Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!mcnc!ece-csc!ncrcae!ncr-sd!hp-sdd!hplabs!sdcrdcf!burdvax!bpa!cbmvax!vu-vlsi!hvrunix!swatsun!greenber
From: greenber@swatsun (Peter Greenberg)
Newsgroups: comp.text
Subject: Re: impress, postscript printers
Message-ID: <1121@babylon.UUCP>
Date: Sat, 9-May-87 06:54:13 EDT
Article-I.D.: babylon.1121
Posted: Sat May  9 06:54:13 1987
Date-Received: Sun, 10-May-87 09:24:18 EDT
References: <6310@hi.UUCP>
Organization: Swarthmore College, Swarthmore PA
Lines: 26
Keywords: impress,postscript
Summary: Are PostScript printers ever bound by communications rate?

In article <6310@hi.UUCP>, cyrus@hi.UUCP (Tait Cyrus) writes:
> When is someone going to make a postscript printer that sits on
> the ethernet?  As far as I know, current postscript printers
> sit on a RS232 port.  This is slowwwwwwwww. 

[ complaint follows about slow bitmap printing ]

I think that the time problem is primarily in PostScript, which because it is
a full programming language interpreter is not blindingly fast. The
problem with the bitmap is that PostScript has to figure out how to half-tone
your image, amongst other stuff, like positioning. A one megabyte bitmap will
take at most 2 minutes to get there continuously at 9600 baud, but the whole 
process will take about 5 or 10 minutes. Evidently, computaion takes more time
than communication. 

Also, I think that PS reads tokens off 
its input stream as it goes along, rather than waiting for the whole file to 
come down first (it uses some protocol to tell the host that its buffers are
full or not, like XON-XOFF). So, I think it is CPU bound and not I/O bound, but 
I wish some PS or LaserWriter guru could enlighten me and the rest of this 
board.
-- 
Peter Greenberg, Swarthmore College, Swarthmore, PA 19081 
AT&T:(215) 328-8384 or 8610
UUCP: ...{{seismo | inhp4}!bpa | {sun | rutgers}!liberty}!swatsun!greenber
ARPA: swatsun!greenber@bpa.BELL-ATL.COM