Adobe Serial and Parallel

Communications Protocols
Specification
Adobe Developer Support

20 November 1992

Adobe Systems Incorporated

Adobe Developer Technologies
345 Park Avenue
San Jose, CA 95110
http://partners.adobe.com/

PN LPS5009

Copyright 19901992 by Adobe Systems Incorporated. All rights reserved.

No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form
or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior
written consent of the publisher. Any software referred to herein is furnished under license and may
only be used or copied in accordance with the terms of such license.
PostScript is a registered trademark of Adobe Systems Incorporated. All instances of the name
PostScript in the text are references to the PostScript language as defined by Adobe Systems
Incorporated unless otherwise stated. The name PostScript also is used as a product trademark for
Adobe Systems implementation of the PostScript language interpreter.
Any references to a PostScript printer, a PostScript file, or a PostScript driver refer to printers,
files, and driver programs (respectively) which are written in or support the PostScript language.
The sentences in this book that use PostScript language as an adjective phrase are so constructed to
reinforce that the name refers to the standard language definition as set forth by Adobe Systems
Incorporated.
PostScript, the PostScript logo, Adobe, and the Adobe logo are trademarks of Adobe Systems
Incorporated which may be registered in certain jurisdictions. Other brand or product names are the
trademarks or registered trademarks of their respective holders.
This publication and the information herein is furnished AS IS, is subject to change without notice,
and should not be construed as a commitment by Adobe Systems Incorporated. Adobe Systems
Incorporated assumes no responsibility or liability for any errors or inaccuracies, makes no warranty
of any kind (express, implied or statutory) with respect to this publication, and expressly disclaims any
and all warranties of merchantability, fitness for particular purposes and noninfringement of third
party rights.

Contents

Adobe Serial and Parallel Communications Protocols Specification 5

1

Introduction 5

Adobe Standard Protocol 5

Definition of Newline 6
Enabling the Standard Protocol 7

Adobe Binary Communications Protocol 8

Functional Description of the Protocol 8
Using the Binary Communications Protocol 10
Character Protocol 10
Differences Between Standard and Binary Protocols 13
Enabling Binary Communications Protocol 14

Adobe Tagged Binary Communications Protocol 16

Functional Description of the Protocol 16
Character Protocol 17
Additional Points 19
Uses in a Language Switching Environment 20
Appendix: Changes Since Earlier Versions Template 21
Index 23

iii

iv

Contents

(20 Nov 92)

Adobe Serial and Parallel

Communications Protocols
Specification

Introduction
This document describes several protocols that can be used to communicate
over a serial or parallel connection to a PostScript printing device, including
standard protocol, binary communications protocol (BCP), and tagged
binary communications protocol (TBCP).
The protocols described here are link-level protocols that are specific to serial
and parallel communications channels. The protocols define special character
sequences to indicate special control functions that are not logically part of
the data stream.
Other communications channels provide such functions in entirely different
ways, such as with special packet types in a local area network (LAN). When
communicating via such channels or when saving a PostScript language program to a file, it is never appropriate to embed any of the control sequences
described in this document. For an elaboration on this topic, see the section
Communication Channel Behavior on pp. 74-75 of the PostScript Language Reference Manual, Second Edition.

Adobe Standard Protocol

The Adobe standard protocol is a simple communications protocol. Data is
sent and received in ASCII. Several character codes are used directly by the
communications driver for communications functions and are not passed
through to the PostScript interpreter. These are listed in Table 1. Throughout
this document, a control character is indicated by a ^ prefix.

