <XMIT> Renumber all lines with a constant

interval between the new line numbers.

or This is useful when you have placed

new lines between the original lines

->RES <XMIT> and now wish to have a uniform series

of line numbers. For example, lines 100, 105, 110, 111, 112, 113, 120, 135 will become 100, 110, 120, 130, 140, 150, 160, 170 if you enter ->RES <XMIT>. This assumes the default values for the line numbers. As with the ->N command, you may specify what the resequencing will be. For example, if you wish to start with line number 10 and increase in increments of 10, type: ->RES 10,10 <XMIT>.

5. Some Utility Procedures

5.1. Printing Hard Copy of a File or an Element

To list the contents of a data file, in ECL:

>@BK1 <XMIT>

>@DATA,L FILENAME. <XMIT>

>@END <XMIT>

>@BK2 [here you may specify printer desired] <XMIT>

If you leave a blank space on the >@BK2, the printer will be the main printer, and copy will go to the Input/Output Room, Administration Building 112. You may direct copy to one of the following remote batch terminal printers: ADMIN, BUSAD, MATH, or ENGR.

To list the contents of a data file, or the contents of all the elements within a program file, in ECL:

>@BK1 <XMIT>

>@FLIST FILENAME. <XMIT>

>@BK2 [specify printer] <XMIT>

To list the contents of a particular element within a program file, in ECL:

>@BK1 <XMIT>

>@ELT,L FILENAME.ELEMENTNAME <XMIT>

>@BK2 [specify printer] <XMIT>

To list the contents of a data file or of a particular element within a program file, in CTS:

->OLD FILENAME. <XMIT> or ->OLD FILENAME.ELEMENTNAME <XMIT>

->SITE <XMIT>

The computer will then ask three questions in succession:

HDG?> Enter whatever you wish printed at the top of the copy.

If blank, simply key <XMIT>.

MSG?> Enter the message you wish printed at the bottom of the copy.

If blank, simply key <XMIT>.

SITE?> Identify the printer desired, as above; then <XMIT>

5.2. Listing of Files and Elements by Name

Ideally you should remember or make a record of any elements that you add to a program file. If you cannot remember their names, the following procedures will list the elements for you:

(1) In CTS:

>@CTS,IN FILENAME. <XMIT>

->LIST SAVED <XMIT> or ->LIS S <XMIT>

This will list the elements by name, with an indication of their type (you will be working in SPSSX with symbolic elements, which will be indicated as ELT, and in DPS with omnibus elements, which will be indicated by O-SYM) and the compiler (if any) that was used.

(2) In ECL:

>@PRT,T FILENAME. <XMIT>

This will provide essentially the same information as above, but the table that is generated will have a different appearance (in this case, omnibus elements will be indicated by OMN). For more information:

>@PRT,TL FILENAME. <XMIT>

This will give more detailed information about the elements, including when they were last referenced (symbolic elements will be indicated by ELT SYMB-Q and omnibus elements by OMNIBUS). If you have made lots of changes in your programs, this listing can be fairly lengthy and you may be asked to pack the file to delete superseded versions of the elements (see the section entitled "Deleting a File or an Element.")

If you cannot remember even the names of your program files or the names of your data files, there are ways of getting the computer to list them for you. The following procedure will give, among other things, information about the size of the files, which will be useful in keeping within the limits of storage allowed to each computer account.

>@L,B <XMIT> or >@L,C <XMIT>

The first command will give a quick short listing. The second command may take a fairly long time, but it will give more complete information.

(CTS allows many ECL commands to be invoked by a special procedure: PXQ, followed by the ECL command without the <@> symbol. The above calls could be invoked in CTS as:

->PXQ PRT,T FILENAME. <XMIT>

->PXQ PRT,TL FILENAME. <XMIT>

->PXQ L,B <XMIT>

->PXQ L,C <XMIT>

5.3. Copying a File to Another File

(1) You may wish to make a duplicate of a data file or a complete program file with all its elements. If the new filename is already cataloged, you may copy the old file into the new file in ECL by:

>@COPY OLDFILENAME.,NEWFILENAME. <XMIT>

or in CTS by:

->COPY OLDFILENAME.,NEWFILENAME. <XMIT>

(2) You may make a duplicate of a data file in CTS through the ->SAVE command:

>@CTS,IN <XMIT>

->OLD OLDFILENAME. <XMIT>

->SAVE NEWFILENAME. <XMIT>

If the new filename has not already been cataloged, you will have to answer the three questions generated by the computer as outlined in the section "Cataloging Files."

