Protocol Standards
Volume Number: 2
Issue Number: 1
Column Tag: Mac Meets Ma Bell
By Bruce Lieberman, Communications Engineer, MacTutor
Contributing Editor
Mac Binary Standard
A file transfer protocol specific to the Macintosh that originated within the MAUG
(Micronet Apple Users Group) on Compuserve has gained widespread acceptance and is
becoming a defacto standard, and is a primary feature of the new MacTerminal 2.0 being
distributed this month.
The "MacBinary" protocol (everything earns a label nowadays!) allows
transmittal of the Data and Resource Forks within a Macintosh application and it also
transfers the Finder information on that particular file ( icons, creation date, size,
"About" info, etc...). This standard is now being utilized by virtually every Macintosh
communications software developer that I have contacted. Look forward to new versions
of MacLine, inTouch and Smartcom that will support the "MacBinary" XModem standard
by the time this article is in print.
Here is a reprint of the technical aspects of the MacBinary protocol, courtesy of
Dennis Brothers ("MacTep" fame) :
The binary format described is independent of the communication protocol used
to accomplish the transfer, and assumes only that an 8-bit transparent transfer can be
accomplished. Such protocols as Christensen (usually referred to as XMODEM or
MODEM7), Kermit, and CompuServe A or B meet this requirement. Because of the
proposed standard's MacTerminal/XMODEM heritage, there is a requirement that the
transmitted data be padded (with nulls) to a 128-byte boundary at certain points, but
this in no way implies that a block-oriented protocol must be used. The basic format
proposed is identical to that used by MacTerminal, by Mike Boich and Martin Haeberli,
and can be used with a version of MacTerminal that has had a patch applied to
"normalize" its implementation of the XMODEM protocol.
In brief, the binary format consists of a 128-byte header containing all the
information necessary to reproduce the document's directory entry on the receiving
Macintosh; followed by the document's Data Fork (if it has one), padded with nulls to a
multiple of 128 bytes (if necessary); followed by the document's Resource Fork (again,
padded if necessary). The lengths of these forks (either or both of which may be zero)
are contained in the header.
The format of the 128-byte header is as follows (offsets and lengths are given in
decimal):
Offset Length Contents
000 1 Zero. This is a "version" byte.
001 1 Length of filename.
002 63 Filename (only "length" bytes are significant).
(the following 16 bytes are a standard Finder Info record)
065 4 File type.
069 4 File creator.
073 1 Finder flags:
Bit 7 - Locked.
Bit 6 - Invisible.
Bit 5 - Bundle.
Bit 4 - System.
Bit 3 - Bozo.
Bit 2 - Busy.
Bit 1 - Changed.
Bit 0 - Inited.
074 1 Zero.
075 2 File's vertical position within its window.
077 2 File's horizontal position within its window.
079 2 File's window or folder ID.
(End of Finder Info)
081 1 "Protected" flag (in low order bit).
082 1 Zero.
083 4 Data Fork length (bytes, zero if no Data Fork).
087 4 Resource Fork lenth (bytes, zero if no R.F.).
091 4 File's creation date.
095 4 File's "last modified" date.
099 27 Zero fill (reserved for expansion of standard).
126 2 Reserved for computer type and OS ID
(this field will be zero for the current Macintosh).
Note that it is the responsibility of the receiving terminal program to resolve
file name conflicts, generally by somehow modifying the name of received file if there
already exists a file with the original name on the target volume.
Note also that, while the original window or folder ID and position may be
transmitted, the receiving terminal program would not normally set these items for
the received file, but would instead accept the values that the File Manager assigns
when it creates the new file.
It is suggested that Macintosh terminal programs implement two modes of
protocol transfer: text and document. Text mode is used for unformatted files of type
TEXT (with only a data fork), and document mode (using the binary format proposed
here) is used for all other files. Document mode may also be used on text files, of
course, if it is desired to preserve such things as the file's creator ID or creation date.
The intent of text mode is to provide compatibility with text files on other
systems. Toward that end, it is recommended that a linefeed be inserted after each
return character as the text file is transmitted, and that, in the case of block-oriented
protocols, the last block be explicitely padded with nulls if the text does not end on a
block boundary. When receiving in text mode, linefeeds and trailing nulls should be
stripped. If a CTRL-Z (Hex 1A) character is received following all other text (and
before any null padding), it should also be stripped (Ctrl-Z is a common text
terminator on CP/M and some other systems). Note that the above discussion applies
only to text files transferred under some protocol, where an exact image of the
transmitted data will be stored in a file on the remote system.
When receiving a file via a protocol, a terminal program may distinguish
between text and document modes by examining bytes 0, 74, and 82 of the first 128
bytes received. If they are each zero (and at least 128 bytes have been received), then
it is a safe assumption that a MacBinary-formatted document is being received.
