[rfc-i] Neanderthals did not wear suits

Bob Braden braden at isi.edu
Mon Nov 29 14:37:09 PST 2010





>>> RFC 60 with its 'modern' 'St. of th. Memo' section while other RFCs of
>>> that period
>>> do NOT contain ANY 'St. of th. Memo' section? Isn't it modifying the
>>> RFC?
>>> A link: http://www.rfc-editor.org/rfc/rfc60.txt
>>
>> It might; if it does, it's a bug that happened when the original RFC
>> was transformed into machine-readable form.
> According to the general statement, RFCs are not to be changed. But
> original RFC did not contain
> the section and, IMO, we have every right to ask the RFC Editor to make
> a correction.
>>

Indeed, the RFC Editor apparently did add "standard" boilerplate to RFC 
60, when it was retyped a few years ago under the "rfcs online" project 
of the RFC Editor. RFC 60 was written in 1970 when the RFCs were 
informal (or sometimes formal) notes passed around in the ARPAnet 
research community. The originals therefore displayed a great variance 
in typographic style, and forget about boilerplate! Those early RFCs 
were created on typewriters (a piece of apparatus you many never have 
seen) ;-) with hand-drawn figures, and then photo-copied and distributed 
by snail mail to the researchers building the ARPAnet. Until the rfc 
online project, they never existed in machine readable form and there 
was no standard format. Some used complex diagrams and arcane symbology 
(say, a letter within a circle). If you examine the series of early 
documents, you will see Jon Postel's influence gradually imposing an 
evolving style for document headers and boilerplate.

Over the past 15 years the RFC Editor has carried out the "RFC online" 
project to manually type in the first 500 or so RFCs. This was done by 
volunteers, so there was considerable variance in the style that was 
used. Why was it necessary to retype the early RFCs? A year or so after 
# 60, FTP was defined and operational, and distribution of RFCs by FTP 
made online copies available. But somewhere around RFC 600, a DEC tape 
containing the archive was lost, so all RFCs up to that point had to be 
recreated by the RFCs online project.

So, the early RFCs had great variance in style, and the rfc online 
process added more variance. There is nothing for the current RFC Editor 
to "correct" about RFC 60.  In fact, I just examined an original paper 
copy of RFC 60 and found that it was apparently typed on MIT's TX-2 
computer, and that it was entirely in *upper case*! The RFC online 
typist apparently supplied appropriate case, a modern header, and a 
Status of this Memo section. The typist also took the first paragraph of 
the original and made it into an Abstract section.  Other typists made 
other choices. The RFC online project had to make judgments on how much 
to modify the typography to preserve the contents but make bring more 
continuity to the series. I believe (hope) that the content of RFC 60 
was unchanged in the process of bring it online.

The early RFCs are very interesting for historians of science, revealing 
how people struggled with the most basic notions of naming, layering, 
flow control, connection state, buffering, etc. But quite obviously 
those early RFCs are so far removed from today's Internet technology 
that it makes no sense to try to apply modern IETF conventions to these 
researcher documents from 40 year ago.

I strongly recommend RFC 61, by the way; a seminal work on inter-process 
communication (that I think was published in a journal subsequently). I 
myself found it an revelation at the time, and it still may be.

Bob Braden




More information about the rfc-interest mailing list