What if Oracle hadn't bought Sun ?

JWJones said:
And basing MacOS X on a robust, mature UNIX-like OS was definitely a better move than the immature BeOS, as cool as it was.
Click to expand...
Whoewer decision this was, this is definitely what shaped Apple to become the giant behemoth that is today, for better or for worse. One of the most important decision in the IT world in the past fifthy years of so for sure.
 
JWJones said:
Yeah, my thoughts exactly. They absolutely needed Jobs to come back when he did, they were teetering on the brink of oblivion, with no vision and a muddled product line. And basing MacOS X on a robust, mature UNIX-like OS was definitely a better move than the immature BeOS, as cool as it was.
Click to expand...
I’ll sound like beating my own drum again, but IMHO one of crucial factors for OS X success was combining ColorSync with Next’s Display PostScript, which they morphed into Quartz 2D which was PDF based, so no more fees to Adobe and happy customers, and a lot of customers then were from the printing industry.

Be had nothing like that. Be had cool multimedia, okay, but who would pay almost Unix WS prices to have decent ROI soon after - for multimedia only in ‘97?

Major market for Apple were pro customers (companies) for almost whole decade, before Jobs started pushing back into consumer market with colorful iMacs and iBooks.
 
Jobs really had the whole thing already developed and a proof-of-concept machine already launched with NeXT when Apple brought him back in, so he had a real story to tell and a code base already in place. The man knew what he wanted. NeXT was the prototype OS-X, to a large degree. I remember seeing a demo of the NeXT cube when it was launched and being pretty impressed by it, it was good stuff.

From https://en.wikipedia.org/wiki/NeXT

"Apple merged with NeXT in 1997 as part of a $427 million deal, including 1.5 million shares of Apple stock. The deal appointed Steve Jobs, then the chairman and CEO of NeXT, to an advisory role at Apple; and OPENSTEP for Mach was combined with the classic Mac OS, to create Rhapsody and Mac OS X."

BeOS was always an outsider, a kind of incompatible copy of windows, they never really had much of a chance at being chosen as Apple's desktop OS. And Jobs definitely deserves full credit as a visionary leader, regardless of what he was like to actually work with. He saved Apple's backside, for sure. I just wish I'd loaded up on Apple shares back then, I'd be living a life of luxury now!

I think if Apple had chosen BeOS instead, they likely wouldn't have got anywhere near as far as they have done, and perhaps would have continued down the plughole. BeOS never got any major traction by itself. As for the BeOS guy demanding a lot of money... I suspect that was only one factor in the decision. Someone made the right decision at the time, for sure.

And... it has to be said, Apple was always Steve Jobs's company, right from the start. We have a couple of sayings, "blood is thicker than water", and "better the devil you know". In other words, Jobs was a known quantity, and he had proved himself capable of delivering the goods in the past; whereas the other guy... not so much.
 
sidetone said:
RiscV is 64bit. Risc is 32bit. I want to see x86 replaced largely by Risc 32bit and Arm 32 bit.

I wish the Solaris or Sun name would be sold or given up to those Illuminos projects. Then, I'd like to see a reusable more universal variant of CDDL1.1 for these projects. OpenOffice was given to Apache Foundation by Oracle.
Click to expand...

That would be good. Oracle did a fair bit of kernel development for Solaris 11.0 (roughly when they closed down OpenSolaris) and later. illumos development seems to be mainly network and disk drivers with an acceptable (though often dated) userland. I don't think that Oracle are doing much other than contracted bug fixes these days.
sidetone said:
And that Sparc gets revived to work alongside RiscV, even for specialized purposes. RiscV is owned by nonprofit RISC-V International. Sparc is owned by a for profit, Sparc International.
Click to expand...
I'm not sure that is a good idea. SPARC was ahead of the pack when it first came out. The register windows were a big advantage over x86 and 68k with only around 8 registers. These days 'registers' are mostly a fiction with the CPU doing register renaming. The AMD Zen 5 now has hundreds of registers, as does ARM Neoverse and I presume the latest Intel CPUs. As I understand it the SPARC register windowing is a kind of explicit register renaming. If you change that you would need to adapt compilers and recompile code to use the extra registers. AMD and Intel with their implicit register renaming don't have that problem. They have been quietly adding more and more registers in the background without requiring any code changes. So what used to be a weakness for Intel and AMD has turned out to be an advantage.

The SPARC architecture _is_ open source. Other than the radhard LEON I've not heard of anyone making use of it.
 
