peterd@opus.cs.mcgill.ca (Peter Deutsch) (09/19/90)
In article <F9ix_c52@cs.psu.edu>, melling@cs.psu.edu (Michael D Mellinger) writes: > > In article <7753@milton.u.washington.edu> wjs@milton.u.washington.edu (William Shipley) writes: > > > Well, today at the big unveiling Jobs announced the NeXTstation (which > will be reviewed to death by others, so I'll be brief), which NeXT > document number N6030 describes as: > > > Anything earthshaking happen at the unveiling? Some of us couldn't > make it :-). Well, I didn't make it either (NeXT offered to pay my way, then withdrew the offer! :-( but I did write a news release that was supposed to be included in the press kit. I include a copy below. If for some reason the event didn't happen, I'm STILL NOT VIOLATING NON-DISCLOSURE as I am merely telling the world I like NeXTen and bought a bunch. Boy, am I sensitive. Anyways, for anyone too lazy to read our blurb below, the interesting bit is as follows: > As one example of this, McGill is currently working on > porting X11 Release 4 of Xwindows to the NeXT machine. A > beta version already runs in cooperation with version 1.0a > of NeXTstep and should be available for public distribution > soon. A stand-alone version is also planned and both ver- > sions will be submitted back to MIT for inclusion in the > standard X11 release. So there, X 11 Release 4 IS RUNNING ON THE NeXT! As for status, I have the best hacker at McGill working on the port for me, he has a working server and basically all the other bits of XWindows compiled and running in conjuction with NeXTStep 1.0a. There are a few little bugs, but the grunt work is done for the "draft" version. Imagine Xeyes on a NeXT, with black background, no borders, in a dark, unlit office. It looks sexy! :-) First, how we did it. We still need NeXTStep, but only to provide us with keyboard and mouse events. These are provided to the XWindows server, which translates them and handles them as appropriate X events. We have asked NeXT for details on accessing these directly and when/if we get this info, we will not need NeXTStep at all. In action, when Xwindows starts up, the first thing it does is draw a DPS window in front of all the existing NeXTstep windows. The X code then opens the frame buffer and talks directly to it, in effect scribbling directly on the screen. It really handles the two bits per pixel correctly and is actually a rework of a colour server to handle 2 bits instead of 8. When we exit X, we clear the DPS window and find ourselves back in NeXTstep. As I said, we still need NeXTStep to get keyboard and mouse events, and hope to bypass this if someone can tell us how. This all works now. One thing we plan to do is to add a toggle so we can switch between X and NeXTstep. The guy doing the port says it should be easy. We'd click on the X icon and go into X, and have some hot-key or icon to click to get back. They can co-exist just fine. The other possibility is to run X standalone, ie kill the NeXTStep window server and run the X server. I would like this option and my hacker says he can do it once we eliminate the event problem. If you know how to do what we want, write. If you know someone at NeXT who knows how to do what we want, wake them up after the big bash and get them to write. I have spoken to our Tech Rep who promises to speak to a couple of people, but hey, why wait!?! Performance? I think it's not blindingly fast on a 68030 cube, but I'm told it's "slower than monochrome on a Sun, but faster than 8 bit colour". This is providing real two bit per pixel support. Things to come? Well, we have 68040s on order and are eagerly awaiting 2.0 of the O/S to make sure nothing breaks, although it shouldn't. If/when NeXT announce colour or can get us beta machines we have offered to do the colour port. Availability? Not yet, but soon. The guy doing the port wants to give it back to MIT for inclusion in the standard release, so expect it at an archive server near you once we're done. Of course, we will post in comp.windows.x and here in comp.sys.next when it's ready. Problems? One problem is that a lot of the public domain X programs are broken, in that they hardwire in an assumption about how many mouse buttons you have (Hey! Not all machines have 3 button mice, guys!). This is not a server issue, it's a programmer education issue, but my suggestion is to map some special hot key on the keyboard to be the missing mouse button. I am not an X hacker so don't know how outlandish a suggestion this is, but my hacker thinks it can be done, so I believe him. The alternative is to fix a lot of broken code, but that seems like more work. One window manager I've seen can work with no mice or buttons at all. This is another obvious way around the problem. :-) I have deliberately omitted the name of the guy doing the port, as I don't want a flood of "Can I have copy NOW, please please, pluuuuueeeese?" letters slowing him down. He's good and is almost done. It will be released soon. And it will be free. Remember my email address and start bothering _me_ in a few weeks if someone haven't followed up. Thanks. One last comment on XWindows. I am not the guy doing the port, I have cooperated to the level of helping to pursuade someone to do it, then giving up my own workstation so the guy would have a testbed. I am not a qualified X (or even NeXTStep) hacker so I may have used the term "window" when I meant "box" or whatever. Sorry, but the gist is sound. I hope to see the hacker doing the port posting on this soon. I have been told (over the phone, not non-disclosed, nothing) that NeXT has contracted with some company to do an XWindows port, including Motif, etc. If you want