SunOS 4.1.1 window system & SunDEW

SunOS 4.1.1 window system & SunDEW

Post by James W. Ada » Thu, 22 Oct 1998 04:00:00



Since some recent posters have speculated about "Sun's original window
system" as shipped in 4.1.1 and how this might or might not allow
remote X display, I thought I'd post a brief explanation.  This is
not intended to be a historically authoritative summary.

Sun's early window system, SunView, was indeed local to the workstation,
and its programming interface became something of a de facto standard
for science/engineering applications.

Realizing that networks were becoming increasingly important, Sun
developed a followon based on James Gosling's work which was initially
called SunDEW, but later changed to NeWS (Network/extensible Window
System).  You may recognize Gosling as the true father of the EMACS
text editor.  

This is a prime example, IMHO, of the UNIX community stepping on its
own crank, as it were, without the help of Microsoft.

By the late 1980s, Sun's NFS had become the de facto standard for
distributed filesystems.

IBM, HP and DEC, among others, were bound and determined that this
recent upstart was not going to establish another such victory.

MIT had long before developed the Athena project as a networked
computing environment for student/academic use.  This employed a
networked window system called X which relied on transmission of
bitmaps.  Given the relatively primitive state of computer graphics
at the time, and the limited application of this protocol, this
seemed reasonable.

Sun enlisted Gosling and others to design a followon product for
SunView which would work over a network.  The SPARC design project
was underway, and graphics were becoming increasingly important
in computing.

Gosling's solution was extremely elegant.  He decided to use
Postscript, a (then) newly developed page description language
as the basic communications medium for the entire window system.
Instead of transmitting raw bitmaps over the network, objects were
encoded in Postscript and that was sent to the window server.

It is amusing to note that NeWS 1.0 on a Sun 3/60 ran about as fast
as X11R6 does on a 300 MHz Ultra.

Under NeWS, one could display a Postscript page by simply cat'ing the
file to the window server.  SGI also adoped NeWS for their highend
graphics systems.  Sun's first X server was actually written in the
NeWS enhanced Postscript language, and running on a Sun-3 actually
blew away every then-existing X implementation, even on superior
hardware.

Unfortunately, IBM, HP and DEC decided to throw enormous amounts of
money at MIT to develop Athena into a marketable, general-purpose
window system to compete with NeWS.

So, now we're stuck with X, arguably the most stupid networked
window system ever designed.

MBA students may want to use this as a case study.

Anyway, the SunOS 4.1.1 XNeWS server is capable of remote display,
but it is based on a very broken version of X11R3/4 and won't work
with Motif applications or many X11R4/5/6 apps.

This is an object lesson that technological superiority is easily
defeated by myopic marketing.

--

        "I became obsessed with angels and ballerinas,
         things of grace and beauty, otherworldly."
C*tesville, VA  22903         --C. Love

 
 
 

SunOS 4.1.1 window system & SunDEW

Post by Thad Flory » Thu, 22 Oct 1998 04:00:00



| [...]
| System).  You may recognize Gosling as the true father of the EMACS
| text editor.  
| [...]

Not wishing to initiate any prolonged debate, I'm curious how you derived
the above assertion.

I was using the MIT TECO-based Emacs before 1980, I still have an original
Version 150 manual handed to me by RMS in John McCarthy's office at Stanford
early in September 1980, and I still have all the TECO and Midas sources for
versions 150 to 163 (but, alas, no system remaining to compile them on :-)

In the "Emacs implementations, list of, regular post [long, FAQ]" posted
to comp.editors, one can find:

        name: Emacs
        last changed/verified: 1994-12-20
===>    original distribution: 1975
        version: 165
        base language: MIDAS (PDP10/DEC-20 assembly language)
        implementation language: TECO
        extension language: TECO
        scope of implementation: extensible
        hardware/software requirements: PDP10/ITS or DEC-20/TOPS-20
        organization/author:
                Richard M. Stallman
                MIT AI Lab/MIT Lab. for Comp. Sci.
                545 Technology Square
                Cambridge MA 02139
                USA
===>    note: this is the original
        free, anonymous FTP from
        [...]

So, if the comp.editors FAQ states RMS' Emacs is the original, how
could Gosling be "the true father of the EMACS text editor" when his
versions (later known as Unipress Emacs) came afterwards?

