It's Motif on VMS, which is not remotely UNIX-like
eventually HPE got tired of dealing with VMS customer requests and sold the rights to VMS Software Inc, who ported it from Itanium to x86 as soon as humanly possible
now VMS Software Inc, is stating that they wish to support ye olde DECWindows and Motif on the modern, x86-enabled VMS
That's a cruel alternate universe, I would've hoped that Motif being in use by more people than just the devs of one homebrew Unix desktop would mean that we wouldn't have suffered through that much versionitis.
The version number was just a joke. But if the "standard unix gui toolkit" was under an open source license back in the 1990s, I am sure people would have run with that rather than inventing something else.
from 1989 to 2005 everyone used more or less the same version (from 1989) because vendors and standards are painful
it wasn't like, meaningfully standardized. just no one ever updated anything. or set a meaningful version string. you just guessed which bugs were un-fixed based on `uname`
I basically meant that we could've avoided the (needless) versionitis of gtk, the toolkit once introduced to rewrite a Motif-based application. (Never understand why they did have to reinvent the Xt part, too, but, well…)
I miss Motif. This is a portal to a time when men were men and UNIX(R)—or in this case, VMS—desktops were utilitarian and did exactly what you needed and nothing more.
Now we live in a time where we allocate GBs of RAM to eye candy that functionally accomplishes nothing. Then we make the case to rewrite the eye candy in increasingly "safe" languages, requiring even more RAM.
Eventually, but it wasn't first class, rather to port UNIX software into VMS.
Just like MS-DOS had plenty of C compilers to chose from, while it was actually written in Assembly, and most folks were programming in Turbo and Quick Pascal, Turbo and Quick BASIC, Clipper,...
Hardly the same kind of C for everything like on UNIX.
That a lot of software wasn't written in C on VMS is beside the point, the fact is that a totally normal K&R C was available on VMS, just the same as e.g. Pascal was available. And the C compiler was popular among people writing TCP/IP software (and I don't mean re-compiling some Unix ftp client), where they could just sit down with Steven's book and code.
On the contrary, that is indeed the point, safe languages have been a thing already before UNIX/C took off, and even the reason why Multics had a higer security score than UNIX, thanks to PL/I.
> Then we make the case to rewrite the eye candy in increasingly "safe" languages, requiring even more RAM.
Indeed, although C kind of felt second-class on VMS since the language has a lot of Unixisms embedded in the standard library and, to a lesser extent, the language itself.
Being able to define command line interfaces using cld files on VMS was really wonderful and you got things like abbreviations of options (and commands) to their shortest unique initial string was quite nice (so, for example, the directory command could be named as such but everybody just typed dir).
"did exactly what you needed and nothing more" You can still do that. Build a config for openbox or dwm. While the wm still compiles you can ignore the fads.
> Now we live in a time where we allocate GBs of RAM to eye candy that functionally accomplishes nothing.
Well, of course it takes more ram when we run 4x the pixels for the same size screen. And we double the refresh rate, but then hold everything back a frame to composite it. :P
I think the one frame delay is due to X11 clients rendering a frame, then the X compositor rendering a frame, both according to the usual timing rules. With Wayland, it should be easier for client and compositor to render in the same frame interval because the compositor properly controls frame timing. It can just tell clients to render slightly early, then use the remaining few milliseconds for compositing.
The Motif default theme was quite handsome, and the "demo" Motif Window Manager worked pretty well, but Motif was something of a nightmare to work with
The API sucks real bad, and even at the height of Motif popularity, the package itself was riddled with bugs because proprietary UNIX vendors never updated that shit
Motif was super-obviously designed by C++ programmers who could not ship a C++ library for technical reasons. So they tried to do a C++ API in C. And it hurts like a pineapple thrust into the wrong orifice, leafy-part-first.
I purchased a copy of OSF Motif for Linux (x86) sometime in the early 1990's (before it was free). I had used it before on SunOS and I liked it.
One of the most annoying things about it was that it did not address the endianness of the arguments to the library functions. So it worked fine on big endian platforms, but not so fine on little endian ones (such as Intel).