Terminal programs implementing possible future versions of the proposed standard
would, of course, accept an appropriate set of version numbers in byte 0. Note also that
compatible extensions of Version 0 of the proposed standard are possible (one such is
suggested below) that involve transmission of additional information following the
information described here. For this reason, a terminal program should be
implemented so as to perform the proper receive procedures for all data sent to it, but
to ignore any data that it does not know what to do with.
Since a text-mode document does not contain a file name, it is suggested that
when text-mode is detected, a file be opened under a dummy name on the receiving
Macintosh, the text written to that file, and the file renamed to a name selected by the
user (via an SFPutFile box) after the reception is completed. This will avoid problems
caused by indeterminate delays for name selection at the beginning of the file transfer.
It is desirable to allow the user to specify the destination volume in advance of
the actual start of transfer for either type of transfer. Two methods are suggested for
this: provide a "Select Destination Volume" menu selection, presumably in the menu
containing the "Receive File" selection; or query the user immediately after the
"Receive File" menu selection is made. Either or both of these methods could be
implemented in a given terminal program - the independent "Select Receive Volume
method is particularly desirable if the ESC-a/ESC-b automatic receive facility (see
below) is implemented. The volume selection procedure should provide the same
functions as the volume selection portion of the SFGetFile and SFPutFile dialog boxes.
First Proposed Extension
It is proposed that the binary format described above be extended to allow the
transmission of descriptive text with a Macintosh document (specifically, the
descriptive text from the Finder's "Get Info" box for the file being transferred). This is
to be accomplished in a transparent manner by assigning bytes 99 and 100 of the
header described above to be used to hold the length of the descriptive text (or zero, if
there is none). The descriptive text, if any, will begin on the 128-boundary
immediately following the Resource Fork, if present, else the Data Fork, if present,
else immediately following the header if neither fork is present. It is hoped that
methods for reading and setting a file's "Get Info" text will be made public at some point.
Notes on MacTerminal's XMODEM implementation, and a proposed extension:
Familiarity with the Christensen (XMODEM) protocol is assumed in the following
discussion.
When doing "Mac-to-Mac" transfers, using the binary format described above,
unmodified MacTerminal does not use a true XMODEM protocol, and is therefore
incompatible with most non-Mac systems. The differences lie in two specifics: the
transmitting Macintosh initiates the transfer by sending the the two characters ESCAPE
(hex 1B) and "a" (hex 61); the receiving Macintosh is expected to reply with the
character ACK (hex 06). The transfer then proceeds using normal XMODEM
procedures, except that each of the header and the two forks (if present) is treated as a
separate XMODEM transfer, with the transmitting Macintosh waiting for a NAK (hex
15), then sending the blocks of that phase (beginning with a block number of one), then
sending EOT (hex 04) and waiting for an ACK (hex 06) from the receiving Macintosh.
It is proposed that a modified procedure be accepted as a standard, to be
implemented instead of or in addition to the above-described MacTerminal
"Mac-to-Mac" protocol in complying terminal programs. The modified procedure,
which is compatible with standard XMODEM implementations, functions as follows: The
transmitting Macintosh sends the two characters ESCAPE (hex 1B) and "b" (hex 62).
The receiving Macintosh may optionally reply with ACK (hex 06) (this is allowed so as
to have minimum impact on existing MacTerminal-compatible implementations). The
transmitting Macintosh then awaits receipt of a NAK (hex 15) (or optionally a "C
(hex 43), if the receiving terminal program supports CRC checking), following which
a single, normal XMODEM transfer occurs. The transfer may be in text mode or
document mode, will begin with block number one, and block numbers will increment
uniformly (modulo 256) throughout the transfer.
It is expected that a patch to MacTerminal making it compatible with the above
proposed procedure will be available in the near future.
Responses to proposals
Please address comments or questions on the above proposals to:
Dennis F. Brothers
197 Old Connecticut Path
Wayland, MA 01778
617-358-2863
CompuServe: 70065,172
Delphi: DBROTHERS
MCI Mail: DBROTHERS
New Communications Programs
Presently there are a number of new programs that incorporate MacBinary,
including Red Ryder 6.2, FreeTerm 1.7, Telescape & VersaTerm. The proliferation of
new Communications software for the Mac is truly amazing ! Most of these are very
functional and flexible pieces of software. Most incorporate command languages for
unattended functions, multiple terminal emulations, etc. Others have nice little
features that make you wonder how you ever got along without them. Like the "view
file" function from within the "inTouch" program that allows you to review a file
(ascii or vidtex) before it is sent or after it is captured; or the pull down menu in
VersaTerm that allows you to store 8 of your most often called numbers and just CALL
them without going through all kinds of hassle to pull up a new file. I guess a dream
program would synthesize the best features of all of these programs, but right now it
isn't as hard to accept what is available as it was a little earlier in the life of Mac
communications.
