Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!ucbvax!ucsfcgl!cca.ucsf.edu!wet!capslock From: capslock@wet.UUCP (Allen Crider) Newsgroups: comp.lang.postscript Subject: Re: display list vs. buffering (was complete PS description) Message-ID: <886@wet.UUCP> Date: 27 Dec 89 04:35:23 GMT References: <28@macuni.mqcc.mq.oz> <1651@intercon.com> <1751@tfd.UUCP> Reply-To: capslock@wet.UUCP (Allen Crider) Organization: Wetware Diversions, San Francisco Lines: 19 In article <1751@tfd.UUCP> geller@tfd.UUCP (David Geller) writes: > >Can anyone assist me in compiling a list of PostScript compatible >devices that use display lists and banding versus entire page buffering? >Thanks. > I believe the only machines that use buffering are Varityper, Autologic and devices using the Ultre*Setter as a recorder. I'm not sure if either of these implementations construct their page buffer from a display list or not. Building the buffer randomly might lead to excesive disk fragmentation, while building a display list, "imaging" the resulting bitmap to a large sequential file, and then passing out the file to the recorder is slow. Autologic has a wonderful solution to the problem of buffering to disk-- use DRAM instead. They sell an interpreter with 48 megabytes of RAM. In any case, I believe full page buffering to disk slows everything down except in cases where the "display list only" interpreter can't keep up with the recorder, such as printing halftones or very complex page descriptions. Full page buffering may be required for recorders that can't maintain pixel- level accuracy during a start-stop-start imaging cycle.