that, then call NeXT. We're offering the public domain, freeware version of an X11 Release 4 server, and I hope/expect to be able to offer it before anyone else does. Finally, one last comment. I was flamed at one point about the non-disclosure rubbish. Someone accused me of "hurting NeXT financially" by releasing info I shouldn't have. I didn't answer because I didn't want to prolong what I thought was a silly thread, but here's my answer to the question I was asked (probably rhetorically) "so what have you do for NeXT financially?" Here's my answer. I bought 63 NeXTen and helped arrange the XWindows port. So what have you guys done? I 'll drop this now if everyone else does. - peterd ------------------------------------------------------------ September 18, 1990. (Montreal) - McGill University announced today that it has purchased 63 NeXT computers to equip a new teaching laboratory for the School of Computer Science. The machines will be used at all levels of the School's undergraduate teaching program, as well as for graduate level teaching and research. ``We began our search for new computer systems with a stringent set of requirements'', says Peter Deutsch, systems manager for the School. ``We had to have machines that could co-exist with a large number of different machines in a networked environment. We needed a machine that could run a large body of existing UNIX-based software. We wanted a fast machine, yet needed as many of them as possible. It just all seemed to come together with NeXT. ``Many students are taking our courses as electives and we see no reason to inflict raw UNIX upon them. At the same time, computer science students need to be able to 'lift the lid' on any conventional graphics user interface and access the raw O/S when required.'' Mr. Deutsch continued. ``Underneath the NeXTstep GUI is a stable implementation of Mach, which does a great job of looking like Berkley UNIX. We have been very pleased with the ease with which we can port software from our existing machines onto the NeXT com- puter.'' As one example of this, McGill is currently working on porting X11 Release 4 of Xwindows to the NeXT machine. A beta version already runs in cooperation with version 1.0a of NeXTstep and should be available for public distribution soon. A stand-alone version is also planned and both ver- sions will be submitted back to MIT for inclusion in the standard X11 release. The new machines join a number of NeXT computers used in individual professors' labs. Current machines are used for research in Speech Recognition, Numerical Analysis and Database Design. Several others are used for systems mainte- nance and monitoring tasks. ``Our success with the earlier NeXT computers helped us decide on this new purchase. We have a lot of faith in NeXT and look forward to having their new machines in our labs.'' said Mr. Deutsch. For information on McGill's NeXT purchase, contact Peter Deutsch by phone: (514) 398-6698, or through elec- tronic mail (peterd@cs.mcgill.ca). -------------------------------------------------------------------------- +-------+ Peter Deutsch McGill University | u # u | peterd@cs.mcgill.ca School of Computer Science |/\/\/\/| | a a | \ a / Go ahead, flame me. I have a /dev/null on my machine... \___/ --------------------------------------------------------------------------
engstrom@SRC.Honeywell.COM (Eric Engstrom) (09/20/90)
In article <2241@opus.cs.mcgill.ca> peterd@opus.cs.mcgill.ca (Peter Deutsch) writes: >[stuff deleted...] > > As one example of this, McGill is currently working on > > porting X11 Release 4 of Xwindows to the NeXT machine. > > [...] > So there, X 11 Release 4 IS RUNNING ON THE NeXT! > > As for status, I have the best hacker at McGill working on > the port for me, he has a working server and basically all > the other bits of XWindows compiled and running > in conjuction with NeXTStep 1.0a. There are a few little > bugs, but the grunt work is done for the "draft" version. > Imagine Xeyes on a NeXT, with black background, no borders, > in a dark, unlit office. It looks sexy! :-) > > [stuff about how X is being done on the NeXT - deleted...] Sounds good, but I've got a question... While I use X (11 rel. 4) on Suns of various denominations (in definite preference to Sun-Tools), I wonder the use of X on a NeXT. Admittedly, X has a large software "base", provides environment homogeneity among different machine types, and is free & well supported, but in terms of usability of the environment, I would still rather use NeXTStep on a NeXT. I'm sure this "usability" stems in part from simply being on a NeXT (machine and environment built together), but if IBM can port NeXTStep, why couldn't NeXTStep be ported to other machines (ignoring the licensing problem) I didn't post this to start a flame-war about the benefits of PD software (personally, I think groups like the X-Consortium and FSF are great! :-), but I'm just curious about how people would/will use X on the NeXT. (email me the flames and spare the group.) Eric (--- Disclaimer: The opinions contained within are entirely my own and (--- do not necessarily represent those of my employer. (--- (but I didn't have to sign any non-disclosure agreement for them.) +--------- Eric Engststrom, Honeywell SRC | [While most peoples' ARPA: engstrom@src.honeywell.com | opinions change, the UUCP: {umn-cs,ems,bthpyd}!srcsip!engstrom | conviction of their MAIL: 3660 Technology Drive, Mpls, MN | correctness never does] Phone: (612) 782-7318 |
olson@sax.cs.uiuc.edu (Bob Olson) (09/20/90)
In article <92180@srcsip.UUCP> engstrom@SRC.Honeywell.COM (Eric Engstrom) writes:
great! :-), but I'm just curious about how people would/will use X on
the NeXT.