It would still work okay if you byte swapped the arguments in and out of the library functions, but it just seemed silly to need to do that, and it made it more difficult to write portable code.
So, I guess the history of 'windows NT' is lost on many. 'NT' started with version 3.1 as the MS/IBM breakup from the joint OS/2 venture happened. It was their first real push into 32 bit protected mode operating systems and supported really crazy cool things like 'multiple processors' and totally different architectures than x86, like DEC. Give the link a look.
the common element between VMS (the subject of this post) and Windows NT, is Dave cutler.
Cutler lived in an extremely overcomplicated world of VMS kernel primitives, and given the chance to let his freak flag fly, he really overcomplicated his past work for Windows NT
In case you ever wonder why your 1 gb/s ssd has ~100 mb/s throughput on windows. there are often quite literally hundreds of layers of filters on even the simplest i/o
but it is super flexible! just slower than iced treacle. aren't you glad you had an object oriented I/O subsystem supporting microkernel services and aspect-oriented programming? i bet you use those features way more often than you read or write files from disk
Edit: the previous poster has since completely rewritten this comment to talk about windows NT. they originally talked about Windows (without the NT) then linked to an NT wiki. Hence my reply.
Except NT 3.1 is still NT and NT didn’t “invent” the 3.1 design. They modelled it after Windows 3.1 (though technically Windows 3.0) and named NT as NT 3.1 for brand familiarity.
I'm not sure what you think I was implying with my post? The history I was pointing out was Windows NT 3.1 supported the DEC alpha processor (3.1 was marketing and implied the UI was similar to windows 3.x, which it was). This was a connection between DEC and windows that I thought was interesting and not the subject of the post, but in some interesting ways ties the two together. DEC did many things for a very long time like make machines, operating systems and processors [1] and is likely good fodder for a top level HN post in its own right. I remember dreaming about running a DEC NT3.1 machine when I heard about it. I had a friend in the AF who showed me the install disks for NT 3.1, a double stack of 1.44mb 3.5" disks. It must have taken hours to install. Anyways, that is the linkage. Take a look. It is fun history.
I put '[edit]' and new comments below that and didn't edit anything above, including the link. The post above the edit is all original. I'm sorry you have been confused through this.
I think NT is correct; NT was designed by Dave Cutler, who famously worked on VMS stuff before working for Microsoft. I think the poster was correct in posting NT.
No they’re not. The GUI design came from Windows 3.0. What NT do was take the design of that and the branding of Windows to help sell NT.
But NT 3.1 is a completely different OS to Windows 3.1.
They might look the same, but one has OS/2 heritage while the other has DOS heritage (to overly simplify their origins, but I’m in a rush this morning so can share more accurate details later if you wish).
Edit: the GP changed their comment. The original copy didn’t reference NT, it just said “Windows 3.1”. Hence my reply.
Even if the GUI design elements originated with Windows, I don’t think it’s incompetent to mention Windows NT when VMS comes up. Due to Cutler’s origins, a lot of people consider NT as the spiritual successor to VAX.
amusingly Motif and CDE were derived from HP attempts to copy Windows 2.x and the betas of Windows 3.0
not windows 3.1 -- windows 3.1 was popular! Windows before 3.1 was distinctly unpopular. It had basically no installed base. The only Windows 2.x applications I know of actually shipped an embedded Windows copy on the floppy disk.
HP was carefully tracking all the much less popular stuff Microsoft was doing in the late 80s because they thought this "WIMP" paradigm had staying powers, even if Microsoft was not exactly selling a lot of units
I'm rather tempted. I'm funemployed at the moment and live over in Stockholm so it's just a (long) train journey away. I used to enjoy messing around on Vaxen in my college days.
I can't see myself ever using VMS professionally though, and it has a somewhat vague description of the actual content. It's not super expensive but I'd have to fork out for the travel and accommodation.
I'd love to hear about it if anyone does a write-up after the fact.
I used to work for a convention planning company, so I shouldn't be surprised, but I somehow am surprised that there are enough VMS users to justify a conference. I have genuinely not heard of anyone using VMS in any context in more than twenty years.