Table 1 Special characters in the standard protocol

ASCII
keyboard

ASCII
name

Value
(hex)

Control function

^A

SOH

0x01

Quote character for BCP & TBCP

^C

ETX

0x03

Generate an interrupt error

^D

EOT

0x04

End-of-file marker

^Q

DC1

0x11

XON in XON/XOFF flow control

^S

DC3

0x13

XOFF in XON/XOFF flow control

^T

DC4

0x14

Job status request

^[

ESC

0x1B

Start of end-protocol for TBCP

In Table 1, the ^A and ^[ are only actually recognized by the standard communications protocol when the tagged binary communications protocol also
exists, since they are used to begin and end the tagged binary communications protocol.
In the standard protocol, there is no way to quote the reserved characters (in
order to pass them through to the PostScript interpreter), nor is there any way
to transmit characters in the high ASCII range (128 to 255) if the high order
bit is used for parity (odd, even, space or mark). However, this causes little
difficulty in normal use since the standard PostScript language character set
consists entirely of printable characters (plus the space, tab, and newline
characters; see section 3.2.2, ASCII Encoding of the PostScript Language
Reference Manual, Second Edition for more information). The language
itself provides a means for encoding arbitrary characters in strings via the
octal \nnn escape sequence. True binary data, such as images and encrypted
programs, must be transmitted in the hexadecimal (base 16) encoding or the
ASCII base-85 encoding (Level 2 only) when using the standard protocol.
Binary data may be transmitted on serial and parallel channels using one of
the binary protocols described later in this document.
The standard protocol is supported on the RS232 and centronics channels of
all devices.
2.1

Definition of Newline

The characters carriage-return (decimal ASCII 13) and line-feed (decimal

ASCII 10) are also called newline characters. A carriage-return followed
immediately by a line-feed are treated together as one newline by the PostScript interpreter when in the standard protocol mode. When a newline character is written to the standard output file, it is translated to the two-character
sequence carriage-return, line-feed.

Adobe Serial and Parallel Communications Protocols Specification

(20 Nov 92)

2.2

Enabling the Standard Protocol

Many printers have only the standard protocol for communicating via the
serial and parallel channels. Therefore, it is not labelled the standard protocol in printer documentation, but merely discussed as serial communications. If it is the only protocol, then there is no need to set it. However, on
printers that support other communications protocols, you may need to establish the standard protocol when you wish to use it.
Note

The proper way to enable the standard protocol differs between Level 1 and
Level 2 printers. However, it is important to note that in both situations, the
protocol change does not take effect until the end of the setup job. Therefore, the protocol must be invoked as a separate job, rather than prepended to
another PostScript language job. These methods are standard across all
devices that support this protocol.
On Level 1 printers, the communications protocol setting is associated with
the current input/output mode. To return to the standard protocol from the
binary communications protocol execute setsoftwareiomode in statusdict
passing in the value of 0. For example,
%!PS-Adobe-3.0 ExitServer
%%Title: (Return to Standard Protocol - Level 1)
%%EndComments
%%BeginExitServer: 0
serverdict begin 0 exitserver
%%EndExitServer
statusdict begin
/setsoftwareiomode known {0 setsoftwareiomode}
end
%EOF

In Level 2, the standard protocol is set with the setdevparams operator setting the Protocol parameter to a value of Normal. Also, set the Interpreter
parameter is set to the value PostScript. This will prevent configuration
errors if the Interpreter parameter was previously set to something else. For
example,
%!PS-Adobe-3.0
%%Title: (Set up Standard Protocol - Level 2)
%%EndComments
currentsysparams
/CurInputDevice 2 copy known {
get
% (%Device%)
<</Protocol /Normal
/Interpreter /PostScript>> setdevparams
}{
pop pop
} ifelse
%EOF