There are those of us who are great fans of the NeXT, and enjoy using
and programming it immensely, but who have to do project development
using X for the work they get paid for :-). Using X on the NeXT would
give me the best of both worlds... now to convince my boss that a
NextDimension is the ideal visualization platform :-).
--bobpeterd@opus.cs.mcgill.ca (Peter Deutsch) (09/20/90)
In article <92180@srcsip.UUCP>, engstrom@SRC.Honeywell.COM (Eric Engstrom) writes: > In article <2241@opus.cs.mcgill.ca> peterd@opus.cs.mcgill.ca (Peter Deutsch) writes: [ my stuff on X deleted ] > Sounds good, but I've got a question... > > While I use X (11 rel. 4) on Suns of various denominations (in definite > preference to Sun-Tools), I wonder the use of X on a NeXT. > > Admittedly, X has a large software "base", provides environment > homogeneity among different machine types, and is free & well supported, > but in terms of usability of the environment, I would still rather use > NeXTStep on a NeXT. I'm sure this "usability" stems in part from simply > being on a NeXT (machine and environment built together), but if IBM can > port NeXTStep, why couldn't NeXTStep be ported to other machines > (ignoring the licensing problem) > I didn't post this to start a flame-war about the benefits of PD > software (personally, I think groups like the X-Consortium and FSF are > great! :-), but I'm just curious about how people would/will use X on > the NeXT. I'd like to say a word or two about this, as the whole issue of X availability was a big one when we were researching our purchase. We are part of a global community of educational institutions, who have collectively produced an incredible environment (the Internet, workstations, Gigabytes of archive servers, Usenet, etc). One of the things that makes all that work is interoperability at the machine level. We don't all run the same hardware, but we keep in mind that whatever we do might have to be ported to some pretty bizarre hardware. Sooo, when I started evaluating workstations (we obtained demos for IBM, Sun, Dec, Data General, Solbourne, Silicon Graphics) we tried several things. We have a whole directory full of P.D. tools and other software that we ported to each machine (or tried, in some cases). We brought over our editors, our UNIX filters, etc. The usual stuff, less, lc, etc. For the record, the NeXTen matched Sun pound for pound on this, better than virtually every other machine. As our primary machines are Sun now, Sun's supremacy was expected. The NeXT performance, although with some problems, surprised and pleased me. As part of fitting into this environment, we pretty much have to accept the supremacy of XWindows. We _don't_ have to endorse the design, but _everyone_ (with the obvious exception of NeXT) had a commitment to it when we started. Frankly it was their major "hit". They had networks, they had UNIX (well Mach, but it really does look like Berkely), they had neat extras like the DSP chip. They had an Object oriented environment, etc. But they didn't have X, and everyone else we deal with did. As part of our negotiations, we received assurances (in writing) from NeXT that this would be addressed. On the strength of this, and given the advantages of the machines they were quoting us (over 50 of the ones announced yesterday, plus eight cubes to make opticals available to anyone that wants it and to give us colour) we went NeXT. Fortunately for me, I am not stuck waiting for NeXT to follow through (although I have no reason to think they wont). I had access to a guy who would do the port for me as a challenge. I'm glad he did, as we now have that much bigger a collection of P.D. stuff to run and our students will be exposed to this obviously important standard. Of course, as I outlined in my previous posting, our server should be able to co-exist with NeXTstep, which gives us the best of both worlds. Sort of like Solbourne's X server, which will emulate either Motif or Open Look by selection as a command line switch. We have greater flexibility to go with the world when we must, but have a nice interface, with some neat features when we want. For what it's worth, I don't actually care if my buttons are round or square, my scroll bars on the left or the right. I like _ALL_ the options. With X on the machine, the one thing I felt they were missing is fixed. Those waiting for Cobol can't say that, but I can. Note that if I was buying 50 or 60 machines for an office, I would have a different analysis, but I might even come to the same conclusions. In any event, _we_ needed it so we did it. If you want it, it should be ready soon. If you don't, have fun with Improv. - peterd ^X ^I .signature -------------------------------------------------------------------------- +-------+ Peter Deutsch McGill University | u # u | peterd@cs.mcgill.ca School of Computer Science |/\/\/\/| | a a | "From MAILER-DAEMON@hq.demos.su Thu Sep 13 00:45:55 MSD 1990" \ a / \___/ The day we made contact.... -------------------------------------------------------------------------- Look! I can type after the signature!