Some folks have had success running it on certain server hardware (Usually HPE Proliant). There are no graphics drivers for x86, so it is X forwarding only.
I dunno what's interesting about this link, but Motif has been LGPL a while and the last release was in 2017.
https://sourceforge.net/projects/motif/files/
https://en.wikipedia.org/wiki/Motif_%28software%29
(in some alternate universe, motif was under the x11 license and you would have motif v13 instead of GTK.)
It's Motif on VMS, which is not remotely UNIX-like
eventually HPE got tired of dealing with VMS customer requests and sold the rights to VMS Software Inc, who ported it from Itanium to x86 as soon as humanly possible
now VMS Software Inc, is stating that they wish to support ye olde DECWindows and Motif on the modern, x86-enabled VMS
That's a cruel alternate universe, I would've hoped that Motif being in use by more people than just the devs of one homebrew Unix desktop would mean that we wouldn't have suffered through that much versionitis.
The version number was just a joke. But if the "standard unix gui toolkit" was under an open source license back in the 1990s, I am sure people would have run with that rather than inventing something else.
motif had the opposite of versionitis
from 1989 to 2005 everyone used more or less the same version (from 1989) because vendors and standards are painful
it wasn't like, meaningfully standardized. just no one ever updated anything. or set a meaningful version string. you just guessed which bugs were un-fixed based on `uname`
> motif had the opposite of versionitis
I basically meant that we could've avoided the (needless) versionitis of gtk, the toolkit once introduced to rewrite a Motif-based application. (Never understand why they did have to reinvent the Xt part, too, but, well…)
I miss Motif. This is a portal to a time when men were men and UNIX(R)—or in this case, VMS—desktops were utilitarian and did exactly what you needed and nothing more.
Now we live in a time where we allocate GBs of RAM to eye candy that functionally accomplishes nothing. Then we make the case to rewrite the eye candy in increasingly "safe" languages, requiring even more RAM.
Safe languages have nothing to do with it, case in point, the choice of programming languages available on VMS.
Which contrary to UNIX did not had the C mistake.
Rather Structured BASIC, Extended Pascal, COBOL, Modula-2, Fortran and Bliss.
It is really sloppy programming nowadays, regardless of the languages.
Hm? VMS also had K&R C, and a simulator I worked with at the time was written (not by me) using that compiler.
Eventually, but it wasn't first class, rather to port UNIX software into VMS.
Just like MS-DOS had plenty of C compilers to chose from, while it was actually written in Assembly, and most folks were programming in Turbo and Quick Pascal, Turbo and Quick BASIC, Clipper,...
Hardly the same kind of C for everything like on UNIX.
That a lot of software wasn't written in C on VMS is beside the point, the fact is that a totally normal K&R C was available on VMS, just the same as e.g. Pascal was available. And the C compiler was popular among people writing TCP/IP software (and I don't mean re-compiling some Unix ftp client), where they could just sit down with Steven's book and code.
On the contrary, that is indeed the point, safe languages have been a thing already before UNIX/C took off, and even the reason why Multics had a higer security score than UNIX, thanks to PL/I.
> Then we make the case to rewrite the eye candy in increasingly "safe" languages, requiring even more RAM.
- Typo: I meant to write "Stevens' book"
Unix software like X11 and Motif…
Yes, they were even available MS-DOS via Coherent, but not the main way to develop software on the platform.
Indeed, although C kind of felt second-class on VMS since the language has a lot of Unixisms embedded in the standard library and, to a lesser extent, the language itself.
Being able to define command line interfaces using cld files on VMS was really wonderful and you got things like abbreviations of options (and commands) to their shortest unique initial string was quite nice (so, for example, the directory command could be named as such but everybody just typed dir).
Much of VMS was eventually rewritten in C to accomodate the Alpha.
By then it was OpenVMS.
"did exactly what you needed and nothing more" You can still do that. Build a config for openbox or dwm. While the wm still compiles you can ignore the fads.
You can use a CDE lookalike https://github.com/NsCDE/NsCDE
The real thing is open source since 2012 https://sourceforge.net/projects/cdesktopenv/
Time to port this to Wayland using Claude code, right?
points finger you son of a bitch, I'm in.
