Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!apple!well!levine From: levine@well.sf.ca.us (Ron Levine) Newsgroups: comp.graphics Subject: Re: Practical Intro to PHIGS (new book) Message-ID: <24484@well.sf.ca.us> Date: 29 Apr 91 15:56:00 GMT References: <1991Apr19.175910.2922@Stardent.COM> <24373@well.sf.ca.us> Lines: 62 In article , Spencer Thomas takes issue with my paraphrase of a sentence in the book: >> In article <24373@well.sf.ca.us> levine@well.sf.ca.us (Ron Levine) writes: >> The second sentence overlooks the fact, that in general, NO >> workstations have the ability to modify parts of a displayed >> picture selectively, at least not correctly. On a CRT, you can ................ > >Not true. Many high-end workstations have a "built-in" display list. >Thus, as far as PHIGS is concerned, the image can be correctly >modified selectively. An old example is the E&S PS300 series (not >sure about their new workstation, as I don't have one.) Even on a >"bit-mapped" workstation, it may be possible to selectively update an >image, under certain conditions - for example, when adding a new >primitive to the end of a structure (depending on structure priority). Yes, on a calligraphic display, such as the PS300, you can selectively modify the display list and get a picture which is correct according to the display list. Now, I have no direct experience with PHIGS implementations for such devices, and so am not very certain about the relation between the device display list and the PHIGS posted structure network, but for several reasons it is hard for me to imagine that they would be identical. And if they are not, then to be sure that the device display list correctly represents the picture defined by the PHIGS posted structure network and workstation state list would require traversing the PHIGS structure network, i.e. "regeneration". (Because of the qualifier "in general" this sentence does not exclude the possibility that for some special picture changes or in some special situations, the display can be made current without traversal, as you point out). Now, this kind of traversal would not be so expensive as regeneration for a raster display because it does not involve scan conversion. Therefore, it may well be more acceptable to use deferral mode ASAP for such devices. Perhaps that is what the authors had in mind in their assumption of ASAP, to which I have objected. Of course, the calligraphic display would be not be a very good typical workstation model for a PHIGS/PHIGS PLUS discussion because it does not support the bulk of the features of the PHIGS and especially PHIGS PLUS system which are more oriented to raster. Do you, by the way, use PHIGS to drive your PS300? Even if I allowed that my sentence may be not completely true, it has not much bearing on my complaint about the paragraph in the book from which it is paraphrased, the paragraph which purports to define "implicit regeneration". This paragraph seems to create the impression that implicit regeneration is something to be used with devices such as laser printers (which would be a bad idea). And it fails to bring out the important fact that, under the book's unstated assumption of deferral mode ASAP, every incremental picture change will produce implicit regeneration, not selective modification, at least on all the most common interactive raster workstations. Ron Levine