IBM's approach for VM (dunno about MVS) is/was to store
PostScript be ASCII, not "plain text" (which would be EBCDIC
with the filesystem handling line breaks out-of-band), but ASCII.
Perhaps this stems from an AR interpretation of PostScript's definition.
(if the doc says "ASCII", they'll interpret it literally) This is a
pain because users have to pass the stuff around as binary. Also,
other vendors don't take this route, but treat PostScript as plain text.
Personally, I'm on the side of these "other vendors".
It makes PostScript a *lot* easier to handle.
Yes, the binary inclusions in Level 2 and beyond
break this simplicity. Drat Adobe for that! Otherwise ...
We can handle PostScript on VM without error. We have had
Apple LaserWriters, NeXT printers, and Xerox printers available to
our VM system for years and never had a problem ... as long as
PostScript is "plain text". Even if PostScript is "binary"
it still works, it's just really annoying and -prone to-
corruption, though not always corrupted, in transit.
I think that plain text formats are a really desireable goal.
Further, I think it's fair to force the issue, allowing certain clients,
servers, applications to break if they fail to handle "plain text".
Rick Troth <troth(_at_)rice(_dot_)edu>, Rice University, Information Systems