TENEX FTP PROBLEM
A warning about line ending problems in Tenex.
Braden, R., "TENEX FTP PROBLEM", RFC 571, DOI 10.17487/RFC0571, November 1973, <https://www.rfc-editor.org/info/rfc571>.
@misc{rfc571,
series = {Request for Comments},
number = 571,
howpublished = {RFC 571},
publisher = {RFC Editor},
doi = {10.17487/RFC0571},
url = {https://www.rfc-editor.org/info/rfc571},
author = {R. Braden},
title = {{TENEX FTP PROBLEM}},
pagetotal = 1,
year = 1973,
month = nov,
} RFC text
Network Working Group Robert T. Braden
Request for Comments: 571 UCLA/CCN
NIC #18974 November 15, 1973
TENEX FTP PROBLEM
Tenex users of FTP should beware of a problem in the current Tenex
implementation which is likely to cause incorrect results when
transferring files to a non-Tenex site. Some Tenex programs create
an ASCII US (Unit Separator) character as an end-of-line indicator,
instead of CR LF. The current Tenex FTP user implementation passes
these US characters over the Network instead of translating them to
CR LF as required by FTP protocol. If the recipient is not a Tenex,
he will not do anything special with these US characters and the line
spacing will be lost. In the case of sending to CCN, we will treat it
as a very long logical record and fold it into successive 360
records.
A circumvention is to pass the Tenex file through some Tenex program
that does the necessary conversion before sending it over the
Network. We know that TECO will remove US's; there are probably
other means. Loading and saving the file with TECO will replace all
US's with CR LF sequences.
[ This RFC was put into machine readable form for entry ]
[ into the online RFC archives by Katsunori Tanaka 4/99 ]
AI ELI5
On this page
§1The Problem
Tenex systems use an ASCII Unit Separator character to mark the end of lines in some files, instead of the standard carriage return and line feed. When users sent these files over FTP to non-Tenex computers, the FTP program did not translate these characters to the protocol-required carriage return and line feed pairs. Non-Tenex recipients would receive the files with no proper line breaks, making them unreadable or scrambled.
Tenex computers use a special character that other computers don't understand to end lines. When Tenex sent files through the network, it forgot to translate that character to what everyone else uses.
§2How It Affects File Transfer
When a Tenex file containing Unit Separator line endings is sent to a receiving system like CCN, the receiver treats the entire file as a very long logical record since it does not recognize the Unit Separators. On systems like the IBM 360, the file gets folded into successive records with incorrect groupings, destroying the file's structure.
The receiving computer thinks the entire file is one super-long line because it doesn't see where the lines should break.
§3The Workaround
Users can convert the file before sending it. The TECO editor, when used to load and save a Tenex file, automatically replaces all Unit Separator characters with proper carriage return and line feed sequences. Other Tenex programs that perform similar translation may also work for this purpose.
Loading the file into an editor and saving it again fixes the line ending problem.