2 Adobe Standard Protocol

Adobe Binary Communications Protocol

The Adobe binary communications protocol is an additional method of communicating between a PostScript printer and a host computer using an 8-bit
wide serial or parallel channel. The protocol allows any of the 256 possible
8-bit values to be transmitted as data, but also allows certain characters to be
used for specifying out-of-band control functions which may be handled synchronously or asynchronously by the communications driver. These control
functions include flow control, status requests, aborting of jobs, and end-offile markers.
The definition of the protocol is completely symmetric between the two communicating devices and does not assume that one system is a slave of the
other. Binary means that all 256 byte values of 8-bit data may be sent and
interpreted as data by either side. In addition, certain 8-bit values are reserved
for out-of-band communications and control purposes. Although these last
two sentences may seem contradictory, they express, in essence, the problem
that the protocol solvessending more than 256 distinct values encoded into
a 256-valued alphabet.
Since any 8-bit value can be transmitted as data, this protocol can be used for
sending PostScript language jobs that contain binary images or for sending to
emulators data with any sequence of control characters. Since both PostScript
language jobs and emulation jobs can be sent, this channel allows software
switching between these modes without having to close and reopen the communications channel.
The binary communications protocol is supported on a few Level 1 printers
and many Level 2 printers. To determine whether or not the protocol is available, applications should check the PostScript Printer Description (PPD) file.
See the PostScript Printer Description Files Specification for more information.

3.1

Functional Description of the Protocol

The protocol is designed to be sent over channels that are logically 8 bits
wide and normally support only a byte stream protocol rather than a packet
protocol. The channel must be 8 bits wide because the fundamental unit that
is being transmitted is a byte. The two most important examples of interest
are asynchronous serial communications and 8-bit parallel communications.
The binary communications protocol is appropriate only over channels that
do not provide the same functions in other ways.
Although the definition of the protocol is completely symmetric, the particular functions of some of the out-of-band information are asymmetric in
nature. The underlying assumption of the protocol is that files of data will
be sent from one computer (referred to as the host) to the other (referred to as

Adobe Serial and Parallel Communications Protocols Specification

(20 Nov 92)

the server) to be executed, each file as an individual job. The specific motivation for the protocol is files sent from a computer to a printer, where each
file is considered a print job. One could envision other uses. For a more complete description of a job execution environment for interpreting files, see
section 3.7.7, Job Execution Environment of the PostScript Language Reference Manual, Second Edition.
The out-of-band information (the non-data communications and control
information) falls into two broad categories: those functions that are asynchronous with the data stream and those that are synchronous with the data
stream. The protocol defines four out-of-band functions that are asynchronous and one that is synchronous.
The asychronous functions are:
job status request. The server should respond immediately to job status
request from the host by sending appropriate data back to the host. The
syntax and semantics of the returned information are not specified by the
protocol.
job abort request. The server should respond immediately to job abort control from the host by terminating processing of the current job and flushing
through the input stream until an end-of-file marker or end-protocol
sequence is encountered. The server should proceed at that point with processing the end-of-file or end-protocol in the normal manner.
XON flow control. The party receiving an XON may resume transmitting
data that was blocked by a preceding XOFF. The XON and XOFF functions are present to support the well-established XON/XOFF flow control
protocol used over asynchronous serial communications channels.
XOFF flow control. The party receiving an XOFF should cease transmitting data as quickly as possible. It may still transmit asynchronous control
functions, especially XOFF and XON. Flow control operates independently for each direction of data transmission.
The synchronous function is:
end-of-file indication
As mentioned above, some of these functions are asymmetric in natureit
would not be expected, for example, that the server would send job abort
request to the host.

3 Adobe Binary Communications Protocol

3.2

Using the Binary Communications Protocol

The binary communications protocol is most useful for sending 8-bit data to
printers over serial or parallel channels, without interrupting flow control
functions. Data typically best suited for use with the binary communications
protocol is binary image data, data output from a compression filter (Level 2
Only), or show strings that access characters encoded in the unprintableASCII range (below 32 decimal). An example is the IBM PC extended character set.
Using the binary communications protocol results in a significant performance advantage for users of serial and parallel ports. Using the standard protocol, the only way to send binary data over these channels is to use ASCII
hexadecimal for image data (a 1:2 expansion of the data), the octal (\nnn)
notation for string data (a 1:4 expansion), or the ASCII base-85 encoding (4:5
expansion) (Level 2 only). The binary communications protocol offers a 1:1
transmission for most characters, with a 1:2 expansion for the few reserved
characters.
Because the binary communications protocol is a device-dependent feature, it
should not be used in Encapsulated PostScript (EPS) files or in print-to-disk
files. In such cases, it is safer to avoid binary data or to use binary data with
no escaping. Note that the binary communications protocol can be applied
later to a job stream without needing any high-level knowledge about the job.
Instead, it should be used only when the printer driver (or host application
sending the job to the printer) can determine that it is connected directly to a
printer. If your application does not have the ability to determine this, then
the choice of using the binary communications protocol should be brought to
the user-interface level.
3.3

Character Protocol

As mentioned in section 2, several character codes are used directly by the

communications driver for communications functions and are not passed
through to the PostScript interpreter.

