This will address some recent printing quirks reported here, mostly
related to 5.5 and HP's network printing, but some issues apply to 5.0
(and earlier) releases and serial printers.  There have been several
quirks reported with respect to "unexpected results" but I'm going to
try to define the underlying causes in an effort to make some
predictable sense of it all.

A spool file holds the FOPEN record, optional "setup" information, and
the actual user data to print.  For traditional lineprinter files there
is no setup info excluding some rare TTYPE or TVFC files, or the
follow-up PENV files for laser printers (2680/2688/etc).  These were all
ironed out over the years so that "setup" data is clearly distinct from
the "user" data -- for these special files the file system generally
writes the env= file name to the spoolfile and the device driver deals
with the files at the time of output.

The "blank page" phenomenon relates to the "device driver" being
enhanced to detect an %61 carriage control in the first user data
record.  If one was found it was ignored since you are already at top of
page.  This depends on the driver detecting "setup info" versus "user
data".

Until 5.5 we had several setup file types:
* TTYPE file
* TVFC file
* PENV file

All the above resulted in spooler records of FDEVICECONTROL or similar
identifiable record types.  Extending to 3rd-party environment we have:
* PENV file from Unison's Formation,
* TTYPE files in RAC's ESPUL valid only for networked printers
controlled
  by their product, containing raw PCL (not a true TTYPE file).

In the above, PENV files in Reflection are caught by an AIF and the
setup PCL is inserted into the spool file as FWRITEs with %320 cctl
codes.  In these cases, the device driver (output spool process) can't
always recognize the start of user data.

The 5.5 release extends ENV= to include any user file with fcode < 1024
which can contain raw PCL setup data (such as the supplied network
printing environment files on 5.5, e.g., [log in to unmask]) and they are
also inserted into the spool file as FWRITEs with %320 cctl codes.

MPE network printing doesn't understand TTYPE files at all, and doesn't
understand "setup" info of any type listed above, so the suppression of
an initial form feed (%61) fails as noted by many.  Further, the serial
drivers don't understand the 5.5 network printing environments
([log in to unmask]) and can also generate blank pages.

The common denominator is identifying the "start of user data".  This
isn't currently clearly (100% accuracy) marked or identified.  The
network printer drivers can be tweaked, but retrofitting this into the
serial drivers is a nasty proposition.

No easy answers, just be aware of the quirks.

Jeff Kell <[log in to unmask]>