1. 21
  1. 11

    PostScript is a page description language and the ancestor to PDF. It was used as the basis of two different graphical display systems: Display PostScript, used by NeXT and OPENSTEP, and NeWS.

    Display PosctScript was the simpler of the two: it defined purely a drawing canvas that accepted PostScript programs to display output; you still had to write windowing-system-specific code to open windows and handle input events. Think of it like Cairo, which is very similar in concept of not execution.

    NeWS is closer to the modern Web: a graphical canvas with a rich “client-side” scripting language. With NeWS, the PostScript layer could handle input events and respond without talking to the application side of things. It was possible to write real applications in nothing but NeWS’s PostScript.

    It’s definitely the most fascinating, IMHO, of the windowing systems that could have been…

    1. 5

      James Gosling talks about the development of NeWS and what happened in his CHM oral history interview.

        1. 4

          In X’s defense it was made available to vendors and didn’t require licenses from Sun and Adobe. Cheap almost always wins over elegant. :)

          1. 7

            It has been years since I looked into this but I don’t think it’s quite this cut and dried.

            X11 had some pretty compelling advantages over NeWS for workstation vendors:

            • X11 is open source and there are no licensing concerns.
              Sun was late to open source NeWS.
            • X11 is an open standard so you can use the reference implementation or build your own.
            • NeWS is owned by a competitor.
              The license terms may change in the future or the license could be revoked.
              They could hold back fixes or updates in the future.
            • Sun will always be first to market with NeWS enhancements.
            • The X11 implementation can be much simpler; I have no idea whether it is.
            • X11 doesn’t have to guard against PostScript DoS attacks.
            • X11 doesn’t require a garbage collector.
            • X11 primitives map more closely to the hardware.
            • It’s easier to upstream fixes for X11.
            • It’s easier to sell and support a system that only does X11 instead of X11 and Display PostScript or NeWS.
              You don’t have to explain the differences to customers.
            1. 5

              Sun was late to open source NeWS.

              Did Sun ever open source NeWS? That would be an incredible project to resurrect.

              1. 3

                NeWS source was available for $995 but it couldn’t be redistributed (source).
                I don’t think it was ever available as open source.

                1. 1

                  That would be awfully cool if only for historical research purposes.

                2. 5

                  A political History of X @ LCA 2020 he goes into NeWs bits in particular about 35 minutes in.

          2. 4

            Back when I was getting started with Linux in the early 2000s, the GUI options sucked. Everything was extremely clunky and slow and crashed a lot. I read some breathless posts about a group trying to clone/port NeWS to Linux, and tried for at least a couple of days to get it working, with no success. There was some brief window where NeWS could have taken over on Linux though!

            1. 2

              Ah, what could have been. It would be cool to see what a parallel universe where NeWS got popular would look like.

            2. 3

              I’ve laid hands on a Sony/NeWS machine once in my life

              1. 4

                NEWS != NeWS - it’s certainly confusing though!

                1. 1


              2. 3

                Does any one know what happened to Noticias? seemed to be an implementation of NeWS?

                1. 2

                  I wonder if they got a takedown notice from Oracle.

                  1. 2

                    Hmm, you are right. They could have; But given it was in Github for some period, I am hopeful there is some source somewhere.

                    1. 1

                      This is an argument for a distributed system like IPFS where the source code Just Keeps Floating Around and as long as SOMEBODY SOMEWHERE has a copy you can still get it :)

                      1. 2

                        One of my hobbies is array programming languages. The only extant complete documentation for k is a copy of the user and reference manuals for k2 (which is substantially similar to k3 and completely different from all later versions, which are the current versions).

                        The manuals in question are just…floating around. You have to find them. They’re not officially available from anywhere and samizdat is basically your only option (though they’re not hard to find in PDF form if you really look).

                2. 2

                  If only there was an implementation of NeWS I could run somewhere :(

                  I really really like postscript as a language, and I wish it was the basis of academic typesetting instead of latex.

                  1. 4

                    If you can find an SGI with version 3 of IRIX, that has NeWS. I used IRIX 3 for maybe a year or so, and was very sad when I had to upgrade to version 4 (which dropped NeWS).

                    1. 1

                      Thanks, good to know.

                    2. 3

                      I wish it was the basis of academic typesetting instead of latex.

                      I’m not sure I follow. TeX/LaTeX can output to .dvi (DeVice Independent file) which can be translated to PostScript.

                      Is anyone creating documents using raw PS? I must admit I’m a huge LaTeX weenie, and I really like the model of compose / compile / proofread as opposed to the more traditional WYSIWYG of word processors. But LaTeX is a set of macros on top of TeX, which was seen as too “low-level” for many users. PS is even more low-level.

                      For example, in a LaTeX doc I can write (example from here):


                      and the rendering engine will ensure it’s positioned correctly on the page. I think that in PS I’d have to manually plase all the elements in the statement myself?

                      1. 6

                        A (niche) convenience of actually writing postscript for scientific computing was that it’s a programming language. I used to work in a physics lab, where one of my colleagues wrote a Fourier transform algorithm in postscript. It could draw to the (Display Postscript) screen on our NeXT workstations, but it was actually faster to print graphs, because they ran on the more powerful CPU in our printer.

                        1. 4

                          PS is not really too low level; it is sort of like FORTH and Lisp in that you can build your own language on top of it. People have been using plain post script for publications for a long time now. Here is one such.

                          If you want to have a taste of how PS is modern, see this which I did some time back. It feels pretty much like modern high level functional programming in point free form. Here is a meta-circular interpreter in postscript

                          As regarding the latex comment, I really meant Tex, in that Latex is the library on top of Tex, and that is exactly what PS lacks for academic publications.

                          1. 1

                            Thanks for expanding, and thanks for the link. Looks very interesting.

                            I’m not sure 100% of the history, but I think maybe that PS wasn’t entirely free to output/print? I.e. it was encumbered by licenses. DVI is a neutral format that could be used to output to other printers than those that implemented Adobe’s drivers.