arnold@freezer.it.udel.edu (frederick arnold) (10/22/90)
Not to clutter up the NeXT group with posts concerning its technical and moral antithesis, X11, however there are two good options to replace that hulking obscenity. 1) Ideal (to the next community) follow the lead of IBM and port NeXTstep. 2) Follow Silicon Graphics lead with the use of the Sony NeWS system. On the iris in our lab this allows a windowing environment that is responsive, easy to program, and extremely customizable. (it better be, as it's running something very similar to postscript(so say the manuals), and can use postscript fonts (so say I, having done it.)) Finally, a glowing ember to the X11R? supporters out there. X is very much an example of what Petr Beckman in his book, "A History of Pi", refers to when he says that what distinguishes the engineering of a healthy society from the Engineering of a Pathological society (in his example ancient Rome and Stalinist Sov Union), is that the engineering of healthy societies gets better, while that of the gangster states becomes bigger and more grandiose. Yes, X will do the job, but it's a performance hog, it bogs down networks, is difficult to program in, and all around a Nasty Thing, a prime product of an unholy alliance between the DoD and MIT. It's time to realize that just because it's popular, doesn't mean it's worth while; it just had better marketing. It's time the producers of better solutions banded together, enforced their own standards, and did away with that monstrosity. In my personal opinion, option 1 above is the ideal path, and I look forward to the day we have it running on the RS/6000 in the office. (now if I can only convince my advisor to buy a cube, and not another Mac) ******************************************************************************* * Wenn die wirkliche Brandstifter Frederick P. Arnold, Jr* * w\"aren, du meinst, die h\"atten Dept of Chemistry, * * keine Streichh\"olzer? University of Delaware * * -Gottlieb Biedermann * * Biedermann und die Brandstifter ARNOLD@FREEZER.IT.UDEL.EDU* *******************************************************************************
ken@dali.gatech.edu (Ken Seefried iii) (10/22/90)
In article <34207@nigel.ee.udel.edu> arnold@freezer.it.udel.edu (frederick arnold) writes: > >2) Follow Silicon Graphics lead with the use of the Sony NeWS system. >On the iris in our lab this allows a windowing environment that is >responsive, easy to program, and extremely customizable. (it better >be, as it's running something very similar to postscript(so say the >manuals), and can use postscript fonts (so say I, having done it.)) > Just a nit: It's the Sun Microsystem NeWS system. Sony News (no StUdLy CaPs) is a series of 680x0 and MIPS based workstations. [ lots of psuedo-intellectual X bashing deleted ] -- ken seefried iii "A snear, a snarl, a whip that ken@dali.gatech.edu stings...these are a few of my favorite things..."
bchen@nntp-server.caltech.edu.UUCP (Bing-Qing Chen) (10/23/90)
> Finally, that brings us to NeXTstep. It is a good environment, except > that it doesn't run over the network. It would be great if I could > run Mathematica on one workstation and have another workstation display > the output. (Why would I want to do this? Suppose somebody else in > the building has a NeXT with 32 MB RAM, while mine has only 8 MB. I > don't want to work in his office, but I'd like to run M on his machine.) > I do this all the time with Mathematica on our HP workstations which use > X. Sure you can do this. Just turn on your local NeXT's public window server flag and run the application from remote NeXT with a flag -NXHost <hostname>. - Bing Chen bchen@pooh.caltech.edu
wisdom@pico.cs.umn.edu (Scott Wisdom) (10/23/90)
tim writes: >Neither NeXTstep nor NeWS let you run over the network. This is a very >important feature of X that very few other window systems ever had. ... >Finally, that brings us to NeXTstep. It is a good environment, except >that it doesn't run over the network. I usually don't get involved in these trivial little battles about who's best, because both NextStep and X windows both have their strong points (NextStep is good, and X is free). BUT: I think you should know that NextStep DOES run over a network quite well, just like X. I also have run mathematica the same way. I developed a few X windows applications, before NextStep came along. Now I'll develop X only if: 1. NextStep fades away *and* 2. X becomes TRULY object oriented with appropriate toolkits MUCH more advanced and integrated than anything currently available (something on the same level as NeXTStep). >I might like the NeXTstation even more if it ran *just* X (skip NeXTstep >altogether) Hmmm. You didn't know NextStep ran over a network? You might like the NeXTstation better if it *just* ran X? Just how well do you really understand NeXTStep? I bet money you don't. -Scott Wisdom wisdom@heckle.cs.umn.edu <-- Send e-mail here, bitte Coda Music Software Hacker