10

Adobe Serial and Parallel Communications Protocols Specification

(20 Nov 92)

Table 2 Special characters in the binary communications protocol

ASCII
keyboard

ASCII
name

Value
(hex)

Control function

^A

SOH

0x01

Quote data character

^C

ETX

0x03

Generate an interrupt error

^D

EOT

0x04

End-of-file marker

^E

ENQ

0x05

(Reserved for future use)

^Q

DC1

0x11

XON in XON/XOFF flow control

^S

DC3

0x13

XOFF in XON/XOFF flow control

^T

DC4

0x14

Job status request

^\

FS

0x1C

(Reserved for future use)

To transmit these characters as data, they must be quoted. Quoting is done by

replacing the character with a two-character sequence: the special character
^A (ASCII hex 0x01) followed by the character itself XORed with 0x40. For
example, to send a byte with the value 0x14 (^T), the two-byte sequence 0x01
0x54 (^A T) is sent since ASCII T is the result of XORing ^T with 0x40. This
method of quoting guarantees that whenever any of the special characters are
received from the host computer, the control function is intended, regardless
of whether the preceding character is a ^A.
The generation and processing of asynchronous control characters, therefore,
can occur at a lower level than the generation and consumption of the data
stream. In particular, on a host computer, the ^A quoting convention can be
implemented by a user program while XON/XOFF processing is performed
independently by the operating system.
All 8-bit values other than those of the special control characters listed in
Table 2 are transmitted by simply sending the value.
After a ^A is received, the next character received that is not one of the special control characters must be the result of XORing one of the special characters with 40 hex. Receipt of any other character is considered an error in
the input. Between the ^A and the XORed character, any number of the special characters can appear, except for special characters that are handled synchronously^D and ^A. Receipt of ^D or ^A between a ^A and the XORed
character is considered an error.
When one of the special characters arrives, unquoted, and it specifies no control function for the channel, the character is discarded. For example, if XON
or XOFF is received and XON/XOFF flow control is not in use, XON or
XOFF is discarded.

3 Adobe Binary Communications Protocol

11

The characters ^E and ^\ currently specify no control functions. They are

included among the characters that must be quoted in case new control functions are added in the future.
The following is a sample implementation for quoting characters in C.
#define UCHAR unsigned char
#define
#define
#define
#define
#define
#define
#define
#define

CTRLA
CTRLC
CTRLD
CTRLE
CTRLQ
CTRLS
CTRLT
CTRLBKSL

(UCHAR)0x01
(UCHAR)0x03
(UCHAR)0x04
(UCHAR)0x05
(UCHAR)0x11
(UCHAR)0x13
(UCHAR)0x14
(UCHAR)0x1C

/* QuoteByte quotes a character if necessary, using the Adobe Binary

* Communications Protocol, and then writes the resulting bytes to
* the printer by calling outbyte()
*/
void QuoteByte(c)
UCHAR c;
{
switch (c)
{
case CTRLA:
case CTRLC:
case CTRLD:
case CTRLE:
case CTRLQ:
case CTRLS:
case CTRLT:
case CTRLBKSL:
outbyte(CTRLA);
outbyte((UCHAR)(c ^ 0x40));
break;
default:
outbyte(c);
}
}

12

Adobe Serial and Parallel Communications Protocols Specification

(20 Nov 92)