NOTE: All the following examples assume that the NEWFILENAME. has been cataloged before the attempt to copy is made.

(3) You may copy a data file into a program file as one of its elements by:

>@COPY,I OLDFILENAME.,NEWFILENAME.ELEMENTNAME <XMIT>

(4) You may copy an element of a program file into a new data file by:

>@COPY,I OLDFILENAME.ELEMENTNAME,NEWFILENAME. <XMIT>

(5) You may copy an element of a program file into a new program file as one of its elements by:

>@COPY,S OLDFILENAME.OLDELEMENTNAME,NEWFILENAME.NEWELEMENTNAME <XMIT>

(This assumes that the element you wish to copy is symbolic. If you wish to copy an omnibus element, the option would be O; if absolute, A; if relocatable, R.)

(Incidentally, the copying of a file to a new file does not change the old file in any way. It is still there, exactly as it was before you made a copy of it.)

5.4. Deleting a File or an Element

To delete a file:

>@DELETE FILENAME. <XMIT>

>@CTS,IN <XMIT>

->UNSAVE FILENAME. <XMIT>

To delete an element:

>@DELETE,S FILENAME.ELEMENTNAME <XMIT>

>@CTS,IN <XMIT>

->UNSAVE FILENAME.ELEMENTNAME <XMIT>

>@CTS,IN FILENAME. <XMIT>

->UNSAVE ELEMENTNAME <XMIT>

The S option in the ECL command assumes that the element you wish to delete is symbolic. If you wish to delete an omnibus element, the option would be O; if absolute, A; if relocatable, R.

(Actually the deleting or unsaving of an element only marks the element for deletion. The actual deletion and release of storage space is made by the additional ECL command: >@PACK FILENAME. <XMIT>) or CTS command: ->PACK FILENAME. <XMIT>.)

(CAUTION: If you have different forms of elements (absolute, omnibus, relocatable, symbolic) with the same name, the ->UNSAVE command will remove all of them, which may be more than you intended.)

5.5. Interrupting Output and Terminating Execution

When you are executing a command file on the CRT screen, the lines of output may print too fast to be read or they may "run off" the top of the screen before you can read them, and you may wish to "freeze" the lines so that you can examine them more closely or at more leisure. Press the <MSG WAIT> key. The computer will stop the output temporarily and note:

*OUTPUT INTERRUPT*

Note that there is no display of a prompt character following this output. To resume printing, press <XMIT>.

If for any reason you wish to terminate the output before it has run fully (possibly because you have discovered an error or have realized the output will run much longer than anticipated), after pressing the <MSG WAIT> key and interrupting the output, type:

@@X <XMIT>

The computer will halt and return a prompt character. This procedure should halt the execution of any SPSSX command file or DPS script. If you find, however, that SPSSX or DPS resumes after you have typed some other command, you may have to type (after getting output interrupted again by pressing the <MSG WAIT> key):

@@X TIO <XMIT>

(NOTE CAREFULLY: <I> and <O> are alphabetic characters, not numerals.) The computer will halt and respond with:

*EXECUTION TERMINATED*

6. For SPSSX Users

6.1. Checking for Errors on SPSSX Runs

To avoid getting error messages on an SPSSX run, there are three procedures that you can use to detect and correct errors before actually executing the command file.

(1) If you plan to run the SPSSX command file on-line (on the CRT screen), you may check for SPSSX syntax errors by doing an EDIT run in SPSSX first:

>@SPSSX <XMIT>

The computer will list a few lines generated by SPSSX. Then type:

>EDIT <XMIT> (NOTE CAREFULLY: do not use <@> with EDIT)

>@ADD FILENAME. <XMIT> or >@ADD FILENAME.ELEMENTNAME <XMIT>

If you are in CTS and already have the command file in the work area (perhaps you have just finished creating and saving it), you may call the EDIT run simply by typing:

->CALL SPEDITX <XMIT>

If the SPSSX EDIT run reports errors, you must go back into CTS, call up the file, correct the errors, and try again. If no errors are detected, you may execute the program as outlined in the section "Running an SPSSX Command File."

