Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!usc!brutus.cs.uiuc.edu!psuvax1!rutgers!aramis.rutgers.edu!geneva.rutgers.edu!hedrick From: hedrick@geneva.rutgers.edu (Charles Hedrick) Newsgroups: comp.windows.news Subject: how are other sites using XNeWS? Message-ID: Date: 15 Feb 90 08:32:15 GMT Organization: Rutgers Univ., New Brunswick, N.J. Lines: 63 I'm curious how other sites are using (or planning to use) XNeWS. At the moment our users are split between SunView 1 and MIT's X (with twm). I think the majority are still using SunView. By the summer we'd like to move everybody to a system that can run X. Our applications are increasingly X-based. Originally I had hoped the XNeWS would be in good enough shape that we could move our user community to it. For those still on 4MB systems, I had hoped to have them use MIT's X with olwm, so that they'd have the same user interface as XNeWS. This plan no longer looks either desirable or feasible. XNeWS doesn't appear to be in good enough shape to give to large numbers of end users. Aside from bugs (like when you iconify a window, it vanishes, and various glitches in focus management), the software simply isn't polished enough for "real" users. If you compare the facilities of twm with XNeWS (the ability to define special function keys, tailor the form of the windows, the way icons are handled, etc.), you see the difference between a production quality system and a rough draft. Of course you can fix the NeWS side yourself in Postscript, but as far as I can tell, there's no way to tailor pswm. (It seems that if you're going to release a window manager that can't be tailored, the least you could do is include the source.) Since most of our real applications use X, missing functionality on the X side is a serious problem for us. At a bare minimum, we'd like the X side to implement the basic L keys, as the NeWS side does. People migrating from SunView 1 are not going to accept a system where you have to go to the title bar and click in order to bring a window to the top. The XView shelltool seems both slow and primitive compared to xterm. Furthermore, an essential part of the original plan was the use of olwm to make MIT's X look compatible with XNeWS. Olwm is still at the "rough draft" stage, and I don't sense much commitment to it from Sun. Currently I'm planning to reverse my original plan. Rather than having XNeWS as the basic system and using olwm to let MIT X look compatible, I'm planning to make MIT X with twm the basic system, and expect people to run XNeWS in a mode that is reasonably compatible. That is, I'm planning to suggest that people who use XNeWS plan to use primarily the X side, with the MIT tools (e.g. xterm and twm) instead of the Sun tools (e.g. shelltool and pswm). With a small amount of tinkering, you can get twm to have its mouse and L key responses sufficiently similar to XNeWS that it isn't too bad to have windows managed by NeWS and twm on the same screen. The question we're still discussing is whether XNeWS provides enough functionality in addition to MIT's X to make it worth using at all. I think it does. We have some users who have been using NeWS 1.1 to do Postscript previewing. It looks like XNeWS is if anything better than NeWS 1.1 in this area. Also, users with graphics accelerators get much better performance from XNeWS than MIT's X. (This is supposedly fixed in X11R4, but that is still too buggy for widespread use.) I have to say that as a NeWS proponent, I feel Sun isn't giving me much help. NeWS has been around an awfully long time to still be at the rough draft stage. Meanwhile, X has turned into quite a nice system. I suspect that it's going to be another year before XNeWS is really ready for widespread use, and by then all of our users will have migrated to pure X. I find it hard to believe that we're the only site in this situation. I'm worried that NeWS has missed its launch window. (I also have this feeling that after all the dust settles in the shootout between the highly-publicized user interface standards, the real winner is going to turn out to be the X stuff produced by the hacker community. I find myself increasingly pleased with twm.)