[fa.laser-lovers] Voluminous ASCII output from LaserWriter

laser-lovers@uw-beaver (08/18/85)

From: Neal Holtz <holtz%cascade.carleton.cdn%ubc.csnet@csnet-relay.arpa>

With a simple ascii terminal hooked to 9600 baud serial line of the Apple
LaserWriter, I executed the following (it prints 100 lines of 62 characters
on the terminal):

/lots 
  { {(abcdefghijklmnopqrstuvwxyz0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ\n) print flush}
    repeat
  } def

100 lots

There were quite a number of characters missing from the output --- perhaps
as high as 1 in 50.  They were missing from random places --- usually no more
than one per line.

Other tests have shown that the terminal has no problem receiving hundreds
of lines of this stuff without resorting to X-OFF/X-ON.  Therefore the
fault must lie with the LaserWriter.

I realize that this is not the primary intended purpose of the LaserWriter,
but still --- it probably shouldn't do that.

laser-lovers@uw-beaver (08/19/85)

From: Stephan v. Bechtolsheim <svb@Purdue.EDU>

I had the same problem.
Even worse: it screws up the program, in my case it just wouldn't
continue. And this is a pain, because I tried to figure out how much
VM space fonts and characters (user defined) need in order to fix up the
downloading of fonts (in dvi2ps). And so I was putting 'vmspace'
instructions everywhere.....

As far as I remember Adobe is on the net. I would appreciate if they
address the problem as soon as possible.

Stephan Bechtolsheim
Purdue, CS

----------

laser-lovers@uw-beaver (08/26/85)

From: adobe!taft@Glacier (Ed Taft)

I too have observed lost or garbled characters in serial output from
the LaserWriter, but only when certain terminals are connected. Other
terminals display perfect output indefinitely. During product test, I
performed extensive tests transferring large files from a Vax host to
a LaserWriter and back, with no errors occurring during the round trip.
(I should add that some terminals seem to have trouble with 9600 baud
output from Vaxes and Suns as well, though I don't know to what extent
this correlates with bad LaserWriter output.)

While I don't deny that there may be a problem, it's not a simple
matter of the LaserWriter discarding occasional characters. (By the
way, the LaserWriter does obey XON/XOFF flow control on the outgoing
channel as well as the incoming one.) There are two features of
LaserWriter serial output that may contribute to lost or garbled
characters:

  1. The actual baud rate used at the 9600 setting is not exactly 9600
baud but is off by slightly less than 1 percent. This means that the
timing of the last bit of a character is off by about 10 percent of one
bit time. It is possible that some terminals are bothered by this.

  2. The transmission dynamics are somewhat unusual. Though the
instantaneous baud rate is 9600, the character rate is limited to 500
per second (an effective rate equivalent to approximately 5500
baud); that is, there is a gap between every pair of characters.
Strictly speaking, there is nothing wrong with this; however, it is
sufficiently unusual that it may foul up the firmware in some terminals.

Since I don't fully understand the nature of the reported problem (if
indeed it is a LaserWriter problem at all), I'm afraid I can't offer a
workaround or a promise of an eventual solution. If anyone has any
additional information about this problem, I would appreciate their
sending it directly to me.

	Ed Taft
	Adobe Systems, Inc.