> Now we live in a time where we allocate GBs of RAM to eye candy that functionally accomplishes nothing.
Well, of course it takes more ram when we run 4x the pixels for the same size screen. And we double the refresh rate, but then hold everything back a frame to composite it. :P
I think the one frame delay is due to X11 clients rendering a frame, then the X compositor rendering a frame, both according to the usual timing rules. With Wayland, it should be easier for client and compositor to render in the same frame interval because the compositor properly controls frame timing. It can just tell clients to render slightly early, then use the remaining few milliseconds for compositing.
The Motif default theme was quite handsome, and the "demo" Motif Window Manager worked pretty well, but Motif was something of a nightmare to work with
The API sucks real bad, and even at the height of Motif popularity, the package itself was riddled with bugs because proprietary UNIX vendors never updated that shit
Motif was super-obviously designed by C++ programmers who could not ship a C++ library for technical reasons. So they tried to do a C++ API in C. And it hurts like a pineapple thrust into the wrong orifice, leafy-part-first.
Not really, because it wasn't a thing when Motif took off.
There are at least, for those wanting a Linux or BSD based Motif fix:
Enhanced Motif Window Manager https://fastestcode.org/emwm.html
and the full-fledged CDE desktop that uses Motif also:
https://sourceforge.net/projects/cdesktopenv/ (note that you want to firewall this somehow as the default settings on the background process ttdb can be a security hole)
I got this patched and up and running on my own LFS distro a few years back, lots of fun. It even has some support for antialiased fonts these days.
UTF-8 support very welcome.
I purchased a copy of OSF Motif for Linux (x86) sometime in the early 1990's (before it was free). I had used it before on SunOS and I liked it.
One of the most annoying things about it was that it did not address the endianness of the arguments to the library functions. So it worked fine on big endian platforms, but not so fine on little endian ones (such as Intel).
It would still work okay if you byte swapped the arguments in and out of the library functions, but it just seemed silly to need to do that, and it made it more difficult to write portable code.
I saw DEC windows and immediately thought of Windows NT 3.1.
https://en.wikipedia.org/wiki/Windows_NT_3.1
[edit]
So, I guess the history of 'windows NT' is lost on many. 'NT' started with version 3.1 as the MS/IBM breakup from the joint OS/2 venture happened. It was their first real push into 32 bit protected mode operating systems and supported really crazy cool things like 'multiple processors' and totally different architectures than x86, like DEC. Give the link a look.
the common element between VMS (the subject of this post) and Windows NT, is Dave cutler.
Cutler lived in an extremely overcomplicated world of VMS kernel primitives, and given the chance to let his freak flag fly, he really overcomplicated his past work for Windows NT
In case you ever wonder why your 1 gb/s ssd has ~100 mb/s throughput on windows. there are often quite literally hundreds of layers of filters on even the simplest i/o
but it is super flexible! just slower than iced treacle. aren't you glad you had an object oriented I/O subsystem supporting microkernel services and aspect-oriented programming? i bet you use those features way more often than you read or write files from disk
You’ve linked to NT, not Win 3.x
They were entirely different OSs.
Edit: the previous poster has since completely rewritten this comment to talk about windows NT. they originally talked about Windows (without the NT) then linked to an NT wiki. Hence my reply.
@OP Poor show on your ninja edit.
No, I linked to Windows NT 3.1. The history here is important. It supported DEC alpha, hence the post.
Except NT 3.1 is still NT and NT didn’t “invent” the 3.1 design. They modelled it after Windows 3.1 (though technically Windows 3.0) and named NT as NT 3.1 for brand familiarity.
So NT 3.1 != Windows 3.1
As you said “history here is important”.