korp@atlantis.ees.anl.gov (Peter Korp) (10/23/90)
In article <tim.656621816@ggumby> tim@ggumby.cs.caltech.edu (Timothy L. Kay) writes: [stuff deleted] >You claim that X bogs down the network? That is like saying that my >workstation gets bogged down by running programs in the background, >so I'd like an operating system that doesn't let programs run in the >background (e.g. MS-DOS). > I have run countless tests over both 9600bps SLIP links and ethernet and have found X to be such a dog that 3 machines running *intensive*(real?) graphics programs brought a subnet to its knees. As an analogy, why drive a car that gets 10 miles per gallon and goes 100 MPH when I have two alternatives that get better gas mileage and go faster? >Neither NeXTstep nor NeWS let you run over the network. This is a very >important feature of X that very few other window systems ever had. NeWS >doesn't have it and NeXTstep doesn't have it. If you don't want to bog >down your network, don't run X over the network. But don't keep me from >doing it. Bzzzttt... Wrong! Both NeXTstep and NeWS are networked window systems, (Its what the N in NeWS stands for), with more efficient network protocols. > >Have you ever used NeWS? It is awful. I have to learn PostScript and I >have to learn the entire NeWS run-time environment. Unfortunately, it >isn't documented. "Well, it is all written in PostScript, so read the >source," you say. Except that all the work gets done by PostScript >primitives which are written in C. For those you don't get the source. >While the idea behind NeWS is a good one, it is implemented very badly. > That is like saying that the only way to understand a C program is to know the assembly language representation. What part of NeWS do you think is good? >I have been using a Silicon Graphics Iris with NeWS. We and many of >their other customers complained so much that the next release of their >OS is throwing out NeWS and moving to X. > There is a long story behind this, one better discussed in another forum. >Also, there is no reason that you couldn't have a PostScript-based >X window manager. This would give you the best of both worlds. > This still doesn't provide a unified imaging model, as both NeXTstep and NeWS do. [ stuff about NeXTstep not being networkable deleted ] >I really think the NeXTstation looks like a good machine. I might buy >one. However, I wouldn't buy one if X weren't available. Thanks to >those who have ported it! And I would certainly buy one if NeXT >supported X. > >I might like the NeXTstation even more if it ran *just* X (skip NeXTstep >altogether). > Over Steve Jobs dead body! :-) >It is my opinion that NeXT will have to eventually support X. That is, >if they manage to see their fifth birthday. > No, someone will just port X to their machine, they will never have to officially support it. >Tim Peter -------------------------------------------------------------------------- These opinions are mine and not the Labs.
bob@MorningStar.Com (Bob Sutterfield) (10/23/90)
In article <tim.656621816@ggumby> tim@ggumby.cs.caltech.edu (Timothy L. Kay) writes:
Neither NeXTstep nor NeWS let you run over the network... If you
don't want to bog down your network, don't run X over the network.
But don't keep me from doing it.
NeWS certainly runs over the network. Just telnet to TCP socket 144
on any machine running a NeWS server, and whisper PostScript in its
ear.
NeXTstep ... is a good environment, except that it doesn't run over
the network. It would be great if I could run Mathematica on one
workstation and have another workstation display the output.
NeXTStep applications certainly do run over the network. They
communicate over Mach ports, which traffic can traverse an IP
internet. Unfortunately, not every machine on the network runs Mach,
and therefore most can't talk to your NeXTStep server. The stubs
still exist in the source (or did, at least during the summer before
the Cube was released) to accept connections via TCP-domain sockets,
having been used during NeXT's development process while debugging the
port stuff. Despite my vigorous protestations and those of several
others, Steve didn't see any advantage in this. The capability was
unfortunately removed from the production version.
