How X started ..

chomwitt · 2023-12-24 06:59:20

First there was : VGTS by Nowicki et al [1] Partitioning of Function in a Distributed Graphics System William I. Nowicki

Then W. W is a bit mysterious as not much public information exist on internet. But from the quote below two persons are credited for W ,Paul Asente and Chris Kent.

We acquired a UNIX based version of W for the VS100 (with synchronous communication over TCP produced by Paul Asente and Chris Kent at Digital’s Western Research Laboratory.

But if W was ported to the DEC's VAX-11 computers in order to be used with the VS100 (which was a graphic terminal) from the V operating system then what was the use of Workstation Graphics Architecture documented in document on WGA from Henry M. Levy who for 8 years from 1975-1983 was Engineer at VAX/VMS ? I get a satoshi vibe on W..

It could be that W is just the initial letter of Workstation Graphics Architecture created for DEC's VS100 ? Maybe Henry M. Levy created W and Paul Asente[ and Chris Kent did the porting of W to Unix (because DEC was also selling a Unix like OS? All three were working at DEC ..

Paul Asente was also involved later in the X toolkit development. (see also X Toolkit Intrinsics - C Language Interface. )
The Western Research Laboratory (WRL) was a computer systems research group that was founded by Digital Equipment Corporation in 1982.

(1983) In May 1983 MIT announced the establishment of a five-year program to explore new innovative uses of computing in the MIT curriculum. This program was Project Athena.
Aside from educational goals additional technical objectives were:
- support of a heterogeneous hardware configuration
- provide a user interface independent of hardware.
- provide a software development enviroment independent of hardware
- maximize the exportability and importability of hardware
Also one umbrella requirement was that the system created should have ''coherence'' . Coherence meant that Athena should act as unifying glue allowing sharing of computational resources . That need couldnt but lead to the requirement for indepedence. So we could say that Athena had in it's dna a distributed soul since it mean to unify diverse computational resources across a networked campus.
Michael Dertouzos is credited for that idea . Dertouzos was also Director of MIT/LCS and is also credited as one of the heads that initiated and supported Project Athena..
It's interesting also to note that for some applications( Athena Laboratory Data System ) the 'coherence' requirement was abandoned and the Unix thin workstation couldn't meet real time requirement of certain hardware (multiple signal processing) in favor of an integrated IBM PC.
I think that means that if you need more control on the peripherals added that can not be done on the thin client and wait for all the signals to propagate to the server to process them. That could be an argument against a fixed division of computational labor implemented in X. Although i am not sure if real time computational requirement can be met by any future distributed system. When you distribute work far away across a shared network you win in sharing and economy but you lose real time requirements.

Finally many many people in MIT, DEC,IBM are credited for making the Athena Project a reality . Combining various things i guess is like genetic merging ... :-) . Some times can be fun and productive ! But lets also not forget that all Project MAC (project was a misnomer in purpose, it was a lab) was funded by DARPA.

( ref )

(1984) X was announced . It was based on W . Initial development by Robert W Scheifler and Jim Getty ( ref )

19 June 1984
From: rws@mit-bold (Robert W. Scheifler)
To: window@athena
Subject: window system X
Date: 19 Jun 1984 0907-EDT (Tuesday)
I've spent the last couple weeks writing a window
system for the VS100. I stole a fair amount of code
from W, surrounded it with an asynchronous rather
than a synchronous interface, and called it X.
The X Window System by Robert W. Scheifler and Jim Gettys

(1985) An X11 License (precursor to MIT License) was added to X version 6 . .X was originally under a proprietary license but what we would now call an open source license was added to X version 6.
(ref ) Computer Systems Research (CSR) Group of the MIT–LCS (precursor to CSAIL) had developed several pieces of network software that were generating outside interest and requests for both information and copies of the code. That was the environment that created the fertile ground for MIT License. Jerome H. Saltzer seem to had a crucial role in the initial push for open licenses. Also that guy had Fernado J. Corbato as Doctoral advisor (Corbato was a pioneer in timesharing) he was a team leader to Multics (the Unix precursor) , led MIT's MIT-LCS/CSR and was techical director to Project Athena that gave X.
Keep in mind that moden MIT's CSAIL formed in 2003 by merger of : The Laboratory for Computer Science (LCS) and the Artificial Intelligence Laboratory (AI Lab). LCS has central role in Project Athena and has its roots at Project MAC and AI Lab was a lab where hacker like Richard Stallman thrived. ( ref ) .