(2) If you plan to send the output from the command file to a printer to get a "hard copy" printout, you may check for SPSSX syntax errors and run-time errors before actually requesting the copy, through the built-in system editor:

>@BK1 <XMIT>

>@SPSSX <XMIT>

>@ADD FILENAME. <XMIT> or >@ADD FILENAME.ELEMENTNAME <XMIT>

>@BK2,E <XMIT>

>LOC DETECTED <XMIT>(NOTE CAREFULLY: do not use <@>)

This will search out and display the phrase " — ERRORS DETECTED" at the end of the SPSSX run without actually printing the run itself on the screen.

>EXIT <XMIT>(NOTE CAREFULLY: do not use <@>)

Even if errors were detected, you might wish to print the run so that you could examine the error message. Printing may be done by typing:

>@BK2 [specify printer] <XMIT>

If you do not wish to get a printout after all, you may delete the output by typing instead of >@BK2:

>@BK2,D <XMIT>

(3) If you plan to send the output from the command file to a printer to get a "hard copy" printout, you may examine on the CRT screen the print file which is created, before actually requesting the copy. That is, you may see on the screen exactly what will be printed, including any error messages, if you decide to print the output rather than deleting it. Type:

>@BK1 <XMIT>

>@SPSSX <XMIT>

>@ADD FILENAME. <XMIT> or >@ADD FILENAME.ELEMENTNAME <XMIT>

>@BK2,C <XMIT>

This will put into the CTS workspace the print file created by >@BK1, >@BK2,C, and you can examine it by using the normal CTS commands (such as ->P A). After examining the print file in CTS, you must decide whether to print the file or delete it. In either case, you must first exit CTS:

->XCT <XMIT>

If you wish to print the file, type:

>@BK2 [specify printer] <XMIT>

If you wish instead to delete the file without printing it, type:

>@BK2,D <XMIT>

6.2. Running an SPSSX Command File

SPSSX can be run in at least four different ways:

(1) On-line (on the CRT screen), in CTS:

->OLD FILENAME. <XMIT> or ->OLD FILENAME.ELEMENTNAME <XMIT>

->CALL SPRUNX <XMIT>

(2) On-line, in ECL:

>@SPSSX <XMIT>

>@ADD FILENAME. <XMIT> or >@ADD FILENAME.ELEMENTNAME <XMIT>

(3) Hard copy to printer, in ECL:

>@BK1 <XMIT>

>@SPSSX <XMIT>

>@ADD FILENAME. <XMIT> or >@ADD FILENAME.ELEMENTNAME <XMIT>

>@BK2 [specify printer] <XMIT>

(4) Hard copy to printer, in CTS:

->OLD FILENAME. <XMIT> or ->OLD FILENAME.ELEMENTNAME <XMIT>

->PXQ BK1 <XMIT>

>CALL SPRUNX <XMIT>

>PXQ BK2 [specify printer] <XMIT>

If errors are reported, it will be necessary to go into CTS, call up the file, correct the errors, and try again. To check for SPSSX errors before actually executing the command file, see the section "Checking for Errors on SPSSX Runs."

7. For DPS Users

7.1. Calling Up the DPS Processor

>@DPS FILENAME.ELEMENTNAME <XMIT>

A DPS script must be entered as an element of some cataloged file; the element that is created will be an omnibus element. If the element is new, DPS will start in INPUT mode:

DPS 1K DATE TIME ELEMENTNAME

INPUT

If the element already exists and you are returning to it to revise or execute it, DPS will start in COMMAND mode:

DPS 1K DATE TIME ELEMENTNAME

7.2. Getting from One Mode to Another in DPS

One of the most confusing things about DPS is remembering what mode you ought to be in, and how to get into that mode from the one you're in at the moment. Here's some help:

If you're in INPUT (>) and wish to go to EDIT (*>):

> <XMIT> (simply enter a completely blank line)

If you're in INPUT (>) and wish to go to COMMAND (?>):

> <XMIT> (blank line)

*>EXIT <XMIT>

If you're in EDIT (*>) and wish to go to INPUT (>):

*> <XMIT> (blank line)

If you're in EDIT (*>) and wish to go to COMMAND (?>):

*>EXIT <XMIT>

If you're in COMMAND (?>) and wish to go to EDIT (*>):

?>EDIT <XMIT>

If you're in COMMAND (?>) and wish to go to INPUT (>):

?>EDIT <XMIT>

*> <XMIT> (blank line)

7.3. The Most Useful Directives and Commands in DPS

There are dozens of directives and commands in the DPS manual. You can process a lot of documents with relatively few of them. Virtually all of this guide was processed with the following ones. If I did it, you can, too.