(Why would I want to do this? Suppose somebody else in the
building has a NeXT with 32 MB RAM, while mine has only 8 MB. I
don't want to work in his office, but I'd like to run M on his
machine.)
Why would I want to communicate via TCP sockets (as do NeWS and X)?
Suppose somebody else in the building, or halfway around the planet on
the Internet, has a Butterfly or a Sequent or a Convex or a Cray. I'd
like to be able to compile NeXTStep applications to run on his machine
but display on mine (just as I do now with NeWS and X). This would
enable the NeXT to become better integrated with the rest of the
existing computing environment. Terminal emulation isn't enough, and
that's part of why people clamored for X.
Oh yeah, NeXT doesn't ship the sources to their libraries, so nobody
can run their applications on anything but a NeXT box. Despite my
vigorous protestations and those of several others, Steve didn't see
any advantage in this ("The PC and the Mac are big successes without
shipping sources.") They still don't ship library sources, but even
if they did, you'd still need Mach ports to use them.
Sigh...
To NeXT: Please provide TCP socket access to your window server, and
please establish some way for "qualified" developers to have access to
the necessary sources to build NeXTStep applications on arbitrary
computers.dd26+@andrew.cmu.edu (Douglas F. DeJulio) (10/23/90)
tim@ggumby.cs.caltech.edu (Timothy L. Kay) writes: > Have you ever used X? You claim that X bogs down the machine. This is a truth. > X runs incredibly nicely on Sun Sparcstation 1 machines. Just because it's responsive and feels quick, that doesn't mean it's running nicely. It's still consuming a disproportionate amount of system resources. I want a window system that will run well on a slow machine, because it will consume very few cycles on a quick machine. > Neither NeXTstep nor NeWS let you run over the network. This is plainly false. NeXTstep lets you run over the network. I've done it lots of times. That's what all the "public window server" stuff is about. The other small and fast windowing system I've used (CMU/IBM's "wm") also lets you throw windows over a network. (As a matter of fact I've got some Sun2s running at home as reasonable "wm" terminals via a SLIP line.) > Also, there is no reason that you couldn't have a PostScript-based X > window manager. This would give you the best of both worlds. No, it wouldn't. Having something as huge as X11 combined with a display postscript would result in something too hideous to contemplate. Sure, it might run well on an ultra-fast workstation, but there's no excuse for spending all those resources on what should be a simple GUI. To improve X11, you need to make it much smaller and less customizable. > I really think the NeXTstation looks like a good machine. I might buy > one. However, I wouldn't buy one if X weren't available. Thanks to > those who have ported it! And I would certainly buy one if NeXT > supported X. Part of my reason for buying a NeXTstation is that it *doesn't* have X11. NeXT is wisely choosing not to help people use the beast on their machines. People at NeXT, are you listening? -- Doug DeJulio dd26@andrew.cmu.edu
brtmac@maverick.ksu.ksu.edu (Brett McCoy) (10/24/90)
In <gb96flu00VI8JTKroU@andrew.cmu.edu> dd26+@andrew.cmu.edu (Douglas F. DeJulio) writes: >> X runs incredibly nicely on Sun Sparcstation 1 machines. >Just because it's responsive and feels quick, that doesn't mean it's >running nicely. It's still consuming a disproportionate amount of >system resources. I want a window system that will run well on a slow >machine, because it will consume very few cycles on a quick machine. The xnews server isn't any faster than the standard MIT X server, and it uses about twice as much memory. I figure that if you take out the X code so that all you have is the NeWS code, you are still going to wind up with a server that uses at least the same amount of memory. As far as CPU usage goes, processing postscript and converting it to pixrect function calls can't be any less CPU intensive than taking the X calls and displaying them. -- Every day you scream at me to turn the music low. But if you keep on screaming, you'll make me deaf, ya know. -- Judas Priest Brett McCoy brtmac@maverick.ksu.ksu.edu
wiml@milton.u.washington.edu (William Lewis) (10/24/90)
In article <BOB.90Oct23112139@volitans.MorningStar.Com> bob@MorningStar.Com (Bob Sutterfield) writes: >To NeXT: Please provide TCP socket access to your window server, ... It's already there; check your /etc/services (or your nidump services, or whateverthehell it is). I forget the port number, but it does work; even lets you use binary encodings and receive events, as well as typing plain postscript in over Telnet. Of course, it *doesn't* work if the window server's port isn't registered, but that's exactly as it should be. -- wiml@milton.acs.washington.edu Seattle, Washington (William Lewis) | 47 41' 15" N 122 42' 58" W "These 2 cents will cost the net thousands upon thousands of dollars to send everywhere. Are you sure you want to do this?"