It seems that Fujitsu's SPARC64, mainframe, and supercomputer processors share a common architecture.
The former supercomputer Kei (ranked number one on the TOP500 at the time) was an extension of SPARC64.
Its successor, Fugaku (ranked number one on the TOP500 at the time), adopted ARMv8 and added extended instructions.
 
I don't think x86 is dead at all, and I don't see why x86 dying would be a "good thing" anyway. Suppose sparc was resurrected, say some company launched a new line of sparc chips tomorrow. So what? Who's going to use it, and why? About the only realistic place I can think of is large multi-core processors for servers, and there are already many other highly competitive players in that arena. Are desktop PC's or laptops going to use it? Seems very unlikely, they never did in the past when sparc was around. Mobile? Even more unlikely, ARM is totally entrenched in mobile. Unix workstations like Sun used to make? That market is stone cold dead, linux on x86 killed that, there is no money to be made there. Embedded? Perhaps, but there is a ton of competition in that space, particularly from ARM, and other very successful companies like microchip; It would be a very hard sell to launch sparc as a new architecture for embedded. Games consoles? Nah. Automotive? Aerospace? See 'embedded'.

Will ARM or risc-V knock out x86? Again, seems very unlikely, at least at present. AMD are doing very well in x86, even if intel are slipping, and it's easy to forget that intel are still shipping an absolute shed-load of x86 product from their fabs. The wildcard would be if the reported Chinese push to develop risc-V results in chips that are as powerful as current x86 at 1/20 the price, and floods the market with them in order to destroy the western industry completely. That could actually happen, IF they can crack the technology, IF they throw enough billions of dollars at it (for comparison the CCP gave out $231 billion in subsidies for their EV program over 15 years, now you know why no western company could compete on EV's) and IF western governments allow their markets to be flooded and are prepared to lose yet another major strategic industry. How likely is that? How do you feel about being ruled by China?

But I don't see where a resurrected sparc is going to fit in; the processor market is already extremely competitive with many very competent players already delivering high quality products. For a new sparc chip to win in that market, it would have to have some very special technical superiority, like a markedly better power envelope, or substantially higher performance than what is already on the market. For example the kind of technical superiority that Apple demonstrated with the M1 when they launched it; although the competition is arleady catching up with them fast. How realistic is it for a putative sparc vendor to achieve that? Apple spent a fortune developing the M1, but they had the money to do it.

IBM currently has all the same dilemmas with POWER and the open power initiative. Another architecture that is struggling to survive, although at least it's still there, and they have the major captive market of the IBM mainframe to keep them going. Whereas sparc was sunsetted years ago.
 
I guess the other thing that could happen to threaten x86's dominance is if there are more snapdragon-X type chips put onto the market to try to take some of x86's market share in the laptop and low power desktop space. Since Apple don't sell the M1-series as merchant semiconductors, their devices are off the table, so it's up to competitors like qualcomm to test the market. So it's possible that some fraction of the x86 laptop market in particular could be taken by these latest ARM chips; it's certainly taken them long enough, they've been trying to get ARM onto the desktop for at least 10 years.

But then you are up against questions of software compatibilty, and whether consumers are content to replace an open architecture with a closed one (many probably won't notice), and above all the level of support from Microsoft and other large software vendors will be critically important. So I don't think it's a sure thing at all. And if intel get their act together (admittedly, that appears to be a big ask!) they can retain a large part of the market, probably the majority share. In the low power desktop/laptop space intel already have some very nice offerings, like the N100 I'm using currently, and the later chips they have replaced it with. So I wouldn't count intel or x86 out at this stage.
 
blackbird9 said:
Are desktop PC's or laptops going to use it? Seems very unlikely, they never did in the past when sparc was around.
Click to expand...
IMG_20200610_221812-scaled.jpg

TOSHIBA AS1000/C80 (SuperSPARC II 75MHz)
 
Sure, there were a few. There was a korean company, I can't remember the name, that were selling sparc laptops with fujitsu sparc processors in them, here in the UK. There was an IBM thinkpad with a powerpc processor in it too at one stage, I seem to remember it ran AIX. They never achieved any more than a very small fraction of market share and were discontinued by the companies that made them. In the final analysis they were no better than the incumbent x86 stuff and came with lots of problems around software compatibility.

