gert@prls.UUCP (Gert Slavenburg) (09/09/87)
From the reactions that were published on the net OR directly mailed to me, it is clear that there is a lot of interest in X-windows on the Atari-ST. To make this discussion a little clearer to outsiders, let me explain (my limited understanding) of what X-windows is : + X-windows is a PROTOCOL that defines how a CLIENT program that wants to manipulate windows on a screen should talk to a SERVER program on some IPC (Inter Process Communication) link in order to do output and receive mouse and keyboard events. + An X-window installation involves writing a SERVER for that machine that will implement this protocol onto the specific hard- ware (notably bitmap display and mouse/keyboard) of the target as well as porting a number of CLIENT programs that are indepen- dent of the hardware, but often depend strongly on Unix support. Also, in the case of the Atari ST, there are (as yet) no networking standards, let alone a distributed IPC standard. Hence we either need to develop a standard for the distributed client/server case or limit ourselves to running client and server on a single ST. My personal goals in getting X-windows going are : 1) I have written one major GEM program using windows/mouse/keyboard etc. according to GEM rules. From this I learned that programs with good user interfaces depend SO STRONGLY on GEM that they are very difficult to port to other environments. The next time I write a large (I'm talking 500k source here) appli- cation, I want it to run on VAXes, SUN's, Atari's, VAXstations, etc. 2) Atari and Digital Research seem to be non talking (Neil : correct me if I'm wrong !). Which means that I own a machine, the manufacturer of which will not (and cannot) maintain its Operating System in the future. They only have a binary license ! If we can port X-windows on top of Minix on the Atari-ST, we will have a state of the art, multitasking, multi window package, ALL SOURCE OF WHICH IS AVAILABLE TO ANYONE AT (ALMOST) NO COST ! Hence, no more dependency on Atari or any other manufacturer. It may also bring our machine to its knees, but I've heard that Minix on the ST is faster than the original release for the PC. So there is hope. Proposal : Minix is currently beta-testing on the ST (I have a copy). The Atari-ST has no networking capabilities, but WHO CARES ! For goal 1 and 2 above, all I care about is running a client and server on one machine. Later, if and when networking hardware for the ST becomes commonplace, we'll put it in ! The implementors of Minix on the ST (Johan Stevenson and others) are interested in getting X-windows (or any other standard) running. They are right now bogged down in the final coding for the real Minix release. But I'm sure they are willing to consult on 'howto' issues. HENCE : VOLUNTEER(S) wanted that have the knowledge and time to work on a non- networking port of X-windows under Minix. I do not have enough knowledge myself, nor enough time. I just act as 'the prophet' and can get serious volunteers in touch with the Minix people, if necessary. What we need is one person making a serious commitment, and willing to allow free dis- tribution of his/her efforts. Any commercial undertaking will not meet my goals and would be contrary to the spirit of both X-windows and Minix. (IDRIS and OS/9 : be warned !) Please comment on the above issues on the net, as long as it doesn't get out of hand. Gert Slavenburg PS : I received many reactions to my original posting. I will not answer you, but will attempt to develop some common consensus. Everyone seems to feel that waiting for Minix and for V11 of X-windows is the thing to do. Some people are working on/thinking of working on a port. Mail me to request an archive with all my mail on this topic, if you want to locate people with common interests. LET's work on our DECLARATION OF INDEPENDENCE from Atari ! Disclaimer : all of the above opinions are entirely my own. And I only repeat rumours anyway, I don't know what I'm talking about.
hays@apollo.uucp (John Hays) (09/10/87)
In article <6083@prls.UUCP> gert@prls.UUCP (Gert Slavenburg) writes: >From the reactions that were published on the net OR directly mailed >to me, it is clear that there is a lot of interest in X-windows on the >Atari-ST. > > The next time I write a large (I'm talking 500k source here) appli- > cation, I want it to run on VAXes, SUN's, Atari's, VAXstations, etc. > Don't forget Apollo [toot my employer's horn], IBM-RT, HP, Adobe [the postscript people], ... who are also on X and actually pushed it into the "STANDARDs" forefront. >LET's work on our DECLARATION OF INDEPENDENCE from Atari ! MISTAKE!!! If you really want this to be successful we must "beat Atari [Neil?] into submission" and get them on the X Window System/[Mi|U]n*x/Open Dialogue* bandwagon. The case can easily be made to Atari that X Window System/Minix is essentially free software for them to port to their machines. The market wants it. It opens the door to a lot more software ... We want people who are not on USENET/ARPA/... to have access to this work as well and if we can get the vendor to "support" it or at least distribute it. [MINIX PROMS?!!?] Atari might even get a jump on Apple on this one and be more interesting to Schools. It has been brought up before but X Window System is large, don't offer to port it unless you have at least a 30-Meg drive. There is some question as to whether the executable will fit into memory. (Perhaps Minix in place of the TOS roms and X Window System on a cartridge.) As far as an IPC ... Phil Karn's (karn@bellcore.com) TCP/IP implementation is being ported to MINIX/UNIX and might be a good candidate... or the addition of Unix Domain Sockets to the MINIX OS. =================== I hope you can figure out the tradmarks (IBM/APPLE/HP/APOLLO/SUN/VAX/X Window System/ADOBE/ ATARI/Open Dialogue/...) ================== * For those who are not familiar with Open Dialogue: Open Dialogue is an extensible User Interface Management System (UIMS) created by Apollo for use with the X Window System (and other window systems). This package gives the ability to create user interfaces with Menus/Scrollbars/Pop-Ups/OneOfs/Number Lines/... and user defined techniques. The application and the interface are logically divided such that multiple interfaces may be defined for a given application and it is frequently possible to generate new user interfaces without recompiling or relinking the application (some implementation issues here). [If you want more info contact Ross Chapman (apollo!chapman_r@eddie.mit.edu)] ================== -- John D. Hays, Consultant, Corporate Systems Engineering | My opinions are my Apollo Computer Inc. - Chelmsford, MA - USA | own -- If anyone ARPA: apollo!hays@EDDIE.MIT.EDU // UUCP: apollo!hays | really cares! PACKET RADIO: KD7UW @ K1UGM // CIS: 72725,424 GEnie: KD7UW // PO Box 21, 01824
bob@wiley.UUCP (Bob Amstadt) (09/10/87)
Posting-Front-End: GNU Emacs 18.47.4 of Tue Aug 4 1987 on candi (berkeley-unix) You've got to wait a little bit longer. I have been communicating with Johan Stevenson and from what he says, MINIX is not really ready for X. I have looked into the X server. It would be foolish to do a port on top of an unstable operating system. I have also determine that it makes no sense to port X into the Atari's native environment. Wait for MINIX to be released. You'll then have my full cooperation. -- ---Bob Amstadt bob@wiley.uucp {csvax.caltech.edu,trwrb.uucp}!wiley!bob