And, FWIW, many/most of the commands I was using in RMS' "Twenex" Emacs
circa late 1970s function [nearly] identically in the latest GNU Emacs.

Thad

 
 
 

SunOS 4.1.1 window system & SunDEW

Post by Barry Margoli » Thu, 22 Oct 1998 04:00:00




>So, if the comp.editors FAQ states RMS' Emacs is the original, how
>could Gosling be "the true father of the EMACS text editor" when his
>versions (later known as Unipress Emacs) came afterwards?

I also found his description of Gosling inappropriate.  The only way to
rationalize it would be if he said "father of Unix Emacs", since his was
the first Emacs editor on Unix.  I think there were also some claims by
Unipress in the early 80's that RMS incorporated some code from Gosling's
Emacs when creating GNU Emacs -- in the end, I think they agreed that the
only thing that was taken was the redisplay, and RMS replaced that module.

Shortly after Gosling Emacs there was also zemacs for Unix, by Steve(?)
Zimmerman, which was commercialized by another company whose name I don't
recall.

They may have been there first, but free software won big in this battle.

--

GTE Internetworking, Powered by BBN, Burlington, MA
*** DON'T SEND TECHNICAL QUESTIONS DIRECTLY TO ME, post them to newsgroups.

 
 
 

SunOS 4.1.1 window system & SunDEW

Post by Thad Flory » Thu, 22 Oct 1998 04:00:00




| >So, if the comp.editors FAQ states RMS' Emacs is the original, how
| >could Gosling be "the true father of the EMACS text editor" when his
| >versions (later known as Unipress Emacs) came afterwards?
|
| I also found his description of Gosling inappropriate.  The only way to
| rationalize it would be if he said "father of Unix Emacs", since his was
| the first Emacs editor on Unix.  I think there were also some claims by
| Unipress in the early 80's that RMS incorporated some code from Gosling's
| Emacs when creating GNU Emacs -- in the end, I think they agreed that the
| only thing that was taken was the redisplay, and RMS replaced that module.
|
| Shortly after Gosling Emacs there was also zemacs for Unix, by Steve(?)
| Zimmerman, which was commercialized by another company whose name I don't
| recall.
|
| They may have been there first, but free software won big in this battle.

Regarding Barry's comment re: Zimmerman, I found this reference in the
comp.editors Emacs FAQ:

        organization/author:
                Steven Zimmerman
        no longer available
        Dated 1983.  Described as a "distant descendant of the one written by
        Warren Montgomery at Bell Labs.  Might be an early, non-commercial
        version of CCA Emacs.

I just went browsing some of the shelves in my office library and found
yet another Emacs manual, this one for "CCA Emacs" by CCA Uniworks, Inc.,
a Crowntek Company, Version 163.0, October 1985.

On its frontispiece can be found:

        Uniworks, Inc.
        20 William St.
        Wellesley, MA 02181

On its page iv can be found this paragraph:

        ``Much credit is due to Richard M. Stallman of the MIT Artificial
          Intelligence Laboratory, who conceived of and wrote the original
          EMACS and who also wrote the original documentation from which
          this manual evolved.
        ''

There is sufficient evidence, both anecdotal and in FAQs and published
documents, attesting RMS's Emacs is the original one.

Thad

 
 
 

SunOS 4.1.1 window system & SunDEW

Post by James W. Ada » Thu, 22 Oct 1998 04:00:00


Perhaps I omitted the critical "UNIX" in the above.  Shoot me now.

Stallman proceeded to turn EMACS into the (IMHO!!!!!) useless monolithic
replacement for the OS, civilization, and Jesus Christ and Mohammad
which it now is.  We used to joke about when /vmunix was going to become
an EMACS macro.  I use MicroEMACS, thank you very much, because  GNU
EMACS is now so bloated that it rivals the entire output of Microsoft
extended to three orders of magnitude.

In case NOBODY noticed, the focus of my post was the X window system,
which is similarly bloated and useless, and was, again IMHO,
a kludge to begin with.

I wasn't trying to post a history of EMACS.  I will defer to your
experience with this.  I was too busy stuffing punched cards into
Control Data supercomputers to care about such things at the time.
UNIX was largely a toy to me in 1975, and only became meaningful
professionally when it became the subject of UNICOS and CONVEXos.