Great photo, BTW! That case looks suspiciously like a Tosh T3200... :)
 
Tadpole was a british company that launched a similar sparc laptop, I remember those, I almost bought one at one point, but it was ridiculously expensive. I think their sparc machine only lasted a few years. Looks like they made a powerpc one as well.

en.wikipedia.org

Tadpole Computer - Wikipedia

en.wikipedia.org en.wikipedia.org
 
Look at this thing again. This is the competition you have to beat. For $112 retail(!) you get 8GB RAM, 256GB SSD, a 64 bit cpu that can support three 4K monitors, bluetooth, wifi, wired ethernet, lots of usb ports, sound card, etc, etc. It's fashionable to knock Intel, but you have to admit that the N100 chipset in this machine is extremely competitive. It comes with an industry standard O/S that has thousands of applications already available, including the ones you already paid for. This little PC is probably good enough for 90% of all consumer and corporate desktops worldwide.

It's an extremely competitive market. How are you going to make any money with a new processor architecture? That is the barrier you have to climb. It's a very difficult proposition. Apple managed it because they can afford to put a large amount of money down to develop the product and they have tight control of their customer base who are prepared to pay a substantial premium for apple products. I personally believe the snapdragon is likely to have a more difficult time getting any substantial share of the market, but that remains to be seen. Why would you as a customer buy an incompatible architecture machine when the x86 one on the shelf next to it in the store is about the same price and runs just about as fast and is compatible with all your existing software? Unless there's some really compelling reason to buy the snapdragon. Well, wait and see what happens, as always.

If it retails for $112, whatever is the factory gate price? That retail price is the aliexpress price so the usual 10 to 1 factor probably doesn't apply, but both the reseller and the aliexpress platform itself must be taking a cut. Then you have a power supply, the packaging (they come in a nice box), some connecting cables that come with it... so what is the actual build cost for the machine? And then what is the cost of the intel SoC? They must sell a lot of them to make any real return on investment, that's all I can think. It would be interesting to see the manufacturer's BOM for this box. Honey, they shrunk my PC :)

1758300231539.png
 
Charlie_ said:
IMG_20200610_221812-scaled.jpg

TOSHIBA AS1000/C80 (SuperSPARC II 75MHz)
Click to expand...
AFAIK AS1000 was the first SPARC laptop, right? Then Tadpole made few models during ‘90s, and latest that I know of was NatureTech 777 GenialStation (~2001/2), with SUN UltraSPARC IIe @ 500 MHz w/256-KB L2 Cache; 15.0" TFT SXGA LCD Panel; 256MB ECC RAM; 80GB IDE disk

Article about release: PC maker ships Sun-based workstation

NatureTech 777 GenialStation, image © Gábor SAMU 2021 (@gabor_samu on X, image copied from his X acc)
FB2BKEwWQAMrJzn.jpg
 
Ha, interesting, not heard of those before. It looks like Sun had quite a good try at getting sparc onto laptops back then. I'm trying to remember the name of the Korean company that were selling them here... I can't remember, it was a long time ago, that was back around 1990 or 1991.

I like the Eizo monitor behind it too, I had a couple of really nice Eizo monitors many years ago, they were very high spec :)
 
blackbird9 said:
I like the Eizo monitor behind it too, I had a couple of really nice Eizo monitors many years ago, they were very high spec :)
Click to expand...
Since Barco stopped to make Personal/Reference Calibrator monitors, IMHO Eizo is the best $ can buy for color accuracy.

BTW, there are still Sparc based notebooks made today, but not exactly with Sparc CPU, Elbrus is based on that tech but diverged from original. Although, all I can find look more like Panasonic Toughbook, I doubt that they are available for civilians:

TS STRONG @MASTER 7020.Р (Тайфун) (Эльбрус-8C CPU; 1.3GHz; 8 cores per CPU, 4 CPU per module; DDR3 ECC)

33e26026614debc7.webp
 
Ah, really interesting, I remember reading about elbus years ago. I guess this is for battlefield/military/industrial applications? It looks highly ruggedised, it looks like an all-metal case.. I wonder what operating system they run? Yeah... not available for civilians...! Well, I guess those cost a lot of money... maybe you can drive a tank over it... :)

I guess that one does not have intel ME 😁

So, of course, if you are talking of special equipment, then there is a market for non-standard CPU, sparc, power, MIPS... But this is not the mass market :)
 
blackbird9 said:
Ah, really interesting, I remember reading about elbus years ago. I guess this is for battlefield/military/industrial applications? It looks highly ruggedised, it looks like an all-metal case.. I wonder what operating system they run?
Click to expand...
AFAIK Elbrus Linux, Astra Linux and Alt Linux and for some models Protected real-time operating system Neutrino-E

Here in #26, I posted link where you can see bunch of Elbrus workstations and servers
vmisev said:
I was just reading about MCST today – I was under wrong impression that they based Elbrus on Vrlogs that I mentioned, but Elbrus 2000 arch predates open sourcing of T1/T2 for a number of years, so I guess they must have some licensing deal with Sun in the late '90s?