In the fall of 1985, the question arose of how to license the X Window System[20] that was being developed by Jim Gettys[21] and Bob Scheifler[22] for MIT Project Athena.[23] In discussions parallel to those of two years earlier, they had noticed that proprietary licensing of early versions of the X Window System was becoming a hassle both for them and for prospective recipients and had the potential of interfering with widespread adoption. There were significant contributions made by early adopters, but that only made it more apparent that it was important to minimize the licensing friction.
Jerome H. Saltzer, The Origin of the “MIT License”

(1988-1993) MIT X Consortium formed as a non-profit vendor group, with Robert Scheifler as director. ( wikipedia )
Also the same year Keith Packard joined in March 1988 as senior developers. (i mention him because 15 years later it seems he had probably central role in the XFreee86 split and X.Org foundation )

(1993-1996) X Consortium, Inc is formed (a non-profit corporation) formed as the successor to the MIT X Consortium. ( ref ) X11 R6 (16/5/1994). According to that analysis of the negatives of the pro-paid work model used were : a) volunteer labor was crowded out b) drop in transparency in X to most of its developers c) change the direction of the project in ways that do not always serve the interests of users.

Because toolkits, desktop environments, and 3D were each controversial or proprietary areas among the funders, staff was not paid to work on these areas. The result was several major gaps in the X11 platform that restricted the ultimate usefulness of the software. Had the workers been working as volunteers, these were problems that would have been addressed early on — there was high demand from users. However, the introduction of paid labor directed the project in other directions.

(1999) X.Org is formed by Open Group.

(1991) X386 was an implementation of the X Window System on x86 arch.
X386 was created by Thomas Roell while at Technische Universität München . ( LJ Interviews Thomas Roell ),( Announcement of release 1.1 (as X386 1.1, based on X11R4) in 11/02/1991 .( ref ) (for reference 25/08/1991 Linux was announced.. so i guess that is a merit of os independence. The kernel came after the display server).

(1992) XFree86 project was formed to continue X386 probably from 1.2+ . initial release.
The project was originally founded by David Dawes, Glenn Lai, Jim Tsillas and David Wexelblat. ( an interesting archived interview with David Dawes that offer valuable insights about the funding of XFree86. Like for example the support from Tungsten Graphics that was co-founded by Brian Paul developer of Mesa library. )

(2003) A split happened in XFree86 core team. Keith Packard an X11 developer from 1988 was removed from XFree86 project ( ref1 ). Some months pass and the core team left sorf of disbands it self! ( Why a history of the X.org fork?)

(2004) X.Org foundation is founded and release X.Org X11 server.( wikipedia/X.Org) . X.Org foundation seem to be the continuation of

Interestingly i found that quote from David Wexelblat (XFree86 initial founders) from that 2003-2004 era:

"X is obsolescent," he wrote in a mailing list posting.
"I've been working in the Windows world for years now, and client-server display systems are utterly irrelevant to the majority of real-world computer users. X needs to be replaced by a direct-rendered model, on which a backwards-compatible X server can be reasonably trivially implemented".

It seems that at some point 3d accelerated graphics in personal computers created a 'gold rush' era that expanded and in the GNU/Linux t. Echoes of that era still makes waves by repeating old arguments and division about X and its role.

I wonder if the solution would simple enough. Those who dont care about the destktop intergrated style of windowing systems (DIWS) to step forward and maintain and improve X.

Ironically the XFree86 and the XOrg seems to represent a majority that favors DIWS. But at the same period 2002-2003 you could find stories about Linux thinclients networks that saved the budget of whole cities!

So i hope that short X history study of mine highlights that X is not only an old an inadequate 2d,3d renderer. X was a first of its kind successful open software distributed-networked windows system designed from day one to be platform independent and network native.
. An X12 or Y should improve on all that accounts. Why not an extensible NeWS like display server that can be part of a microkernel of a distributed OS ? (see also Why X is Not Our Ideal Windows System.) But surely an all in one solution that puts as priority the rendering aspect is for me a different window system. Not X like but Windows like.