Otherwise, I used G(e)COS and, occasionally, MULTICS to do routine
work, with an occasional foray into TOPS-10 which usually ended
when I went out to lunch and returned only to find that the prompt
had yet to return.  We already had equal or better tools.

Again, if you want to follow up to this thread, please comment on
the subject instead of re-initiating the whole tired Stallman debate.

The subject is Sun WINDOW SYSTEMS, not EMACS.

--

        "I became obsessed with angels and ballerinas,
         things of grace and beauty, otherworldly."
C*tesville, VA  22903         --C. Love

 
 
 

SunOS 4.1.1 window system & SunDEW

Post by Anthony Mandi » Fri, 23 Oct 1998 04:00:00



> Stallman proceeded to turn EMACS into the (IMHO!!!!!) useless monolithic
> replacement for the OS, civilization, and Jesus Christ and Mohammad
> which it now is.  We used to joke about when /vmunix was going to become
> an EMACS macro.  I use MicroEMACS, thank you very much, because  GNU
> EMACS is now so bloated that it rivals the entire output of Microsoft
> extended to three orders of magnitude.

        At least he didn't try and include the kitchen sink. He left that
        to others to try - most notably Bill Gates. And if you ever run NT
        you'll notice that it certainly feels like its dragging one around!

Quote:> In case NOBODY noticed, the focus of my post was the X window system,
> which is similarly bloated and useless, and was, again IMHO,
> a kludge to begin with.

        I had noticed and, as a NeWS lover, was interested. I also noticed
        that the only respondants to your post wanted to discuss EMACS.

Quote:> Again, if you want to follow up to this thread, please comment on
> the subject instead of re-initiating the whole tired Stallman debate.

> The subject is Sun WINDOW SYSTEMS, not EMACS.

        And its one I'd like to comment on. Considering that the other
        players are either jumping ship or moving on, and considering
        that Motif was designed to be Windows 3.1-like, and considering
        that things have moved on now, why does Sun still bother with it?
        XView and NeWS were/are far superior. I saw a comment once that
        Java was Gosling's NeWS reborn. Certainly Java has picked up
        quite a head of steam and very quickly too, I might add. But
        how long before it is the desktop rather than just on it?
        (Note I'm not considering Javastations here but the ordinary
        Sun workstations that the rest of us all use.)

-am

 
 
 

SunOS 4.1.1 window system & SunDEW

Post by Thad Flory » Fri, 23 Oct 1998 04:00:00



| Perhaps I omitted the critical "UNIX" in the above.  Shoot me now.

Naw, we're civilized.  The honorable solution is hara-kiri.   :-)

| [...]
| In case NOBODY noticed, the focus of my post was the X window system,
| which is similarly bloated and useless, and was, again IMHO,
| a kludge to begin with.

And the focus of your article regarding the windowing systems WAS
very interesting and informative.  What kind of response were you
expecting to see?  Something like:  <AOL> Me Too!!!!!!!! </AOL> ? :-)

| [...]
| Again, if you want to follow up to this thread, please comment on
| the subject instead of re-initiating the whole tired Stallman debate.
|
| The subject is Sun WINDOW SYSTEMS, not EMACS.

Precisely.

There really was no reason to slip-in that inaccurate one-liner crediting
Emacs to Gosling when your article concerned windowing systems, so you
should expect to reap what you sowed.

Thad

 
 
 

SunOS 4.1.1 window system & SunDEW

Post by Winston Edmo » Fri, 23 Oct 1998 04:00:00




Quote:>So, if the comp.editors FAQ states RMS' Emacs is the original, how
>could Gosling be "the true father of the EMACS text editor" when his
>versions (later known as Unipress Emacs) came afterwards?

   Richard Stallman (RMS) developed the original Emacs.  It was a set of
TECO macros.  He also further extended the already highly modified MIT
AI-Lab TECO as needed to support the new functionality.  TECO itself was
written in assembly language and so wasn't easily portable to the new
(typically Motorola-68000 series) Unix workstations.

   James Gosling wrote the first Unix Emacs, using "C" and creating "Mock
Lisp" as its extension language.  Later, GNU Emacs appeared.  Like
Gosling's Emacs, GNU Emacs was written in "C" and used a Lisp-based
extension language (but not Mock Lisp).  GNU Emacs also came with a
Mock-Lisp-to-ELisp conversion program (which didn't work very well for me).