Here is a page with a range of current offerings, List of Russian computers based on Elbrus (E2 K) architecture processors, (Russian only, but page is simple enough that auto translate is not bad), but no prices – I found few online dealers offering desktops, but price is on request only. In their online classifieds Elbrus desktops (older models) are quite cheap.

Afaik, China based their efforts on MIPS – Loongson, and newer versions even have fast x86 and ARM binary translation, but IDK much about them.
Click to expand...
 
I'im wondering about EMP protection... (!) But perhaps its better if we don't talk of these things :oops: :'‑(

Hmm, elbrus has a large range or machines. I remember, maybe around year 2000, Elbus was mentioned in the western press. At that time, they thought to sell advanced CPU's to the west, I think it was original russian design, before they licensed sparc. But maybe my memory is mistaken, it's quite a long time ago. I remember reading some press about it back then...
 
blackbird9 said:
Hmm, elbrus has a large range or machines. I remember, maybe around year 2000, Elbus was mentioned in western press. At that time, they thought to sell advanced CPU's to the west, I think it was original russian design, before they licensed sparc. But maybe my memory is mistaken, it's quite a long time ago. I remember reading some press about it back then...
Click to expand...
Well, MCST is acronym for Moscow Center of SPARC Technologies, and they are spin-off of Lebedev Institute of Precision Mechanics and Computer Engineering which was first to develop Elbrus brand which had few different instruction set architectures. MCST continued to use Elbrus brand both for evolution of original Elbrus 3 and 2K and for SPARC, so it’s not so easy to track down what’s what. For example, Elbrus-90 is definitely based on SPARC, yet it runs S-400. Go figure?
 
Yes, that's what I remember reading. There was some hype around that time that super-cpus from russia were going to come to the west. It never happened, of course. The name elbrus stuck in my memory. Well, it says they got some western patents, and some similarity to IA64, so perhaps some of their ideas were used in western designs. That paper says they didn't have the process technology though, so basically they couldn't make them in Russia, although they had some architecture. So, who knows what happened. It looks like they licensed sparc fairly soon after that anyway.

You can tell how long ago that was because the paper talks about IBM as a major chip manufacturer. Ie, before they sold their fabs off to global foundaries.
 
As I see it, biggest hurdles for them are twofold: on one hand development wasn't biggest priority because they could buy Intel CPUs and second, what they developed was produced by TSMC, from which they were banned. IDK for sure, but I assume that domestic development is priority since '22 and that they are counting on SMIC and Chinese made DUV Lithography machines, and looks like that they are progressing really fast.

A lot of news reported about this few days ago, but this is only one I found that's not behind paywall:
SMIC Said to Test Chinese-Made DUV Lithography Tool from SiCarrier Affiliate Amid AI Chip Push

It will probably take another year or two to make it large scale production ready, but one thing is for sure, they will get there, if nothing else, out of spite for banning them from buying new ASML machines.

So, this industry can get very interesting to watch over next few years... 🧐
 
Yeah from what I've read china is throwing billions of dollars at companies like SMIC and other chinese foundaries, to develop that strategic industry. The "cpu wars" are hotting up... time will tell, as always.
 
fmc000 said:
Whoewer decision this was, this is definitely what shaped Apple to become the giant behemoth that is today, for better or for worse. One of the most important decision in the IT world in the past fifthy years of so for sure.
Click to expand...
No, this is wrong. The choice of OS didn't matter too much here to make Apple the giant we do know today.

What really was changing the world is that with buying NeXT Apple also got Steve Jobs back. More importantly, Jobs became CEO of Apple again.

And in that position Jobs discontinued a lot of products, like Apple Newton and streamlined the rest of the line up.

Also Jobs realised one important thing: Macs need to appeal to coporate environments as well, and in order to do so Mac needs a native port of Microsoft Office. This is why he made a deal with Microsoft about it, even had a video call with Bill Gates in his keynote where he announced that. Many were not really happy about that controversial decision.

The first product Jobs launched then was the colorful iMac. Which still ran old MacOS, but was popular and brought in money back. Another important milestone was the launch of the iPod with iTunes.

New MacOS happened somewhere along the lines then. But it was this transformation from computer manufacturer only to music player devices and mobile phone producer that changed Apple big time.

It was mostly Jobs' decision making and priorities back then which changed Apple. Having NeXTStep was a nice bonus, but let's not forget FreeBSD already was around. So Apple could have also totally adopted FreeBSD instead.
 
You must log in or register to reply here.
Back
Top