Raptor Engineering Helping To Improve POWER Support In Wine, Eyes Hangover

[edwaleni]

As for all that historical narrative on IBM blunders, and baffoonery, of course a lot of strategic mistakes were made as the micro market took off. Geez, so did Digital (Rainbow), so did Apple (Lisa), so did NCR, and so on and so on.

IBM ESD in Boca Raton (where the PC and PC-XT were developed) is a classic textbook example of what happens when corporate knowledge & skill comes together with little interference.

The baffoonery started when long time IBM clients wanted micro platforms with the similar types of predictability and supportability as the System 36xx/46xx/System 34 they use. IBM ESD was then folded back in under new executives who only knew "old Blue" ways. That is when the hubris started.

If ESD had been left to its own devices, they would have worked with Compaq and a few others to develop open platform standards, but after the AT, ESD got their wings clipped. That is what got us MCA, EISA, VL-Bus and some others. It wasn't until the PCI standards came along, the segmentation stopped. IBM's cost plus model of micro pricing made them uncompetitive in the public marketplace on purpose. They wanted the enterprise, not consumer.

I supported IBM PS/2's and Compact Deskpro's in their prime. They were excellent platforms for a purpose and could do things many people weren't' aware of that we take for granted today. They were platforms for their time, but what they meant then has no bearing on the needs of today.

As for OS/2 on POWER, it was not any kind of financial drain on IBM at all. It wasn't even strategic. Ellen Hancock spent way more pushing ATM25 as a token ring replacement than IBM did on OS/2 POWER. OS/2 for POWER actually came up in a customer focus session initially and it was taken back as a POC. IBM approached me to see if we had an interest in merging the POWER 615 with OS/2 to run Windows applications using the WoW (Windows on Windows) capability. Our answer (as well as other large corporates) was no. The train had left the station and they couldn't go back.

IBM just couldn't provide the level of engineering and support with the margins x86 micros were demanding. That is why they sold it all off to Lenovo.

As for patent trolling by IBM, I don't buy it. (As an example) The coding scheme for your music compact disc was developed by IBM back in 1956. The Sony-Philips consortium had no problems negotiating a royalty agreement with IBM to use it.

I haven't heard of anything in recent times where they are cracking peoples nuts over some IP.

[Djhg2000]

See Michael's write up on the Libre Graphics effort using RISC-V and Vulkan.

https://www.phoronix.com/scan.php?pa...-RISC-V-Vulkan

Looks more like a proof of concept to me. Let's face it, $250k is peanuts compared to what either of the three major PC players are spending on R&D.

[edwaleni]

Considering what known "bad guys" do (Microsoft, Oracle, Intel, AMD following Intel's footsteps), IBM does sure look good though.

Those "bad ole days" of IBM arrogance is pretty much behind us now. There are still a few groups who feel pretty good about themselves but they are a minority from my perspective.

But the company overall has turned over so much and so often in the last 25 years that many of those "old Blue" employees have retired or died.

I am excited about the OpenPOWER initiative because IBM sees the open-source light. With such weakness and micro-segmentation of ARM these days, a strong ISA is needed to counter x86.

[Vistaus]

It depends on your definition of "fine". According to Intel's documentation the firmwares enable more features/performance and power saving technologies.

It's probably a less severe version of the nouveau nVidia driver situation - it will "work" but without reclocking so you get terrible performance, for example.

My definition of "fine" is "great performance" and "good battery life" and "very stable". And those all apply. But I will try the tips from the Arch Wiki article to see if things improve even more.

[ynari]

The immediate reason OS/2 failed was that IBM invested in OS/2 for PowerPC rather than OS/2 x86, and mismanaged their relationship with software vendors that could have created a solid application base. Win16 compatibility didn't help, either, particularly when a number of OS/2 versions were less functional than their Windows alternatives.

The longer term reason OS/2 failed/would have failed anyway was poor architectural decisions. The kernel remained (still is, in Arca OS) almost entirely 16 bit and single user. Not to mention Presentation Manager was not overhauled, it got a 32 bit GDI in SP1 of OS/2 2.0, 32 bit Windowing had to wait till Warp 3. NT did not make these mistakes, and managed to create win32 with an extremely high level of win16 compatibility, then it largely managed it again when x64 versions of Windows were released.

There were other reasons too (poor install, poor vendor preloads, drivers, memory requirements, graphics origin being opposite to Windows) but they probably weren't as much of a factor.