I'm not sure what you think I was implying with my post? The history I was pointing out was Windows NT 3.1 supported the DEC alpha processor (3.1 was marketing and implied the UI was similar to windows 3.x, which it was). This was a connection between DEC and windows that I thought was interesting and not the subject of the post, but in some interesting ways ties the two together. DEC did many things for a very long time like make machines, operating systems and processors [1] and is likely good fodder for a top level HN post in its own right. I remember dreaming about running a DEC NT3.1 machine when I heard about it. I had a friend in the AF who showed me the install disks for NT 3.1, a double stack of 1.44mb 3.5" disks. It must have taken hours to install. Anyways, that is the linkage. Take a look. It is fun history.
[1] https://en.wikipedia.org/wiki/Digital_Equipment_Corporation
I'm old enough to have lived it. :)
I put '[edit]' and new comments below that and didn't edit anything above, including the link. The post above the edit is all original. I'm sorry you have been confused through this.
With the greatest of respect, if you’d put “Windows NT 3.1” originally then I wouldn’t have commented. But you didn’t.
Anyway, I think we’re both plenty experienced on both platforms so I don’t see any need for us to argue over NT or not to NT.
I think NT is correct; NT was designed by Dave Cutler, who famously worked on VMS stuff before working for Microsoft. I think the poster was correct in posting NT.
No they’re not. The GUI design came from Windows 3.0. What NT do was take the design of that and the branding of Windows to help sell NT.
But NT 3.1 is a completely different OS to Windows 3.1.
They might look the same, but one has OS/2 heritage while the other has DOS heritage (to overly simplify their origins, but I’m in a rush this morning so can share more accurate details later if you wish).
Edit: the GP changed their comment. The original copy didn’t reference NT, it just said “Windows 3.1”. Hence my reply.
Even if the GUI design elements originated with Windows, I don’t think it’s incompetent to mention Windows NT when VMS comes up. Due to Cutler’s origins, a lot of people consider NT as the spiritual successor to VAX.
spiritual successor? how about "ghoulish horror" ;)
the worst example of "second system effect" i have ever heard of
To be fair, I actually think that the NT kernel is fine, and arguably better than Linux. It’s the rest of Windows that is terrible.
Fair point.
amusingly Motif and CDE were derived from HP attempts to copy Windows 2.x and the betas of Windows 3.0
not windows 3.1 -- windows 3.1 was popular! Windows before 3.1 was distinctly unpopular. It had basically no installed base. The only Windows 2.x applications I know of actually shipped an embedded Windows copy on the floppy disk.
HP was carefully tracking all the much less popular stuff Microsoft was doing in the late 80s because they thought this "WIMP" paradigm had staying powers, even if Microsoft was not exactly selling a lot of units
Anyone here going to the VMS bootcamp? [1]
[1] https://events.vmssoftware.com/bootcamp-malmo-2026
I'm rather tempted. I'm funemployed at the moment and live over in Stockholm so it's just a (long) train journey away. I used to enjoy messing around on Vaxen in my college days.
I can't see myself ever using VMS professionally though, and it has a somewhat vague description of the actual content. It's not super expensive but I'd have to fork out for the travel and accommodation.
I'd love to hear about it if anyone does a write-up after the fact.
I used to work for a convention planning company, so I shouldn't be surprised, but I somehow am surprised that there are enough VMS users to justify a conference. I have genuinely not heard of anyone using VMS in any context in more than twenty years.
Red Hat recently removed Motif from their distribution. I wonder if it's still in AIX.
https://access.redhat.com/solutions/6113101
"I wonder if it's still in AIX."
I think aix still ships CDE, too
which would be unremarkable except none of the UNIX vendors has produced a new build in like 30 years
so if they are still shipping CDE it is probably 32 bit binaries from AIX 4 or AIX 5L
"deprecated" is not "removed"
it can take literally decades for a deprecated package to actually get removed, because customers get mad
I think it is removed from rhel10. I can check if you want.
The entire 32-bit subsystem (multilib) is also removed from rhel10.
It is as of 2 years ago, the last time I had access to AIX. many proprietary applications on AIX used Motif Libs.
Dual booting OpenVMS (not virtualization) would be a really cool thing.
Some folks have had success running it on certain server hardware (Usually HPE Proliant). There are no graphics drivers for x86, so it is X forwarding only.
I still have my Vax 780 pen.
I have a Vax 730. But you can lift a pen. :-)
yes;1