Experiments in 8-bit printing

                               2025-06-24

In 2024-10, I read a discussion on Hacker News where people talked about
how back in the day, you could just pipe text to the line printer (lpr).
Someone said that this still works. So I tried it out.

I  have  a  Samsung  M2070 Series printer/scanner. I use Arch Linux with
CUPS. I have configured the printer with  system-config-printer  on  one
system  and  over localhost:631 on another. I have set the printer to be
my default printer. I use the "Generic PCL Laser Printer" driver.

I ran my experiments on 2024-10-12, 2025-06-18, and 2025-06-24. I  found
out that I could indeed just pipe text to "lpr -l". (The "-l" option ap-
parently  stands for "-o raw".) I got printouts in Courier with 64 lines
per A4 page and 78 visible  columns  per  line.  (For  lines  with  more
columns,  the CRLF at the end still worked, but the text was cut off, as
if there had been no characters between the 78th column and the CRLF.)

A backspace moved one column to the left. A carriage return moved to the
beginning of the line. This let me overstrike characters: I  could  pro-
duce  underlines  with  the  underscore and fake umlauts with the double
quote character. Overstriking a character with itself  did  not  produce
bold characters, though.

A  line feed moved down to the next line, but stayed in the same column.
To go to the beginning of the next line, I needed the classic CRLF  com-
bination.

A  form feed moved to the first line of the next page, but stayed in the
same column (like a line feed). To go to the beginning of the first line
of the next page, I needed to replace the last  CRLF  combination  of  a
page with a "CRFF" combination. (The pattern "CRLF, form feed, CRLF", as
used  in  various  RFCs,  moved  to the second line of the next page in-
stead.)

A horizontal tab moved to the next column just after a multiple of 8 (as
usual). The veritcal tab had no effect.

Apart from the mentioned characters  (backspace,  horizontal  tab,  line
feed,  vertical tab, form feed, carriage return), the (hexadecimal) byte
values 00 (null), 07 (bell), 0E (shift out), 0F (shift in), and 1B  (es-
cape)  produced  no  graphical  output.  I  have  not tested whether the
printer would react to escape sequences with the escape byte.

All other byte values produced the corresponding  characters  from  Code
Page 437 (even 7F, "delete", that showed a "little house"). So I did not
have  to  use  overstriking  to produce umlauts, and I could even do box
drawing and some math.