Except for the next subsection on sending escape sequences to the terminal, this section mainly presents the details of setting up the terminal manually by sitting at the terminal and going thru menus. If you haven't already done so, you should read Terminal Set-Up (Configure) Overview. It's best if you have a terminal manual, but even it you don't there is information here on many of the options which you might possibly need to set.
The communication parameters such as its baud rate must always be set up at the terminal since if this is not done there can be no communication with the terminal. Once communication is established you have two choices for doing the rest of the terminal configuration. You may continue to configure manually at the terminal and save the results in the terminal's non-volatile memory or you may do this by sending escape sequences to the terminal from the computer each time the terminal is powered on (or the like).
If you know how to set up and save a good configuration inside the terminal it may be the best way. If you don't, you might want to just send the init string from terminfo to your terminal each time you use the terminal. Perhaps doing nothing will still give you a usable terminal. You (or an application program) can always change things by sending certain escape sequences to the terminal.
Once the communication interface is established, the rest of the configuration of the terminals may sometimes be done by sending escape sequences to the terminals from the computer. If you have a large number of terminals, it may be worthwhile to write (or locate) a shell script to automatically do this. There may (or may not) be a command you can send to a terminal to tell it to save its current set-up in its non-volatile memory so that it will be present the next time the terminal is powered on.
There is an simple way to send these escape sequences and a complex way. Using the simple way, you never look up escape sequences but issue commands that automatically find an appropriate escape sequence in the terminfo database and send that. Unfortunately, not all the escape sequences which you might want to send are always in the terminfo database. Thus the more complex (but possibly better) way is to directly send escape sequences.
For this complex method you'll need an advanced manual. Old terminal manuals once included a detailed list of escape sequences but newer ones usually don't. To find them you may need to purchase another manual called the "programmers manual" (or the like) which is not supplied with the terminal. A Esc Sequence List for some terminals is on the Internet but it's terse and likely incomplete.
Even without a manual or the like, you may still send commands to configure the terminal by using the programs "tput" and "setterm". See Changing the Terminal Settings. You could just send the terminal an init string from the terminfo entry if the init string sets up the terminal the way want it. See Init String. Unless you plan to have these sequences sent from the computer to the terminal each time the terminal is powered on, you must somehow save the settings in the non-volatile memory of the terminal.
On older terminals look at the keyboard for labels just above the top row of numeric keys. If they exist, these labels may be what these keys do in set-up mode. Some older terminals may have only one "set-up" menu. Still older ones have physical switches. In some cases not all the switches are well labeled but they may be well concealed. Of course, if you set something with a switch, it's "saved" and there is no need to save the setting in non-volatile memory.
To select options (configure) at the terminal, you must first enter "set-up" mode and then select options (i.e. configure) using menus stored inside the terminal and displayed on the screen. To do this, the terminal does not even need to be connected to a computer. How to get into set-up mode is covered in the terminal's manual, but here's some hints that may help:
If there's a "set-up" key try pressing it. Also try it shifted.
To move around in the set-up menus, try the arrow keys. Use Return, Space, or a special key ("toggle" on old terminals) to select. To exit set-up mode select exit from a menu (or on some older terminals press the set-up key again).
For the terminal to work at all, speed, parity, :its/character, and communication mode must be set correctly. Incorrect flow control may cause loss and/or corruption of data seen on the screen. The essential communication options were dealt with (for both the terminal and computer) in another section: See Communication Interface. The following list provides some links to that section, as well as some additional communication options set only at the terminal.
Your set-up must be saved in the non-volatile memory of the terminal so that it will be effective the next time you turn on the terminal. If you fail to save it, then the new settings will be lost when you turn off the terminal. Before you go to the trouble of setting up a terminal, make sure that you know how to save the settings. For modern terminals the save command is done via a menu. In some older terminals, only the manual tells how to save. For many of these you press Ctrl-S to save.
What follows in this section describes some of the options which are available in the set-up menus of many terminals. Options are also called parameters or features. Many options may be called "modes". Setting options is often called "configuring". Many of these options may also be set by sending certain escape sequences to the terminal. Different models and brands of terminals have various options and the same option may be called by different names (not all of which are given here) Terse names used by Wyse are enclosed in {...}. Names used mostly for VT terminals are enclosed in {{...}}.
Most modern terminals can emulate several other terminals. The terminal can likely do more if it is set to emulate itself (actually no emulation) {native personality}. Sometimes there are 2 different emulations for the same model of terminal. For example VT220-7 emulates a VT220 with 7-bits/byte while VT220-8 emulates a VT220 with 8-bits/byte (256 possible characters).
