> http://eclipsed.net/~gr/April1.txt

I've got some comments on this; at least two of the people quoted in
that file are bit confused about things.  But those I will send
off-list, as they are not relevant to tech-userlevel.

> I'm not near a CUPS-managed printer right now, but CUPS is only
> really "broken" if their lpr(1) gets things wrong that ours doesn't.
> (Oh, wait, do they always guess, even when given an explicit argument
> about what to assume the file is?  That wouldn't be very neighborly
> of them...)

I don't know what CUPS does.  But our lpr, at least as of the manpage I
quoted the version number of (which was current as of my last update),
doesn't even _have_ explicit arguments for "this is text" or "this is
PostScript", which is my actual beef with it.  (I can deal with
guessing by default, but when the guess (a) is done with an
undocumented algorithm and (b) cannot be overridden, it crosses the
line into "broken" for me.)

> But lpr sort of *must* do Postscript guessing because the general
> case of feeding it a Postscript (ASCII text) file is that you want
> the output of its Postscript engine, and the general case of feeding
> it a random text file (also ASCII text) is that you want the contents
> of that text file spit out on a printed page.

Well, s/general/common/ and I'd agree with you.

I just consider it egregiously arrogant, unacceptable, broken,
whatever, for it to provide no way for the human to override that
guess, because it _will_ guess wrong at least sometimes.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B