   Gosling's Emacs was initially the better of the two, but once it mutated
into "Unipress Emacs" and began costing real money, users began converting
to GNU Emacs, which didn't have copying or usage restrictions.  IIRC, it
was, in fact, the imminent change of Gosling's Emacs from free to for-sale-
with-restrictions that prompted the development of GNU Emacs.
 -WBE
(an Emacs user since the TECO days)
[courtesy copies of followups discouraged; I read the newsgroup]

 
 
 

SunOS 4.1.1 window system & SunDEW

Post by Philip Bro » Sat, 24 Oct 1998 04:00:00



>    And its one I'd like to comment on. Considering that the other
>    players are either jumping ship or moving on, and considering
>    that Motif was designed to be Windows 3.1-like, and considering
>    that things have moved on now, why does Sun still bother with it?

"it"? I will assume you mean "X11".

Well, for starters "Motif != X11".
Then there's "X11 in some form now can run on just about every UNIX system
on the planet, AND optionally for MS-DOG, AND other systems, so why the heck
wouldn't they "still bother with it"?

Quote:>    XView and NeWS were/are far superior. I saw a comment once that
>    Java was Gosling's NeWS reborn.

Funny, I thought that Java was elisp reborn. Plus, we finally get the
"kitchen sink" factor of it finally having its own graphics calls. :-)

Also, I presume you meant "SunView", not "XView". Since XView I believe is
the X library implementation of the basic SunView API [ or something that
supposedly was "incredibly easy to port to" from SunView], and can still be
gotten in various places.

Most commonly, /usr/openwin/lib/libxview.so on solaris 2.6

--
[trim the no-bots from my address to reply to me by email!]
 --------------------------------------------------
"initiating.. 'getting the hell out of here' maneouver" - Lennier, babylon5

 
 
 

SunOS 4.1.1 window system & SunDEW

Post by Anthony Mandi » Sat, 24 Oct 1998 04:00:00




> >       And its one I'd like to comment on. Considering that the other
> >       players are either jumping ship or moving on, and considering
> >       that Motif was designed to be Windows 3.1-like, and considering
> >       that things have moved on now, why does Sun still bother with it?

> "it"? I will assume you mean "X11".

        Sorry, I should have made myself clearer. I was actually thinking
        of CDE and its Motif interface (along with the crock that Display
        Postscript is).

Quote:> Well, for starters "Motif != X11".
> Then there's "X11 in some form now can run on just about every UNIX system
> on the planet, AND optionally for MS-DOG, AND other systems, so why the heck
> wouldn't they "still bother with it"?

        Do you really think that Sun cares all that much what other systems
        can do?

Quote:> >       XView and NeWS were/are far superior. I saw a comment once that
> >       Java was Gosling's NeWS reborn.

> Funny, I thought that Java was elisp reborn. Plus, we finally get the
> "kitchen sink" factor of it finally having its own graphics calls. :-)

        I understand its so that it can tie into JavaOS, but I could be wrong.

Quote:> Also, I presume you meant "SunView", not "XView". Since XView I believe is
> the X library implementation of the basic SunView API [ or something that
> supposedly was "incredibly easy to port to" from SunView], and can still be
> gotten in various places.

        Acutally, I should have said OpenWindows/Open Look. Its when NeWS
        made its appearance. But so far, Sun have gone thru three interfaces
        an appear to heading for a fourth. (Note that XView implements the
        Open Look look and feel and so isn't SunView, but uses the same
        API approach.)

-am

 
 
 

SunOS 4.1.1 window system & SunDEW

Post by James W. Ada » Sat, 24 Oct 1998 04:00:00


Quote:>> Also, I presume you meant "SunView", not "XView". Since XView I believe is
>> the X library implementation of the basic SunView API [ or something that
>> supposedly was "incredibly easy to port to" from SunView], and can still be
>> gotten in various places.

>    Acutally, I should have said OpenWindows/Open Look. Its when NeWS
>    made its appearance. But so far, Sun have gone thru three interfaces
>    an appear to heading for a fourth. (Note that XView implements the
>    Open Look look and feel and so isn't SunView, but uses the same
>    API approach.)

XView indeed is essentially an extension of the SunView API.  I ported it
to CONVEXos.  SunView was a good local window environment, and a lot of
useful software was written for it.  It was worth porting.