Older models of terminals usually have fewer features than newer models. Suppose one wanted to emulate an old terminal but also wanted some of the advanced capabilities of the later model terminal they are sitting at. This is sometimes possible (to some degree). This feature is sometimes called {Enhance} (or Enhanced ??).
This is the size of the cell in which a character fits. It is measured in pixels (=tiny dots). The more dots, the better the resolution. 10x16 is 10 dots wide by 16 dots high (16 rows and 10 columns). Note the notation is inverted as compared to the notation for matrix dimensions which gives rows (height) first.. Also, the character cell includes rows and columns of pixels allocated for the space between adjacent characters so the cell size which defines the boundaries of an actual character may be smaller.
Usually 80 columns and 24 or 25 lines. This means that there may be up to 80 characters in a row (line) on the screen. Many terminals have a 132 column option but unless you have a large screen, the tiny characters may be hard to read. {{Set 132 column mode}}. If you set 25 lines, make sure that this is in the terminfo. You should also put "export LINES=25" into /etc/profile and also use: "stty -F /dev/ttySx rows 25". If you don't it might result in a scrolling problem (see Terminal doesn't scroll
The cursor may be set to appear as a rectangle (= block) {Blk}. Other options are underline {Line} or blinking. I prefer non-blinking {Steady} block since it's big enough to find quickly but there is no distractive blinking. If you set it invisible (an option on some terminals) it will disappear but new letters will appear on the screen as you type at the invisible cursor.
Display Attributes may either be magic cookies or be attribute bytes assigned to each character. For magic cookies, there is a limit to their extent: Are they in effect to the end of the line or to the end of the page? It's best to use attribute bytes (which could actually be half-bytes = nibbles).
May be called various names such as "Display Controls". When off (normal) it's "Interpret Controls". When set on, you see the escape sequences from the host (which you normally never see on the screen). So that these sequences may be viewed in sequence on a line, they are not acted upon (interpreted) by the terminal. Except that a CR LF sequence creates a new line. See Control Codes.
Some terminals can have their characters double width and/or double height. This feature is seldom needed. When changing a line to double width (DW) the right half (RH) is pushed off the screen and there is the question of whether or not to delete (erase) it. "Preserve" means to keep the RH of DW lines. When in double height mode, it may be necessary to send each such line twice (the 2nd time down one row) in order to get a double-height line on the screen.
Normal video is light (white, green, amber) letters (foreground) on a dark (black) background. Reverse video {Display Light} is the opposite: black text on a light background. This is easier on the eyes (unless the room is dark).
A status line is a line at the top or bottom of the screen that displays info about the application program you are running. It's often highlighted in some way. With a status line enabled, an application can send the terminal a special escape sequence which means that the text that follows is for the status line. However, many applications don't use this feature but instead only simulate a real status line by direct cursor positioning. The ordinary user looking at it doesn't know the difference.
When switching the number of columns from 80 to 132 (or conversely) should the data displayed in the old format be erased (cleared) or preserved? {80/132 Clr} {{Screen Width Change}}. It should make no difference how you set this option since if an application program uses 132 columns, it should set this option appropriately via a control sequence.
For a Wyse terminal to be able to access multiple pages of display memory {Multipage} must be set to on.
The terminal memory may be divided up into a number of pages. See Pages and Pages (definition) for explanations of pages. You may partition the page memory into a number of pages of selected length. Linux applications don't seem to use pages at present so it shouldn't make much difference how you set this up.
The terminal memory may be divided up into a number of pages. See Pages and Pages for explanations of pages. When the cursor is moved to a location in video memory not currently displayed (such as another page, or on the same page but to a location not displayed on the screen) should the display change to let one view the new cursor location? If so, this is called "Coupling". For cursor movement within the same page there is "Vertical Coupling" and "Horizontal Coupling". For movement to another page there is "Page Coupling".
The terminal will identify itself and its state, or send out a pre-recorded message in response to certain escape sequences.
You may write a short message during set-up which may optionally be sent to the host at power-up or be sent to the host in response to a request from the host (perhaps the ENQ (inquire) control character).
If set, sends the answerback message to the host at power-on without the host asking for it. Do any "getty" processes look for this ??
If set, will never let anyone see the answerback message (except of course the host computer). If it needs to be changed, deselect "answerback concealed" and the formerly concealed message will be destroyed so you then may enter a new message (but you don't get to see the old one).
The terminal sends this reply in answer to a request for identity.
When set, pressing any key makes a click (broadcast by a tiny loudspeaker in the keyboard). These clicks annoy some people and I think it's best to set keyclick off.