Table 2 shows the actions taken or the data put in the input buffer for byte
sequences received from the channel by the communications driver.
Table 3 Actions when bytes are received from hardware
Read from
hardware

Data in
input buffer

Type

Action

^C

Asynchronous

Generate interrupt

^D

End-of-file mark

Synchronous

^E

^Q

Asynchronous

XON (flow control)

^S

Asynchronous

XOFF (flow control)

^T

Asynchronous

Printer status request

^\

^A A

^A

Synchronous

^A C

^C

Synchronous

^A D

^D

Synchronous

^A E

^E

Synchronous

^A Q

^Q

Synchronous

^A S

^S

Synchronous

^A T

^T

Synchronous

^A \

^\

Synchronous

^A <any other char>

Comm-Err mark

Synchronous

<any other char>

<char received>

Synchronous

3.4

Differences Between Standard and Binary Protocols

This section highlights some of the differences between the binary communications protocol and the standard protocol for serial and parallel channels.
The binary protocol is entirely symmetrical. The control characters and the
quoting conventions apply in both directions, although most of the control
characters have no defined meaning in the printer-to-host direction.
In contrast to the standard protocol, there is no mapping between end-ofline conventions in binary protocol. The end-of-line characters (CR, LF, or
CR LF) sent by the host are received by the interpreter or emulator in the
printer. The PostScript interpreter handles the different end-of-line con-

3 Adobe Binary Communications Protocol

13

ventions in a uniform way, but a program that reads data from the channel
directly (using read or readstring) receives the characters as sent by the
host.
Whatever output is generated by a PostScript language program (using
print or ==) is sent unchanged. Note that the standard PostScript language
end-of-line (corresponding to the \n escape sequence in strings) which is
normally carriage-return, line-feed, now simply becomes line-feed. This is
especially noticeable in the executive mode of the interpreter.
3.5

Enabling Binary Communications Protocol

Before using the binary communications protocol, a driver must determine if

doing so is appropriatethat is, communications are via a serial or parallel
channel connected to a product that supports it. The binary communications
protocol is never supported over communications channels that are inherently
binary to begin with, such as Appletalk or SCSI. Additionally, the protocol is
not available in all printers. To determine whether or not the protocol is available, applications should check the PostScript Printer Description (PPD) file.
See the PostScript Printer Description Files Specification for more information.
Note

The proper way to enable the binary communications protocol differs

between Level 1 and Level 2 printers. However, it is important to note that in
both situations, the protocol change does not take effect until the end of the
setup job. Therefore, the protocol must be invoked as a separate job, rather
than prepended to another PostScript language job. These methods are standard across all devices that support this protocol.
On Level 1 printers, the binary communications protocol is associated with
the current input/output mode. This is set using setsoftwareiomode in
statusdict. For example,
%!PS-Adobe-3.0 ExitServer
%%Title: (Set up Binary Protocol - Level 1)
%%EndComments
%%BeginExitServer: 0
serverdict begin 0 exitserver
%%EndExitServer
statusdict begin
/setsoftwareiomode known {100 setsoftwareiomode}
end
%EOF

14

Adobe Serial and Parallel Communications Protocols Specification

(20 Nov 92)

On Level 2 printers, the binary communications protocol is set by setting the

Protocol device parameter to Binary with the setdevparams operator.
%!PS-Adobe-3.0
%%Title: (Set up Binary Protocol - Level 2)
%%EndComments
currentsysparams
/CurInputDevice 2 copy known {
get
% (%Device%)
<</Protocol /Binary>> setdevparams
}{
pop pop
} ifelse
%EOF

This sets the device corresponding to the current communications channel.

Note

It is important to check the PPD file to confirm the existence of the binary
communications protocol on a per-printer basis, because this is a devicedependent feature. More information on supporting device features can be
found in Technical Note #5117, Supporting Device Features.
Also note that if the system parameter password (SystemParamsPassword)
for the printer has been changed from the default, a Password entry will be
required in the dictionary passed to setdevparams.