The bad news is that it seems that from day one, in Linux land,those who worked to bring X to x86 , probably had in mind not X but a Wayland-like solution investing more of their energy mainly in the rendering-gpu-support aspect. Who knows.. maybe that was a job that HAD to be done before exploring other venues.. Its nice to have cool graphics in your windowing system in you GNU/Linux pc. But that is not in X tradition. That is something else.

Now seeing the more slowly rate of Wayland adoption it could mean that the Wayland-type window-system push although is a welcomed addition to the libre land and should be welcomed and encouraged nevertheless it could be that libreland and it's inhabitants are not an ms-window alternative land.
I think some want that . Steam wants that. Propably RedHat and many developers since there is where the more users are. But i think that over-done push could be also a mark of a certain stagnation..

On the other hand what did we expect ? When Thomas Roell started porting X11 to x86 PC i don't think he had in his mind the VS100 connected to a VAX 11 superminicomputers. IBM PCs were never meant to be thin clients.

It could be the case that X11's distributed soul in the process of being ported to a PC-oriented platform has been ''hijacked'' in favor for a PersonaWindowSystem PC's PWS . And it makes more sense to pay more attention to having fast and eyecandy graphics in a PWS than pay any attention to (as the wayland proponets call it) niche use cases that nobody cares for!.. And it makes sense in developing a PWS to want WS developers that can deliver that aspect.

Reading emails and articles from the Xfree86 - X.orgs 2003 (''split'') i havent found (so far) any trace of arguments in between Xfree86 developers regarding the use-cases that must be worked on the most. Instead the differences were on licenses and the organization efficiency to keep pace with rapid changes in the gpu - market.

In this view and context its seems logical (and a dream come true if you read the aspirations of the PWS developers mentality 30 years back) for a Wayland PWS to emerge.

What i dont like is that there are not any Distributed Windowing System (DWS) devs to come forward and claim X11's legacy. Since that was not the case from day 1 of the X386 port that explains the fact that X is being abandoned in favor for Wayland!!!.
That is happening because of the circumstances that i describe here. While it would make much more
sence for PWS proponent devs to split and create their own PWS project and dont tarnish X11 DWS aura.

Now we are stuck with X11/amd64 users (minority) that rely on X11/amd64 's DWS dna and indepedence form day1 (1980-83) era and X11/amd64 developers with a PWS mentality and abilities that are controlling (or that it seems to me so far) its development fueled by a strong flux of money and interest by GPU mammoths that see in games and AI huge market grow opportunities.

So the minority seems to be slowly getting thrown out of the AMD64/x86 Linux bus...

I wonder why Linux desktop virtualization (or DaaS) solutions of companies like NX-based solutions (nomachine), vmware ,teamware dont have any interest on the development of X11 i mean in a way to feel the necessity to be part of X.Org.

If they build X11 forwarding proprietary protocols shouldn't they want also to have a voice in the X.Org BoardOfDirectors ? Isnt it strange to have Valve and Codeweavers as guardians of X11 ? Or as i said before , it could be the case the X11's distributed spirit is long dead. That hypothesis could be tested also easily by looking at the sources to see the changes in various areas. Or could it be the case that Wayland can also function without any deficiances as Distributed Diplay server by its own extensions mechanisms ?

Speaking of distributed souls have you heard of Sun Ray. Sun's thinclient solution (1999-2013) that was based on separate network display protocol, Appliance Link Protocol (ALP)? SunRay was a successor to JavaStation a previous attempt (1996-2000) from Sun to create a thinclient market.

Is there any plan by Wayland to incorporate ssh forwarding? 2019 , Interesting discussion about X11 forwarding and Wayland . I think the discussion centers on one 'architectural' difference between X11 and Wayland. X11 has graphics primitives and that can beat a framebuffer 'forwarding' approach. But still couldnt wayland offers as extension Cairo primitives ?

A short timeline of X

Last edited by chomwitt (2024-01-03 13:25:58)

zapper · 2023-12-26 03:23:55

I honestly wonder if the corporatization of software on linux, isn't the reason why xorg and other older standards are not secure today.

It isn't exactly the corporate motto to have no bloat in operating systems. They go the other direction, microsoft is a perfect example of this.

dbus, systemd, similar nonsensical crap like that, probably has only hurt security for linux.

I sometimes wonder another words, if they are fixing a problem they created themselves.

Hmm...

The officially official Devuan Forum!

#1 2023-12-24 06:59:20

How X started ..

#2 2023-12-26 03:23:55

Re: How X started ..

Board footer