Our next few areas of discussion deal with some of the latest developments in
telecommunications : 2400 baud modems and some of the new protocols for higher
speed data transmission.
"V.22 bis" is the defacto international standard for 2400 baud full duplex
asynchronous data communications. Currently modems from Racal Vadic, Hayes, U.S.
Robotics, Novation, AT&T and Microcom support this standard. It is a creation of the
CCITT and as has earned the stamp of approval from Bell. The CCITT is actually tied into
the United Nations and it produces recommendations for communications standards in
the "X" and "V" categories. The "X" series (such as the X.25 standard) applies to packet
switched networks such as Telenet, Tymnet and Uninet. The "V" series applies to
ordinary switched telephone lines; the 1200 baud standard is V.22 and the 2400 baud
standard is now V.22 "bis" (which is French for secondary, kinda like the European
version of revisionitis). In the USA there are some slight modifications made to V.22bis
to accomodate some of our own domestic quirks, ie: switching between Bell 212a and
V.22bis (rather than V.22); and with call-waiting as a widespread feature they have
changed the loss of carrier disconnect delay from 40-65 milliseconds to 600 or more
milliseconds.
Telephone circuits normally have a frequency bandwidth that is limited to around
300-3000 cycles per second (also known as cps, Hertz.). The carrier signal that the
modem modulates must fit within this bandwidth, to exceed the capacity of this
frequency " window" we resort to another type of modulation, PHASE encoding. Going
beyond representing 1 and 0 with two different frequencies, we now switch between
multiple states of phase (for 2400 baud there are 16 states of phase, the current Bell
212a standard uses 4 states of phase) and divide our information within each state. In
V.22bis things like "Quadbits" (four consecutive bits) come into play. The first two
bits in a quadbit define the change of quadrant (0, 90, 180, 270 degrees of phase)
relative to the quadrant occupied by the preceding signal element. The last two bits of
the quadbit define one of 4 signalling elements related to the new quadrant.
Here is a chart of common baud rates and their appropriate carrier states :
Bit Rate (bps) Baud Rate Carrier States
300 300 2
1200 600 4
2400 600 16
4800 1200 16
9600 2400 32
Notice how none of the baud rates exceeds the frequency bandwidth limit.
Remember that this is for FULL DUPLEX transmission. There is a "High" and "Low
channel for simultaneous transmission and reception. Each channel operating at HALF
the overall speed of the unit. At baud rates above 2400 line equalization is necessary to
combat the poor quality of phone lines. The next step beyond V.22bis will be V.32.
V.32 will allow full duplex asynchronous transmissions of up to 9600 baud.
Telebit of Cupertino (where else ?) is entering into an agreement with a number of
heavy telecommunications companies to market their new 10K-baud modem called the
"FastLink" (you'll probably see this under the "IRMA" label first). While it is not
presently V.32 compatible, they intend to submit what they feel is a superior standard
to the CCITT and see if they can have their method adopted in lieu of the current V.32
(V.32 "bis" maybe ?). Concord Data Systems of Waltham, Mass. (among others) intends
to release a V.32 standard modem by the end of the year. Both the CDS and the Telebit
technologies fall into the $2,000 per modem bracket. For a LOT of corporations, these
technologies will pay for themsleves VERY quickly (for FAX transmissions among
many other things). And WE are the folks that will inherit this technology as the prices
fall. Come on now, admit it ..... wouldn't you just LOVE to have one ??
NOW, onto the protocols that will be necessary to support these high speeds .....
the following two Protocols are in contention. X.PC and MNP. X.PC is being promoted
by Tymnet (a major packet switched network) and MNP is being promoted by
Microcom (a major modem manufacturer). X.PC is being placed in the public domain and
MNP requires a $2,500 licensing fee.
X.PC is a subset of X.25, which is a CCITT standard for packet switched networks
such as Tymnet. It is a full duplex asynchronous error correcting protocol that
supports up to 15 logical channels. X.PC is a Link Level protocol. The link level is one
of the seven levels defined in the ISO / OSI (International Standards Organization Open
Systems Interconnect) model. It maintains transparent full time automatic error
correction during any communications session between two or more machines (right
now this is limited to multiple PC's and/or machines on a Network that supports the
X.PC protocol). X.PC protocol (as does MNP) groups bits of data into packets, does a
high level statistical analysis of the packets (16 bit CRC) and transmits the results to
the receiving end. The X.PC on the receiving end will also perform a statistical analysis
of the received packet and if there has been an error in transmission it will request
that the originator retransmit the packet. When connected to a Host Network that
supports X.PC (such as Tymnet or Telenet) you can execute up to 15 different
concurrent sessions (simultaneously talk to 15 different Host Computers) with just
one phone link. Each packet of data is encoded with both an address as to which computer
system it is destined for and an address as to which system the packet originated from.