3 Adobe Binary Communications Protocol

15

Adobe Tagged Binary Communications Protocol

This section describes the Adobe tagged binary communications protocol, a
protocol for bidirectional binary communications between two computers.
The tagged binary communications protocol is very similar to the Adobe
binary communications protocol though they are not compatible with each
other. Note that neither protocol is an extension of the other. See section 3,
Adobe Binary Communications Protocol for a description of the original
binary protocol.

4.1

Functional Description of the Protocol

This section details the differences between the Adobe binary communications protocol and the Adobe tagged binary communications protocol.
Upon entering the tagged binary communications protocol, a sender should
precede the data by a begin-protocol sequence. The characters that encode
the begin-protocol sequence (see section 4.2, Character Protocol) are an
illegal sequence in the Adobe binary communications protocol.
This choice was deliberate so that if one side is using the binary communications protocol and the other is using the tagged binary communications protocol, an error will be generated immediately.
The protocol may be thought of as a connection between the sender and the
receiver. The sender begins the connection by sending the begin-protocol
sequence. During the connection a sequence of files each separated from the
next by an end-of-file marker is sent from the sender to the receiver. The
sender terminates the session by sending an end-protocol sequence. Note that
the last file in the connection should be followed immediately by an end-protocol sequence. If the last file is followed by an end-of-file marker which is
then immediately followed by an end-protocol sequence an empty file will
be processed, which although in most environments should be benign, is not
recommended.
A sequence of files sent to the server that does not start with a begin-protocol
sequence (that is, it does not establish a connection) does not conform to the
tagged binary communications protocol. The data may be encoded according
to some other protocol, but the identity of that protocol is not part of the specification of the tagged binary communications protocol (see section 4.4,
Uses in a Language Switching Environment).
The end-protocol sequence not only signifies the end of the file data but also
indicates termination of the protocol (that is, it closes the connection between
host and server) and that a change of interpretation context may occur. The
protocol does not specify precisely the semantics of this last action. The

16

Adobe Serial and Parallel Communications Protocols Specification

(20 Nov 92)

common intended use is to switch seamlessly from one job language to

another. If further files are to be sent using the tagged binary communications
protocol, a new connection must be established by preceding the files with a
begin-protocol sequence. Typical usage of the tagged binary communications
protocol on a printer which switches languages seamlessly is discussed in
section 4.4 of this document.
4.2

Character Protocol

Table 4 lists the control characters that are treated as control functions rather
than as data by the communications driver when they are received from the
hardware:
Table 4 Special characters in the tagged binary communications protocol
ASCII
keyboard

ASCII
name

Value
(hex)

Control function

^A

SOH

0x01

Quote data character

^C

ETX

0x03

Generate an interrupt error

^D

EOT

0x04

End-of-file marker

^E

ENQ

0x05

(Reserved for future use)

^Q

DC1

0x11

XON in XON/XOFF flow control

^S

DC3

0x13

XOFF in XON/XOFF flow control

^T

DC4

0x14

Job status request

^[

ESC

0x1B

Start of end-protocol sequence

^\

FS

0x1C

(Reserved for future use)

To transmit these characters as data, they must be quoted. Quoting is done by

replacing the character with the two-character sequence ^A followed by the
character itself XORed with 0x40. For example, to send a byte with the hex
value 0x14, the two-byte sequence 0x01 0x54 is sent. This method of quoting
guarantees that whenever any of the nine special characters are received from
the hardware, the control function is intended, regardless of whether the preceding character is a ^A. The generation and processing of asynchronous control characters, therefore, may occur at a lower level than the generation and
consumption of the data stream. In particular, on a host machine, the ^A quoting convention may be implemented by an application program or job
spooler while XON/XOFF processing is performed independently by lower
level communications code.
All 8-bit values other than those of the nine special characters are transmitted
by simply sending the value.

