In '97, the OpenStep API was hardly being used at all. Objective C was considered dead. I think it was pretty even - both NeXT and BeOS had terrible third-party support at the time. Adobe ditched NeXT early on; I don't think Be EVER attraced any major developers. (I don't consider Opera a major developer )) I think there was an X server for BeOS, wasn't there? If not, writing one would have been pretty easy.) rather than Be was the better choice. I believe this for one reason: Steve Jobs. In my opinion, if Apple had not gotten Steve back, it would no longer be in business today. Gil Amelio was in the process of running Apple into the ground by licensing Mac hardware to clone makers who built faster Macs than Apple did and sold them for lower prices. If Apple had not bought either NeXT or Be, Amelio would have been fired several months later for his poor performance and replaced by another braindead Rick Belluzo-esque "tech whiz" who would have finished the job.
Isn't OS X a proprietary Unix-like OS? Excepting Darwin, of course, but I'm not sure how that's relevant. As long as you can write software for it that works, isn't that all that matters? It's not like OS X users regularly hack their kernels.
Gasse't = Gassé? If so, then are you speaking of the time long ago in the '80s while he still worked at Apple? He played no part in allowing Mac clones, since he had nothing to do with Apple after the late '80s. The guy who ****ed up Apple in the mid-'90s by doing that was Gil Amelio. (He is not the only guy who played a part in ****ing up Apple, but he was significant.) Apple buying Be would not necessarily have made Gassé Apple's CEO anyway.
Yeah, but as it played out, Nextstep eventually became as slow as it is today. I was only comparing what might have happened with what really did happen.
That's very true. BeOS could have been further Unixized... maybe. I guess close Unix compatibility is a nice thing to have, even when you're talking about a situation where it's not really seen as a big advantage (the desktop market).
Originally posted by alex_ant
I don't think BeOS was any more different than OpenStep. Both were great, new object-oriented software infrastructures. OpenStep's was a bit better but that's not to say Be's wasn't far superior to Mac OS's or Windows' as well.
But NeXT was better in that it had near full POSIX compliance and was setup much more like a traditional Unix for which there were many more developers.
Well, Apple was in the market for a desktop OS, not a server OS. OpenStep was never a great server OS anyway. (And still isn't, in its current form) BeOS was far faster, and at the time had much more robust audio & video subsystems than OpenStep did - which should have appealed to some of Apple's most die-hard markets. BeOS had some problems (networking, printing, incomplete POSIX compliance [if you consider that a problem]), but none that couldn't be fixed.
The claim of "faster" can be a bit misleading. Latency was practically non-existant on Be. NeXT had latency but basic tasks, such as copying files and other hardware related things can't be sped up much by even the fastest OS. Sure, if one person is playing a first person shooter and serving files simultaneously, then file-sharing is slow. But file sharing on its own is something OS X and NeXT are good at. Just like all other Unix's:stable and good at handling mutliple things at once.
Except that OS X and NeXT use Mach, which isn't quite as fast as some other kernels. Again: not dramatically slower, but slower none-the-less.
And who is to say Apple didn't consider the OS's use as a server?? For all you know that could have been a selling point for a new OS. They might have seen it as a potential future serving OS. I think you're being a bit presumptuous.
Wow... Mail and OmniGroup's apps
Yeah, but NeXT was bigger and arguable more developed in its heyday. I know of many companies (including Swiss Bank) that invested in heavily and continued to use the technology because of Unix compliance.
Also, apps like WebObjects (and the NeXT customers history of using them) made NeXT a more attractive serving environment. Could have made a difference.
In '97, I don't think the ability to easily port Unix software to the new Mac OS was much of a consideration, since the open-source movement was MUCH less significant then than it is today. (And that's saying a lot, considering it's not even very significant today, with a few exceptions.
Sure it was easy in 97. Not all of the big apps had been ported but they were being worked on. MkLinux had been around for a while and so had LinuxPPC (in fact that was a time when the two were via-ing for "market share" and there was an unsuccessful push to get mklinux on PPC.
Furthermore, porting software is a snap between LinuxPPC (or any Linux) and OS X. Its almost always just a recompile away. The same can NOT be said of BeOS.
Listen, I think BeOS is a great OS and should be picked up by someone, someday. But whether you want to believe it or not, I think its being too different might have been why Apple didn't pick it. It innovated and strayed away from what a lot of people were going into at the time--namely a *nix style OS (like Linux, PPX Linux, MkLinux, and other ports like FreeBDS and the rest). Unix was the trend of the time. BeOS was and is very promising, but it was different and uncharted territories. NeXT represented safe and popular territory and offered everything Apple was looking for, already tried and looked at.
I think it was the right choice by Apple and I hope BeOS isn't dead and forgotten.
Taft
I was under the impression that Apple Quartzized OpenStep because they wanted to replace Display PostScript with something for which they didn't have to pay royalties. In doing so, they were forced to completely rewrite the display layer. BeOS's display layer was not proprietary, so I'm presuming that there would have been no inclination on Apple's part to change it substantially.Originally posted by arn
But that's what I'm saying... it's not that NeXTStep was slow either... it's that Apple Aquaized/Quartzized it...
arn
) Amelio was monumentally stupid, but I doubt he was that stupid. (Although I could be wrong, and I guess I wouldn't be surprised if he were.)