Directives in INPUT mode:

.FONT

.CONTROL TAB

.CONTROL UNDERLINE

.ON and .OFF FILL,COMPACT,JUSTIFY

.ON and .OFF CENTER

.SKIP

.EJECT

.INDENT

.0, .1, and .2 titling directives

.TOC (table of contents) directive

Commands in EDIT mode:

TOP

LOCATE

CHANGE

DELETE

INSERT

INSERT BEFORE

RETYPE

Commands in COMMAND mode:

DEVICE

OUTPUT FILE

FORMAT

FINAL

LIST

PROOF

7.4. Exiting from the DPS Processor

In COMMAND mode, type:

?>@EOF <XMIT>

The computer will respond:

END DPS

8. Logging-Off

If you fail to log-off properly after using the computer, someone else may be able to use your account for his own purposes and may even be able to change or delete your files. It is strongly recommended, therefore, that you log-off properly.

>@FIN <XMIT>

This command may also be given in CTS (manual control) and in EDIT and COMMAND modes in DPS, so the following prompt characters are also possible:

->@FIN <XMIT>or *>@FIN <XMIT> or ?>@FIN <XMIT>

The computer will then give some statistics about the run and the account and report:

*TERMINAL INACTIVE*

You should then type:

>@@TERM <XMIT>

There may be times when the computer will log you off on its own accord. If the system "crashes," you of course cannot log-off even if you wish to leave the system. If it is "down" for an extended period, it will automatically log you off and there is nothing you can do about it (except to log back on when the system comes "up").

If you fail to enter any lines for a lengthy period (you might be pondering what to do next, or you might be engaged in a conversation, or you might have left the area to get a drink of water), after several minutes of inactivity, the computer may log you off automatically. The natural way to avoid this is to enter a line often enough to prevent it. Always enter something as soon as you see the TIMEOUT WARNING flash on the screen — if you don't, you will very shortly be logged off for inactivity. If you must be inactive for several minutes, you may avoid being logged off automatically by going to ECL and typing:

>@H <XMIT>

The computer will respond:

*** THIS TERMINAL IS IN USE BY: USERID DATE TIME

About every four minutes it will print a new line with the new time. To end the hold, type: <@> <XMIT>. (You may have to enter it twice; the first entry may give you another message.) (The @H command may also be given in EDIT or COMMAND mode of DPS, in which case it will take you out of DPS to ECL.)

9. About the Author

The original manuscript of this guide came to light one Saturday afternoon in May 1982, when a Memphis State University professor was strolling on the beach near his summer cabin at Lake Woebegone, Mississippi (not to be confused with the more famous lake with a similar name in Minnesota, it is a nice lake nevertheless). The manuscript was found inside a bottle which had washed ashore, and although it was somewhat damp and water-stained, it was perfectly readable. Intrigued, the professor floated a message back out into the lake and was delighted the next weekend to find a reply awaiting him on the beach. As a result of many interchanges of messages, the professor received permission to show the manuscript to the staff of M.S.U. Computer Services. The staff made numerous suggestions for corrections and improvements, which were graciously accepted by "A. Survivor," and eventually Survivor gave his consent for publication of this guide.

Little is known of the author. His bottled messages have revealed that his real name is "Addled Survivor," but he prefers "A. Survivor" as being more dignified and British-like. His former profession remains a mystery, although he obviously at one time had some contact with the academic world. He is now, by deliberate choice, a hermit living on a secluded island far out in the middle of Lake Woebegone. He says that he lives largely on government food stamps, supplemented by the trapping of wild animals and the growing of a vegetable garden, and hereafter (he hopes) the royalties which will come to him from this publication.