[Oberon] RE: Unix line delimiters

W B Hacker wbh at conducive.org
Mon Oct 30 00:29:21 MET 2006

Chris Burrows wrote:
>> Just a pedantic point, Ken and Dennis used KSR-33 to invent 
>> Unix, along with some model 37's and a lot of glass ttys 
>> shortly after.
>> Here's a link to an image of the first pdp-11 model 20 they 
>> shared with the patent department at Murray Hill.  I don't 
>> know why they used just newlines as line deliminters, but I 
>> don't think it wasn't because of the output devices.
> I agree. Just the opposite in fact - they deliberately designed text files
> in a way that was independent of any particular input / output device. Why
> use a pair of characters to denote a line delimiter when one was sufficient?
> The following extract from 'A Retrospective' by Dennis Richie, The Bell
> System Technical Journal, July / August 1978, Vol 57, No 6 Part 2. 
> "A text file, for example, is stored as a sequence of characters with
> new-line characters to delimit lines.  This form of storage is not only
> efficient in space when compared with fixed-length records, or even records
> described by character counts, but is also the most convenient form of
> storage for the vast majority of text-processing programs, which almost
> invariably deal with character streams. Most important of all, however, is
> the fact that there is only one representation of text files. One of the
> most valuable characteristics of UNIX is the degree to which separate
> programs interact in useful ways; this interaction would be seriously
> impaired if there were a variety of representations on the same
> information."
> Although this might seem obvious today, it was quite a departure from common
> practice at the time. On other systems the applications (not systems)
> programmer had to meticulously format text in different ways, depending on
> where it was coming from or going to. 
> --
> Chris Burrows
> CFB Software
> http://www.cfbsoftware.com/gpcp
> --
> Oberon at lists.inf.ethz.ch mailing list for ETH Oberon and related systems
> https://lists.inf.ethz.ch/mailman/listinfo/oberon

Good point, and it solves the mystery, as other info reminds us that specific
bits were (once) specific *wires*, not even patterns of wires, so the above was
a sensible departure from a very mixed bag of legacy hardware dependency.

Pity more printer makers did not add 'firmware' support for recognition of it
early-on - drivers for at least 'raw' text could have certainly been more
universal if more had done so.


NB:  When I pitched the old OKI-82A a month or so ago, there were two
momentary-on toggle switches protruding from the front apron. Dim memory
reminded me that its feeble mind *scanned* the DIP switches at power-on, then
retained what it had scanned in scratchpad ROM.

These were wired across the DIP switch, and ISTR that one switch selected
parallel/serial, the other CR-LF or just LF if/as/when held down while the power
switch was cycled. Kludges! can't live without 'em, but now we do 'em in code...

More information about the Oberon mailing list