jfreem@uncecs.edu (Joe Freeman) (10/25/90)
>In article <tim.656621816@ggumby> tim@ggumby.cs.caltech.edu (Timothy L. Kay) writes: > Neither NeXTstep nor NeWS let you run over the network... If you > don't want to bog down your network, don't run X over the network. > But don't keep me from doing it. > I probably came into this late but NeXT step is client server and does run over the network. The 1.0 version ran between NeXT and IBM machines and I assume that 2.0 will do the same. <joe>
richf@adiron.UUCP (Rick Fanta) (10/25/90)
dd26+@andrew.cmu.edu (Douglas F. DeJulio) writes: >> Also, there is no reason that you couldn't have a PostScript-based X >> window manager. This would give you the best of both worlds. >No, it wouldn't. Having something as huge as X11 combined with a >display postscript would result in something too hideous to >contemplate. Sure, it might run well on an ultra-fast workstation, >but there's no excuse for spending all those resources on what should >be a simple GUI. To improve X11, you need to make it much smaller and >less customizable. >> I really think the NeXTstation looks like a good machine. I might buy >> one. However, I wouldn't buy one if X weren't available. Thanks to >> those who have ported it! And I would certainly buy one if NeXT >> supported X. >Part of my reason for buying a NeXTstation is that it *doesn't* have >X11. NeXT is wisely choosing not to help people use the beast on >their machines. People at NeXT, are you listening? >-- As much as I hate to admit it, I do agree that Next should support X. On a personal level (ie. having a Next at home, or working on an independent project where user interface compatibility isn't important) then I don't care what GUI a machine uses as long as it's fast and easy to program. From a buisness standpoint, however, I probably want a full port of X available. Don't get me wrong, I think it's a great machine and Nextstep probably is much better technically and in speed. But we use Suns and Decstations running X around here as do many other people. In fact, our customers (sigh, we're a defense contractor, but please don't hold it against me) are starting to demand X just so that they can run our apps on all their machines. Also, consider the universities that Next is wooing. They probably aren't exclusively Next-based either. Now your asking them to accept and support a machine with a non-standard (albeit, possibly better) interface. That's just adding more work (granted Interface Builder probably is much preferable to working with any Toolkit or widget set under X). Basically, even though the new Nexts may wind up outperforming even machines like Sparcstations (it's quite possible, I'm waiting to see benchmarks, especially floating point), I can't recommend to my bosses that they buy some to replace our more ancient machines. I'm sure the Next wins hands down on price-performance vs. any Sun out today, but if I can't run all of X on it the same as any other machine, it loses out. There maybe some hope, though. Peterd says that someone has done a port of X11R4. I wonder how complete it is. Also, in the new Next "Software and Peripherals" catalog (e.g. third-party stuff) they list a company called Pencom Software as working on a port of X11R4 with (Augh!) Motif that should be available damn soon (4th Qtr, 1990). I still hope to buy a Next within the next year or so for myself. If Next ever decides to support (WELL!) the same stuff that the big boys are going to, then they'll sell a ton of them to businesses like this one. ------------------------------------------------------------------------------ Rick Fanta PAR Technologies "... She said that she was working for the ABC news, it was as much of the alphabet as she knew how to use ..." - Elvis Costello