This is interesting: it turns out there was a NextStep release for IBM AIX workstations. From the initial, archived press release (via Steven Troughton-Smith):
AIX PS/2 NextStep Environment Version 1.1 is a state-of-the-art graphical user interface and programming environment for AIX workstations, designed to be compatible with the same application programming interface (API) as the NextStep product, Software Release 1.0, provided by NeXT, Incorporated.
AIX PS/2 NextStep Environment Version 1.1 provides icons and menus to facilitate access to system utilities and applications. The AIX NextStep Interface Builder is designed to provide a rich set of well-defined objects and graphical cut-and-paste capabilities for designing and implementing application user interfaces. The Objective-C (3) Compiler provides the benefits of object-oriented programming for developers who choose to design additional objects for the application development environment. AIX PS/2 NextStep Environment can help increase the productivity of programmers and end users.
Steven Troughton-Smith, who has a thing for collecting NEXT/early OS X builds and versions, is now looking for this piece of software history, but not a whole lot can be found about this online. I did ran into a thread in comp.sys.next.advocacy from 1995 in which a Robin D. Wilson sheds some more light onto the fate of this product:
And we ran it on an RS/6000 model 540 (with 63MB of RAM no less) — it was pretty fast. The thing that killed it is Steve Jobs wanted IBM pay more money for 2.0. They had only _just_ finished porting 1.0 to AIX (it did run on top of AIX — and there were several hacks made to accomodate it — but it did run fine). When NeXT was shipping 2.0, IBM felt they wouldn’t be able to sell 1.0 (there we some rather dramatic improvements between 1.0 and 2.0). They also didn’t want to spend more money on it (as SJ was demanding for 2.0), and they didn’t feel like porting 2.0 would take any less time (meaning they wouldn’t get done until NeXT released a newer version). All that considered — IBM abandoned NS.
This wasn’t a “bad decision” by SJ (per se), but I can see IBM’s view on
this easier than I can see NeXT’s…
Steven also stumbled upon a very, very long FAQ about NextStep/AIX, which contains tons of information. This will probably be very hard to find, but for the sake of digital archaeology and preservation, we really need to find it somewhere and preserve it. Absolutely fascinating.
A bit later, Next ported their OS to x86, and 32bits Sun and HP workstations, using the original MACH kernel.
There was also later OpenStep running on Solaris (a bit like this NextStep on AIX)
In another parallel universe, Apple released a Macintosh/MC68k emulator (“MAE”) for Sun Sparc workstations before the 68k emulator for PowerPC MACs.
https://www.youtube.com/watch?v=AL3ND_T4bQQ
http://temlib.org/pub/s.JPG
Lets not forget OpenStep Enterprise, an OpenStep implementation NeXT made for Windows NT.
Yep, there was a version of OpenStep that ran on NT. Also a version for the Rhapsody SDK too. Really not very useful, but they definitely exist – I have/had a copy somewhere in storage. The big kicker was that the UI needed to be defined for the NT operating system in a different nib file. If it was the same for all targets, that would have been a lot slicker. That and the fact about 3 or 4 additional services were needed to make the run-time function correctly.
I’d be very surprised if these didn’t still exist internally at Apple, forming the underpinning of iTunes/Safari for windows.
Not sure. I think that stuff was all Carbon wasn’t it?
The issue was that the services that ran, like the servers in BeOS, were on top of the OS. It’s fine if they exist on your target OS as a native layer, but installing they and maintaining them is not trivial. Stuff like .Net gets away with it as it is a vendor extension…
I don’t remember exactly how many there were, but I remember there was one called something like the “paste book server” or “pbserver” and another that did something to do with window management. I think there was a least one more. If they failed, the apps stop working.
Part of this framework was available for those who had purchased web objects to do web app development. Not only did it include the apple version for macs, but also the NT version. I had it running on Windows XP at one point. It included a very basic version of project builder.