peterd@opus.cs.mcgill.ca (Peter Deutsch) (09/20/90)
In article <OLSON.90Sep19122115@sax.cs.uiuc.edu>, olson@sax.cs.uiuc.edu (Bob Olson) writes: > In article <92180@srcsip.UUCP> engstrom@SRC.Honeywell.COM (Eric Engstrom) writes: > > great! :-), but I'm just curious about how people would/will use X on > the NeXT. > > There are those of us who are great fans of the NeXT, and enjoy using > and programming it immensely, but who have to do project development > using X for the work they get paid for :-). Using X on the NeXT would > give me the best of both worlds... now to convince my boss that a > NextDimension is the ideal visualization platform :-). Well, now for you guys to convince NeXT to get us the colour machines we ordered quickly so we can start the colour port. I know, why don't you all write NeXT and tell them you like this! And tell them to thrown in a 16 bit one or two, as well! :-) ^^^^^ ||||| [ Puuullleeessseee note the smiley! ] - peterd -------------------------------------------------------------------------- +-------+ Peter Deutsch McGill University | u # u | peterd@cs.mcgill.ca School of Computer Science |/\/\/\/| | a a | \ a / "Love my work, hate my job..." \___/ --------------------------------------------------------------------------
hunt@boulder.Colorado.EDU (PUT YOUR NAME HERE) (09/20/90)
In article <92180@srcsip.UUCP> engstrom@SRC.Honeywell.COM (Eric Engstrom) writes: >> [stuff about how X is being done on the NeXT - deleted...] > > >Sounds good, but I've got a question... > >While I use X (11 rel. 4) on Suns of various denominations (in definite >preference to Sun-Tools), I wonder the use of X on a NeXT. > >Admittedly, X has a large software "base", provides environment >homogeneity among different machine types, and is free & well supported, >but in terms of usability of the environment, I would still rather use >NeXTStep on a NeXT. While I agree the NeXTStep is a nice GUI, what I don't think you understand is the flexibility of X window. The "look" and behavior (window border, gadgets, pop-up window, etc.) are all aspects of the X window _manager_, which can be anything one desires (eg: twm, uwm, motif, openlook, etc.). So I'm sure than someone could make a NeXTStepesque window manager to work pretty much like it does now (although it would most likely be a bit slower). The X window _server_ is what is common to all X window setups -- it's the device that lets you communicate with X window clients (X window applications) and other X window servers. This brings up a very neat concept of X -- the idea that all X terminals are "equivalent" such that you can open a window on any other X terminal (as long as you have permission). This is quite handy when you want to show a window to someone remotely for their pursal. The flexibility and gererality of X makes it a Good Thing. >Eric | --Lee | | Dave Haynie is Bob! | | hunt@spot.colorado.edu ...!ncar!boulder!spot!hunt |
doug@eris.berkeley.edu (Doug Merritt) (09/20/90)
In article <OLSON.90Sep19122115@sax.cs.uiuc.edu> olson@sax.cs.uiuc.edu (Bob Olson) writes: >There are those of us who are great fans of the NeXT, [...] but who have >to do project development using X for the work they get paid for :-). Similarly, my company must support several hardware platforms so that our software will run on the customer's machines. X gives us a way to develop just one piece of software, rather than writing brand new software for each target platform's windowing system, saving mucho time and money. Naturally we need to have X on our development machines. Once X is available for NeXT, I can make a case for using NeXT as our development system. Until then, there's nothing I can say, and we just keeping buying more Sparcs. As usual, compatibility in one form or another turns out to be the overriding concern in the real world. Doug Doug Merritt doug@eris.berkeley.edu (ucbvax!eris!doug) or uunet.uu.net!crossck!dougm
rca@cslab5a.cs.brown.edu (Ronald C.F. Antony) (09/21/90)
Why must X blank the screen? I thought XNeXT was so handy in that it allowed the two window systems to coexist on the same screen. Why not keep that, please!! Ronald ------------------------------------------------------------------------------ "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man." Bernhard Shaw | rca@cs.brown.edu or antony@browncog.bitnet ------------------------------------------------------------------------------