WINE/WinOS/2/other mostly solid emulation options can be a mixed blessing. Whilst there was a solid contributing community of OS/2 users, a large proportion of the user community expected a higher quality Windows like marketplace. This was not realistic : some of the best OS/2 applications were from small companies, the expectation should be to pay more (than Windows) for a lesser amount of functionality, but at a higher quality. The distinct danger is that users don't accept this and opt for Windows apps under emulation.

Linux is in a better position than OS/2, because it's open source and doesn't have to succeed. However, if you want companies contributing to Linux you need to buy their stuff..

[rmoog]

Absolute heroes!

[jacob]

Economic textbooks? It's not like those practices didn't exist in the physical product world already.

They sure did, but in the computer industry (both hardware and software) IBM has always been, and remains, the world's dirtiest player and the world's largest patent troll.

Because of the usual IBM comical mismanagement and bureaucracy. Which is the actual IBM's trademark.

I don't think you can accuse IBM of mismanagement. Overly burdensome and bureaucratic management maybe, but generally whatever they do, they do it competently and forcefully. The exception would be the PC Jr, that really was a tragicomical tale of mismanagement and moronic execution. But it remains an exception in IBM's history. They had many more famous duds (the PS/2, OS/2, the PowerPC to name just a few) but those projects died by alienating IBM's partners and clients alike, not because the products themselves were inherently bad (which is not to say that they were necessarily brilliant either).

The thing where SCO accused IBM of stealing their UNIX code during a failed joint venture they did with IBM and contributing it to Linux?

Wasn't that a cash grab from the start? Imho there was very little "certain IBM breach of contract" there and someone really had to prove that claim.

It has been many years but IIRC it went something like this. IBM got SCO to sign up for Project Monterrey, which was to create a new version of UNIX for the Itanium platform. Then shortly after that, IBM changed their mind, lost interest in creating any new UNIX, killed off the project and decided to push Linux instead. SCO was furious because not only IBM let them down, but in fact started investing big time in Linux, a product that was competing against SCO's own Unixware. In a nutshell that's all it should ever have been about. But then McBride got the idea that if he could "sell" it to the public and the courts as if SCO "owned" UNIX and IBM was "stealing" it and giving it to Linux, none of which was ever remotely true, he could yet turn it into a money making scam. The rest is history, as the cliche goes.

[starshipeleven]

Where do you think Oracle, Intel, AMD and especially Microsoft - particularly old Microsoft of the Gates-Ballmer era, the worst kind - learnt their tricks?

Economic textbooks? It's not like those practices didn't exist in the physical product world already.

Why did OS/2 fail?

Because of the usual IBM comical mismanagement and bureaucracy. Which is the actual IBM's trademark.

Also BIOS and the IBM modular PC failed for the same reason. Comical mismanagement caused both to spin off and become a major world-changing thing outside of IBM's original dumb project.

And remember the SCO Affair?

The thing where SCO accused IBM of stealing their UNIX code during a failed joint venture they did with IBM and contributing it to Linux?

Wasn't that a cash grab from the start? Imho there was very little "certain IBM breach of contract" there and someone really had to prove that claim.

[jacob]

Considering what known "bad guys" do (Microsoft, Oracle, Intel, AMD following Intel's footsteps), IBM does sure look good though.

Where do you think Oracle, Intel, AMD and especially Microsoft - particularly old Microsoft of the Gates-Ballmer era, the worst kind - learnt their tricks? FUD, Embrace & Extend, client lock-in and extortion, patent trolling, deliberate anti-features, they all have the IBM trademark all over them. IBM started doing it all before everyone else and they never stopped.

Why did OS/2 fail? It was not Microsoft's fault, but IBM's. IBM insisted that all OS/2 should do was to allow vertical IBM applications to connect to an IBM mainframe (from an IBM PS/2 desktop) and not much else.

And remember the SCO Affair? At the heart of the case, which didn't really have anything to do with Linux at all, was a breach of contract on IBM's part. If SCO stack with that, they would have had a legitimate case and yes, they were wronged by IBM, but instead of seeking any sensible resolution, IBM moved to simply crush them. SCO then sealed its fate by embarking on a demented plan to try to parasite off Linux by spreading BS about "stolen" code and Steve Ballmer's conspiracy theories about a bunch of evil communist coders who are out to destroy capitalism. But Incredible as it sounds, IBM were the evil guys even against SCO.

[starshipeleven]

Remember, IBM aren't the "good guys". They never were.

Considering what known "bad guys" do (Microsoft, Oracle, Intel, AMD following Intel's footsteps), IBM does sure look good though.