4 Adobe Tagged Binary Communications Protocol

17

After a ^A is received, the next character received that is not one of the asynchronous special characters must be the result of XORing 0x40 with one of
the special characters or with ASCII character CR (0x0D). Receipt of any
other character is considered an error in the input. The sequence ^A M (ASCII
M is the result of XORing 0x40 with ASCII CR) indicates begin-protocol
(see table 5). Between the ^A and the XORed character, any number of the
asynchronous special characters may appear. Receipt of a synchronous special character between a ^A and the XORed character is considered an error.
The end-protocol sequence is the following 9 character ASCII sequence
appearing within the quotation marks: ESC%-12345X. Note that ESC is
the same as ^[. The receiver must parse for this sequence upon receipt of the
initial ESC character. If the full sequence is not received all prior characters
of the sequence received are passed through as data. If the full sequence is
received it becomes an end-protocol sequence and is treated accordingly.
Note that the end-protocol sequence may have asynchronous special characters interspersed, and this should not interfere with parsing the sequence.
Because of the requirement described above that the receiver must pass
through as data all sequences that begin with ESC except for the end-protocol
sequence, the sender may take one of two strategies in quoting the ESC character. It may quote all instances of ESC characters occurring as data. This is
especially simple but may expand the data more than is desired. Instead it
may choose to quote the ESC character only if it is followed by the remainder
of the end-protocol sequence, %-12345X. An intermediate strategy might
also be used. For example, the ESC character might be quoted only if it was
followed by an ASCII %.
When one of the special characters arrives unquoted, and it specifies no control function for the channel, the character is simply discarded. For example,
if XON or XOFF is received and XON/XOFF flow control is not in use, it is
discarded. If a ^A M sequence (begin-protocol indication) is received after the
initial one it will be discarded. The characters ^E and ^\ currently specify no
control functions. They are included among the characters that must be
quoted in case new control functions are added in the future.
Table 5 shows the actions taken or the data put in the input buffer for byte
sequences received from the hardware. Note that ^[ is the same as ESC.

18

Adobe Serial and Parallel Communications Protocols Specification

(20 Nov 92)

Table 5 Actions when bytes are received from hardware

Read from
hardware

Data in
input buffer

Type

Action

^C

Asynchronous

Generate interrupt

^D

End-of-file mark

Synchronous

^E

^Q

Asynchronous

XON (flow control)

^S

Asynchronous

XOFF (flow control)

^T

Asynchronous

Printer status request

^\

^[ %-12345X

TBCP end-protocol

^A A

^A

Synchronous

^A C

^C

Synchronous

^A D

^D

Synchronous

^A E

^E

Synchronous

^A M

TBCP begin-protocol

Synchronous

^A Q

^Q

Synchronous

^A S

^S

Synchronous

^A T

^T

Synchronous

^A \

^\

Synchronous

^A [

^[

Synchronous

^A <any other char>

Comm-Err mark

Synchronous

<any other char>

<char received>

Synchronous

4.3

Additional Points

The tagged binary communications protocol is entirely symmetric. The

control characters and the quoting conventions apply in both directions,
although some of the control characters have no defined meaning in the
server-to-host direction.
The data that passes through the protocol is raw. The intent is that both
sides can send pure binary data to each other. No data transformation
(such as mapping line-feed to carriage-return, line-feed or vice versa) is
carried out by the protocol. If these transformations are desired they must
be done at a level above the protocol itself.

4 Adobe Tagged Binary Communications Protocol

19

4.4

Uses in a Language Switching Environment

Although not part of the tagged binary communications protocol specification, certain printer software environments typify those which inspired the
tagged binary communications protocol. The tagged binary communications
protocol works well, and Adobe recommends its use, on printers that have a
language independent software component and arbitrates which language
will interpret an incoming job, such as Hewlett Packards Printer Job Language (PJL). In such an environment, when the arbitrating code invokes the
PostScript interpreter to handle a job using the tagged binary communications protocol, the job stream should have the form
^[ %-12345X