Originally posted by alex_ant
File sharing is really hardly dependent upon OS speed. You could be running Windows NT 4 on a 500MHz PII and as long as you've got Ultra160 RAID and the Pentium 4 round the corner doesn't, you'll be alright. Have you ever used BeOS? I have a 550MHz PC running BeOS and a 550MHz Mac running X.1.4, both with 256MB of RAM, and the speed difference is simply staggering. BeOS is faster in every way - all aspects of filesystem performance (reads, writes, throughput),
[/b]
I can't tell if we're agreeing or not. I guess my point is that a system reaches limitations on how fast it can be because of hardware. The OS can more efficienty use that hardware, but the speed increases (from my experience anyway) are nominal. I ran LinuxPPC, OS 9 and OS X for considerable amounts of time. Even though LinuxPPC was much faster overall, certain hardware dependant things like copying a file just didn't get the speed boost that other operations did.
I believe BeOS utilized a microkernel as well, which wouldn't explain kernel speed differences unless Mach is for some reason slower than it should be.
That was my point...a lot of people think that Mach is slow when compared to other kernels. This is why multithreading isn't as good as it could be and latency is higher. Be did some neat things at the kernel level.
I don't think I'm being presumptuous. In 1997, Apple was dying. I know the phrase "Apple is dying" is a cliché in Apple-land, but it really was true. They were hardly in a position to explore a market in which they had consistently been an utter joke for the past decade plus. They needed to regain a solid hold on their niches, not jump into a pool of sharks with IBM, Sun, HP, Microsoft, SGI etc., and I think they knew this, which would explain why Mac OS X Server in 1999 was barely marketed and used more as a technology demo instead.
You're right, they were dying. And they needed to regain what they had lost, which was the consumer market.
But you're still making the assumption that they weren't making a decision based on all the benefits a technology could offer. Unless they were running in panic mode and barely looking past how the OS would effect them in the short term, I submit that they must have considered the potential for the OS to move them into other business areas. Not doing so would just be bad business.
NeXT's heyday was in like 1991. Sure they managed to hold onto some small niches, but then again, so did OS/2 and Amiga. NeXT's technology was largely irrelevant by 1997. (It made a remarkable comeback, though, didn't it?
I think it extended a bit later than that...could be wrong though. Anyway, the point is that they were a technology that quite a few people picked up in their heyday. They were a known and tested entity with strong inroads to the Unix community.
...the promotion of an obscure, barely-used web development technology is probably not going to be my first priority.
Again, I don't see it as stupid. The internet was hot at the time. WebObjects offered something good, even at the time. Its about making rounded, forward looking business decisions that are able to look past what Apple had offered in the past.
Also, it doesn't matter the total number of Webobjects developers there were. The point is, did the people who used it like it? Was it a good technology? If so, it might be something we can sell, and help sell hardware. Get where I'm going??
That's true, but name me one open-source application circa 1997 for which there was significant demand on the Mac platform. Most applications cherished by Mac users were (and are) commercial and had no equivalents on any Unix platform. LinuxPPC and MkLinux were both irrelevant. If you can get Samba and Apache working on OpenStep out of the box that's great, but if it takes a little porting work, I don't see what the big deal is, since there was no real clamoring at the time for the ability to run any open-source software. (Even today, the demand for this I'm guessing is probably quite small, and does not originate from the typical Mac devotee crowd.) And POSIX-izing BeOS would have been an option, if the need were really there.
Again, I strongly disagree. It doesn't matter if Mac users were clamoring to have all of the open source apps. SOMEBODY wanted the apps ported and they were ported. I was part of the LinuxPPC community then. They were doing good things and they continue to do so. Its a good community that, IMO, forwarded the cause of Unix on a Mac and Unix on PPC in general.
My point is that a few experiments into running Unix on PPC had already been run. And they turned out pretty successfully. Whether or not people were using them, solid Unix style OS's existed and there were apps for them.
And contrary to what you believe, I think Linux on PPC *was* relevant back then. That was during the whole Linux on everything push which has subsequently subsided. I think this could've played into Apple's decision. The fact that many companies were going Linux or Unix and home users were starting to pick it up as well...Apple could've been following current trends of the time.
In 1997, Linux was irrelevant in the Mac market. LinuxPPC and MkLinux were even more irrelevant. They're still irrelevant today, in fact. (On the desktop and in the customary Mac market, that is.) Apple could survive without GPL software without any problems whatsoever. BSD software would still be there for the taking, so there would be no problems. Microsoft nicks BSD software for use in Windows all the time - no reason Apple couldn't or wouldn't do the same. I don't think Unix was a trend in 1997 any more than Windows NT was. Right now, in 2002, it's been clarified that Windows blows on the server, and so the hype surrounding it has died off a bit. But NT was actually growing quite rapidly in 1997. That said, I don't think there was a general trend in '97, unless you consider the desktop, where the trend would be that Apple was close to death and Windows was moving from 95% to 97% market share.
Well, I guess we disagree on a number of points. Oh well. This thread has been pretty lively hasn't it? I like these kind of discussions.
Taft