When the Caps-Lock key is down, should only the alphabetic keys generate shifted characters? If set to {Caps} or upper-case-only then hitting a number key with the Caps-Lock on will type the number. To get the symbol above the number one must manually hold down the shift key. This is the normal mode. If set to {Shift} then all keys type the shifted character when Caps-Lock is on (hitting the 5 key should type % without holding down Shift, etc.).
If a key is held down then that key is repeatedly "typed". This is handy for repeatedly typing the same character to create a line across the page.
When the cursor is 8 columns away from the right side of the display, a bell is rung (like on an old typewriter). Almost all editors will automatically create a new line if needed (no need to hit the Return key) so this feature is seldom needed.
The code sent to the host when a key is pressed is normally the ASCII code for that key (depends also on Shift and Control key). On some terminals you may make any key send any code you wish. That is, you may completely remap the keyboard by setting up the terminal that way. This may be useful for some foreign languages and Dvorak keyboard layouts, etc. which permit one to type faster. Even for terminals that don't have the feature, there is software to remap the keyboard (and screen also). It's something like a device driver which uses a pseudo terminal. See Character Mapping: mapchan
Wyse terminals have a key near the lower left corner which may be set to do various things. Its may be labelled "Funct", "Compose Character", "Alt", "Hold" or "Scroll Lock". Early models don't have all of the following options:
The numeric keypad (the rectangle of mostly numeric keys to the right of the main part of the keyboard) can be set to send special codes which will do special things in certain application programs. Ditto for the arrow keys. There is thus a "normal" mode where they send what is shown on the keycap (or the normal code sequence for an arrow-key) and an "application" mode where special escape sequences are sent. In some cases there is a "hex" numeric mode which is almost like normal numeric mode except that 6 non-numeric keys send the letters A-F. Thus one may type for example "B36F" on the numeric keypad.
Depending on how they're set up, shifted-del sometimes sends the control character CAN and shifted backspace sometimes sends DEL.
Many terminals can emulate a PC keyboard by sending PC scancodes (see Keyboard-and-Console-HOWTO) instead of ASCII codes. This is mostly used with special Multiuser DOS OSs. It won't work with ordinary MS DOS. See Non-Linux OSs However, hardly any Linux programs that run via the serial port can accept scancodes. If this is the latest version of this HOWTO, let me know if any programs do this. I think Foxpro can do it. You need to define smsc and rmsc in the terminfo, and perhaps pctrm.
When using scancodes it's best to use hardware flow control since normal software flow control conflicts with some of the codes (??). If you do use software flow control, you must use the XPC type of flow control. It uses 0x65 and 0x67 for on and off characters. It must be set this way both in the terminal and by stty for the PC.
Some keys may have alternative letters on them. When keys is set to "Typewriter" they send what they would normally send on a typewriter. When keys is set to something else the alternative characters are sent.
In this case "New Line" means a new line starting at the left margin below the current line. In Linux and C "new line" (NL) may have a different meaning: the line-feed control character LF also called new-line or NL. This is because in Linux text files, the LF character means a "new line starts here" so it's labeled NL. Normally, a LF (NL) sent to a terminal only results in the cursor jumping down one line below where is was and does not move the cursor back to the start of this "new line".
If Auto New Line is set, the above "normal" situation is canceled and a physical new line is created on the display upon receiving a LF from the host. This is exactly what one wants in Linux. Except that (when Auto New Line is set) the Return (or Enter) key sends a CR LF sequence to the host (for Wyse and VT100, but for VT420 ??). Since Linux uses LF as a "new line" marker in files, Linux would like only a LF to be sent (and not a CR LF). Thus the "New Line" option is seldom used. Instead, the required translations are made by the serial port device driver by default. It is as if one gave the command "stty onlcr icrnl". But you don't need to do this since it's the default.
This is just another type of "Auto New Line". When a CR (carriage return) character is received, a LF (line feed) action is added resulting in a new line being displayed. Since Linux marks the end of lines with LF, this option is not used.
If off, the DEL character received by the terminal is ignored. If on the DEL performs a destructive backspace. Null characters are usually ignored in any case. Both DEL and NULL are sometimes used for padding. See Padding
Also called "Auto Wrap(around)". What happens when the right edge of the screen is reached (col. 80, etc) and no return character (or the like) has been sent from the host? If Line Wrap is set, then the rest of the line displays on the line below, etc. Otherwise, the rest of the line is lost and is not seen on the screen. Any good application should handle the situation correctly (provided the terminfo knows how Line Wrap is set). Thus even if Line Wrap is not set, the application should either wrap the screen for long lines or provide another way for you to view the cutoff tail of long lines (by use of the arrow keys, etc). But a raw copy command (and other situations) may not do this so it's often best to set line wrap.