^A M

<PS job encoded in TBCP>

^[ %-12345X

In case the previous job left the software in an ill-defined state, the leading
end-protocol sequence ensures that control returns to the arbitrating code,
which is a well-defined state, before the next job begins. There are two possible cases of this:
1. The leading ^[ %-12345X is the end-protocol sequence for a previous
tagged binary communications protocol connection (not shown in the
example), and is processed as described in the protocol specification.
2. The leading ^[%-12345X is not part of a tagged binary communications
protocol connection but appears out of the blue. In this case, the protocol
specification doesn't really have anything to say about what ^[ %-12345X
means. As a practical matter, the ^[ %-12345X is received by the arbitrating code, which is responsible for handling it properly.
The ^A M causes the protocol to change from the default protocol (typically
Adobe standard protocol) to the tagged binary communications protocol.
Note this change occurs instantaneously without data loss or errors which
could occur when a data stream encoded for one protocol arrives at a different protocol. When placed in the job stream, ^A M should immediately precede the first byte of a PostScript language job. The end-protocol sequence is
overloaded in the sense that it causes the protocol to end, the job to end, and
control to return to the arbitrating code, which may invoke a different language or return to the PostScript interpreter to interpret the next job.
As discussed in section 1, use of the tagged binary communications protocol
(including the begin-protocol and end-protocol sequences) is appropriate
only when using channels that do not provide the same functions in other
ways. Control sequences such as ^A M and ^[ %-12345X are not a part of the
PostScript language and have no special meaning if encountered by the PostScript interpreter.

20

Adobe Serial and Parallel Communications Protocols Specification

(20 Nov 92)

Appendix: Changes Since

Earlier Versions Template

Changes since October 14, 1992

In all appropriate tables, corrected the definition of ^C to be that it generates an interrupt error.

Changes since February 14, 1992

In section 2.2, Enabling the Standard Protocol, added code to set the
setdevparams parameter Interpreter to a value of PostScript in the
Level 2 example.
Fixed minor typographical errors.

Changes since August 8, 1991

Converted Technical Note #5081, Adobe Binary Communications Protocol to this new specification, Adobe Serial and Parallel Communications
Protocols.
Added a section on the standard protocol.
Added a section on the tagged binary communications protocol.

Changes since May 4, 1991 version

Changed section Enabling Binary Communications Protocol. Applications should check a PPD file for existence of protocol, rather than how to
enable it.
Fixed minor typographical errors.

Changes since July 17, 1990 version

Added section Uses of this protocol.

21

 Added example C implementation.

Added section Enabling Binary Communications Protocol.

22

Appendix: Changes Since Earlier Versions Template

(20 Nov 92)

Index

A
Appletalk 14
ASCII base-85 6, 10
asychronous functions 9

14
end-protocol 16, 18
ESC 18
executive 14

H, I, J, K
hexadecimal 6

BCP. See binary communications

protocol
begin-protocol 16
binary communications protocol 5
character protocol 1013
definition of 8
enabling 14
using 8
binary data 6

C
carriage-return 6
control characters
A 6, 11, 13, 17, 19
back slash 6, 11, 12, 13, 17, 19
C 6, 11, 13, 17, 19
D 6, 11, 13, 17, 19
E 11, 12, 13, 17, 19
Q 6, 11, 13, 17, 19
S 6, 11, 13, 17, 19
T 6, 11, 13, 17, 19

D
differences
between protocols 13

E, F, G
enabling
binary communication protocol

L, M
line-feed 6

N
newline 6

O
octal 6

P
Password 15
PostScript Printer Description (PPD)
Files 8, 14
print and == 14
Printer Job Language (PJL) 20
Protocol 15

Q
quoting 11
sample implementation 12

R
read 14
readstring 14

23

S
SCSI 14
separate job 7
setdevparams 7, 15
setsoftwareiomode 7, 14

standard protocol 5
synchronous functions 9

T, U, V, W
tagged binary communications
protocol 5
character protocol 1719
definition of 16
TBCP. See tagged binary
communications protocol

X, Y, Z
XON/XOFF 6, 9, 11, 17
XORing 11, 18

24

Index

(20 Nov 92)