alt Hacker NewsSome former Be employees have commented here in the past saying that Apple made the right decision to buy Next and not Be from a technical perspective, as BeOS had poor code quality and was far behind NEXTSTEP in important areas like networking and printing. They also implemented the OS API directly in C++, which tied them to a specific C++ compiler (some mid-90s version of GCC) due to the ABI getting broken in newer versions. The 32-bit version of Haiku (modern BeOS reimplementation) supports BeOS software, and as a result still needs to be compiled with that mid-90s GCC release.
Beyond that, I'm not sure if buying Be would have fixed the managerial problems at Apple. Would they have been able to add the required features to BeOS to turn it into a production OS without the project turning into another Copland? Would the existing Apple management have slimmed down the massive Apple product line that had dozens of nearly identical models? Would it have gotten rid of dead-end projects like Newton and Pippin? Etc.
Replies
Former Be person here. Excellent summary. To further elucidate the management issue, some of the Be management went to Palm and they weren’t super effective there either. A fun alternative universe would be Apple buying Be, Apple going bankrupt or eaten by the clones, NeXT powering the internet with WebObjects, Microsoft crushed by antitrust and we are all running Linux using Eazel/GNOME.
Pre-standard C++ is an even worse programming language than C++ 98, but also the big difference wouldn't have been the technology, it was the people. If Apple buy Be Inc they're getting JLG back whereas if they buy NeXT they get Jobs back.
Jobs is what makes or breaks Apple in that critical period. I didn't like him but that's relevant to the success of the company. JLG did not have whatever it was Jobs gave the company.
I recall the C++ "fragile base class" problem already starting to rear its ugly head by the end. Existing base classes couldn't be reasonably extended without breaking compatibility with existing applications, and the BeOS API was routinely designed in a way that promoted subclassing. e.g. to implement a window, you'd subclass BWindow and fill it with your own subclasses of BView. It's difficult to see how they would have dug themselves out of this hole in the long term. Perhaps eventually they would have needed to ship multiple versions of their system libraries corresponding to each ABI breakage.
One minor correction: the Haiku developers worked some magic with GCC binary compatibility and modern Haiku builds are done with a modern GCC yet still preserve the ability to run and interoperate with software built using the old release which is important for preserving the use of commercial software built for BeOS without released source code.
It was reported that Apple offered $75M for Be and almost $500M for NeXT and in retrospect those seem like entirely fair valuations. One was a mature OS and the other was a demo of an OS.
> They also implemented the OS API directly in C++, which tied them to a specific C++ compiler
We paid Cygnus(?) millions for that. The multiple virtual inheritance needed for MediaKit was a real joy to get working.
I'll note that Mac OS X used a C++ ABI in the _kernel_....
Best example of how far BeOs fell short was that for a very long while it lacked printer drivers of any kind --- by contrast, not only did NeXTstep essentially get printing "for free" since it used Display PostScript, the implementation of that was such that Adobe demanded a modification to the user agreement limiting the output resolution so that print shops would not be able to purchase NeXT computers and use them as RIPs for imagesetters.