For an 80 col. screen, most terminals only wrap if the 81st character from the host is a graphic (printable) character. This allows for the case where 81st character from the host might be "return" or a "newline" (non-graphic characters) which means that the application is handing the wrapping OK and intervention by the terminal is not needed.
Scrolling {Scrl} is where all the lines on the screen move up or down. Its also called "panning" which includes movement sideways. In ordinary scrolling lines roll off the bottom or top of the screen and disappear, and new lines from the host appear at the opposite edge (top or bottom). There are 3 types of this: smooth, jump, or burst. Burst is not really scrolling since its an instant replacement of an old screenfull by a new one (although some lines on the new screen may be from the old screen). Jump is where new lines jump into view one at a time. Smooth {Smth} is where the text moves at a steady speed upward or downward. If the smooth scroll rate is slow enough, one may read the newly visible lines when they are still scrolling (in motion).
Smooth scrolling on slow terminals was once useful since one could continue reading as the display was scrolling. But with higher baud rates, jump scroll is so fast that little time is lost as the new display appears. Since it takes a little longer to read scrolling text than fixed text, it may actually waste more time if smooth scrolling is selected.
If (auto)scrolling {Autoscrl} is disabled, then new text from the host must go somewhere so it is put at the top of the display. If the old text is not erased, the new text merges (nonsensically) into the old. If the old text is erased, then the new text is out of context. So keep (auto)scrolling enabled.
See Pages and Pages for explanations of pages. When the current page is full (the last line is finished) should the page scroll, or should a new page be created (leaving the previous page stored in the terminal's display memory). If {Autopage} is set, then a new page is created. Since you are probably not using pages, you should probably set this to off.
These are the keys labeled F1, F2, etc. On older terminals they may be labeled PF1, PF2, etc. where the P stands for Programmable. Some keyboards have both. One may program (redefine) these keys to send out a string of user-defined bytes. This doesn't work for older terminals such as vt100 or for the Linux command-line interface (how sad). They may often be easily "programmed" using a certain set-up menu {FKey}. On some terminals, one may also specify where this string is sent to when the key is pressed. In "normal" mode, pressing the key is just like typing the string at the keyboard. In "local" mode pressing the key sends it to the terminal (just like if the terminal was in local mode). This may be used to send escape sequences to the terminal so as to configure it in a special way. In "remote" mode, the string is always sent out the serial port to the host computer (even if the terminal is in local mode).
Some options are only for the case of Block Mode. This option is powerful since it provides forms and takes load off the host computer by transmitting in bursts. But it's more complicated to set up and is thus not used too much.
In block mode some regions of the screen are for the text of forms and are thus write-protected "Prot" {WPRT}. Options may set the characters in these regions to appear dim, reverse video {WPRT Rev}, and/or underlined {WPRT Undrln}. {WPRT Intensity} may be set to dim, normal, or even blank (invisible)
Should write-protected text (the original text in the form) be sent to the host upon transmission of a block: {Send All} or is write-protected text also read-protected: {Send Erasable}
Should the entire screen be sent or just the scrolling region? {Send Area}. Should the sending stop when the current cursor position is reached? If {Xfer Term} is set to Cursor, only the data on the screen up to the cursor is sent.
What is the termination symbol to be appended to a block of data? {Blk End} or at the end of a page {Send Term}ination.
There are various types of Locks. One is the Locked keyboard due to flow control. See Keyboard Lock Another lock {Feature Lock} is that which prohibits the host computer from changing the terminal set-up by sending certain escape sequences to the terminal. Placing such a lock may result in unexpected behavior as application programs send escape sequences to the terminals that are ignored. Not all set-up parameters lock. Unless you have a good reason to do so, you should not enable such locking.
A Function Key lock will prohibit the computer from redefining what a programmable function key sends. You may want to use this if you have something important programmed into the function keys.
Also called "CRT Saver". This turns off (or dims) the screen after the terminal is not used for a period of time. It may prolong the life of the screen and save some energy. Hitting any key will usually restore the screen and may "execute" that key so it's best to hit the shift-key, etc.
For Wyse, if there is no {Printer Attached} set it to Off. It's not essential to do this, but if you do it any escape sequence to send text to the printer (instead of the terminal) will be ignored.
Setting up the printer port is about the same (usually simpler) as setting up the communications on the main port. There are a couple of options specific to the printer. Is the printer a serial or parallel printer? If it's parallel it should be designated as such in setup and connected to the parallel port on the terminal (if there is one). Should a FF (form feed) be sent to the printer at the end of a print job? If {Print Term} is set to FF, this will happen.