I am frustrated by what I see: Java on top of CDE/Motif, on top of X11.
Bloat upon bloat upon bloat.  Most vendors can't even deliver a working
Imake, which was in itself another example of bloat.

At least OpenLook was aesthetically pleasing.  That seems to have been
its downfall.  Ovals aren't "professional."  (I.E., they don't look like
Microsoft Windows.)  BTW, OpenLook was a AT&T product, their window manager
for X.

NeWS was a beautiful, technically superior window *system* built on open,
published standards.  If we had NeWS, we wouldn't NEED Java.

Perhaps that was its undoing.  It was too elegant, and it permitted too
much creativity.  Microsoft is attempting to destroy Java in the same
fashion.

The only thing elegant about NT was its internals, originally based on
the internals of VMS.  Now all that is rapidly fading in the face of the
need for Microsoft to control everything, lest some remotely intelligent,
talented or creative human innovate.  We must prevent that at all costs.

For any grammatical bigots, "internals" is a singular noun.

People sometimes seem to forget that computers are tools to do real
work which doesn't involve "computing" per se.  Sometimes it seems
like computer hardware and software vendors are doing the equivalent of
redesigning nuts and bolts every couple years in order to force people
to buy new tools to so the same jobs.  Hmmm, perhaps they are.  Torx and
all that.

Jerking people around to create a new market seems more important than
honest competition in a market oriented around what people really need.

I'd better start listening to those Saturday afternoon infomercials...

--

        "I became obsessed with angels and ballerinas,
         things of grace and beauty, otherworldly."
C*tesville, VA  22903         --C. Love

 
 
 

SunOS 4.1.1 window system & SunDEW

Post by Philip Bro » Sat, 24 Oct 1998 04:00:00



Quote:

>I am frustrated by what I see: Java on top of CDE/Motif, on top of X11.
>Bloat upon bloat upon bloat.

yah. Interesting things are happening at kaffe.org. A direct Xlib
implementation. But unfortunately, the AWT stuff isn't quite up to*yet.

Quote:>NeWS was a beautiful, technically superior window *system* built on open,
>published standards.  If we had NeWS, we wouldn't NEED Java.

Okay, let's not get carried away on a wave of nostalgia here :-)
Two completely different things here:
1) a windowing system
2) a portable operating environment, INCLUDING a _partial_ window system.

Quote:>Perhaps that was its undoing.  It was too elegant, and it permitted too
>much creativity.  Microsoft is attempting to destroy Java in the same
>fashion.

Eh. Actually, the way I heard it, it was to a great extent shot down because
of the "everyone else has X11, we want it too" factor. Nothing to do with
the merits or failings of XNeWS.
(Although come to think of it, I do remember there was some complaints
 specifically ABOUT XNeWS. But let's not speak ill of the dead :-)

--
[trim the no-bots from my address to reply to me by email!]
 --------------------------------------------------
"initiating.. 'getting the hell out of here' maneouver" - Lennier, babylon5

 
 
 

1. Windowing systems with older versions of SunOS?

Hi all,

Can some one please give me an authoritive listing of the different
windowing systems that came with the early releases of SunOS?  So far
I have:

        SunOS 2.x       SunView
        SunOS 4.1       OpenWindows
        Solaris 2.5     CDE

Is that right - and what, if anything, predates SunView (SunOS 1.0)?
Where did NeWS fit into all this?

TIA,

--
Rich Teer

President,
Rite Online Inc.

Voice: +1 (250) 979-1638
URL: http://www.rite-online.net

2. newbie's quesion about solaris printer setting

3. FXHTML Pilot Released for Linux, Windows 95/NT & SunOS

4. NiceTry remote

5. Network Monitoring & Apache & SunOS & Virtual Hosting

6. le0 - packets coming only in one direction

7. system calls in Solaris2.5 & SunOs 4.x

8. Mail filters and .forward files

9. Unix utility tr (translate) behaves differently on SUNos 5.4 & SUNos 5.5.1

10. &&&&----Looking for a unix shell------&&&&&

11. ppp && PPPoE && ADSL && net && buffer(s)

12. Dell Inspiron 7000 & Linux Redhat 6.0 & Kernel 2.2.10 & X-Windows

13. SunOS 4.1.3 && Kadb