GNUStep could have been the big chance for Linux on the desktop. They could have built a Mac clone and we would have a free desktop environment that really cut it.
http://etoileos.com/etoile/ was a very promising approach but late.
Unfortunately a lot of effort had gone into subpar GTK and Qt which eventually turned out to be rather a dead end.
Don’t know about GTK, but Qt is going strong.
GTK is still used for software running on Linux such as Firefox.
Or you could argue that it is better for the free software world to innovate on themselves and offering new alternatives instead of just cloning proprietary systems.
In general we have to admit that (apart from the hugely successful GNU) almost none of these clone projects succeed in the long term. They all suffered from never being on par with their proprietary sibling, and more often than not also lagging behind.
In contrast, while KDE and Gnome never found a strong market penetration, at least they are still relevant and thriving, especially KDE. Gnome was also adopted in commercial systems, while innovations from KDE influenced modern commercial desktops.
Edited 2017-02-03 13:57 UTC
Agreed. I was only kind of interested in gnustep to see if mac code could be run. It was cool, and it was kind of wine like where a very simple program might work, but most obviously didn’t. Plus the native applications for the environment were pretty poor quality when compared to the gnome/kde equivalents at the time ( maybe like 10 years a go, so think kde3 & gnome 2).
A lot more time and effort went into Gnome and KDE. If all that had been spent on GNUStep it might have come close to Cocoa.
Of course it would never cover 100% of the API – by defintion, if you are catching up. But that is not the point. If there is enough functionality there to build a professional application the primary goal has been reached.
GTK has practically stalled in development and Qt has gone astray with Qt Quick. Both have become bloated but not better over the last decade.
GNUStep was always very incomplete and was chasing the OPENSTEP spec back int he day, never really got there. They Apple started to deviate from the spec with OS X and Cocoa, and they literally had almost zero chance of every getting to a workable compromise between OPENSTEP and Cocoa. A lot like ReactOS/Wine, they are forever playing catch up. Haiku has the benefit that the OS is no longer a moving target, and the pace is still glacial in many respects.
There was also Cocotron http://stackoverflow.com/questions/tagged/cocotron . That looked promising for a while, but I have no idea what is happening with that anymore. I managed to get it to work and could run apps under Wine on a Mac that were compiled using it.
There’s also this: https://github.com/Microsoft/WinObjC which with work might bootstrap Mac apps on other OS.
This “big chance” of cloning a system with tiny market share does not look that big.
10 years after I started using it and GNUstep still hasn’t got a web browser. The single most crucial component of any desktop OS and it’s completely MIA. There isn’t even a wrapped browser so you can at least alt-tab between your gnustep applications and the browser and have a somewhat coherent desktop. This is why GNUstep isn’t used by anyone. As an API it’s fine (although there are many apis competing with it for attention). But as a desktop it’s MIA so noone uses it as it’s easier to just use GNOME and learn GTK or use KDE and learn QT. The only use people have for GNUstep is trying to relive the glory days of NEXT which is a shame since GNUstep can be put into OSX mode which makes the horizontal Apple menu bar, or put into Windows mode which makes it work like Windows 9x. The lack of apps is killing GNUstep.
In my experience, they’re also weird about upstream patches.
Most of the guys are pretty friendly though.
Hello,
Thanks for this article. This is a great share. In addition, a book was published a few years back named, “iFailed.” The link is below for the text on Amazon. I too was interested in learning more about the IBM deal as the author had made reference to the agreement but did not provide more details. Maybe the author (if contacted) would be able to provide more insight into this topic?
https://www.amazon.com/iFailed-true-inside-story-NeXT-ebook/dp/B007L…
thanks,
Matt
Speaking of rare operating systems, i have a system running Solaris/PPC here, but have had no luck getting hold of a working compiler for it so it’s fairly useless.
What about an old version of GCC? I’m sure you can get an older copy cross-compiled if you were determined enough.