Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!swrinde!elroy.jpl.nasa.gov!decwrl!deccrl!news.crl.dec.com!jg From: jg@crl.dec.com (Jim Gettys) Newsgroups: comp.windows.news Subject: Re: NeWS "open" / "standard"? Message-ID: <1991Jun25.222909.2648@crl.dec.com> Date: 25 Jun 91 22:29:09 GMT References: <1991Jun21.165649@aquarius.sfu.ca> <1991Jun25.210259.19555@watcgl.waterloo.edu> Sender: news@crl.dec.com (USENET News System) Organization: DEC Cambridge Research Lab Lines: 28 I knew I'd regret any comments :-). Remember that the decisions that decided NeWS vs. X were being made some years ago. The point about colormaps/rasterops stuff was true when the industry was making up its mind... That it has been fixed since is irrelevant to the discussion about why people went the way they did. More of an issue (and to my knowledge, still unsolved in NeWS w.o. X) PostScript has very poor semantics at the boundary between filled areas; the easiest example where this hurts badly is ECAD, where parts of a drawing must be able to be updated without disturbing adjacent areas. You must have a tie-breaking rule, and not fill both edges of a filled area. This is irrelevant when painting on paper, as the issue never even surfaces. Moral is: pixels on screens aren't quite small enough to take the view that you can't notice each and every pixel touched; the fact is, you can. At 300 dpi, where PostScript was designed, you can (usually) get away with it. And I'm not going to get sucked into a more general discussion, particularly on general systems design philosophy; if you want more detail about our opinions, go read the special issue on X in Software Practice and Experience that just came out recently (I got my copy last week), though this doesn't go into specific criticisms of NeWS (afterall, it is papers about X, not critique of NeWS). You'll have lots of fun with it; it goes into all the things we really did wrong, given the general design point. Availability was the other major issue. - Jim Gettys