Open Source Computing

Depends on how you measure "best". If you measure it by democratic means, namely market acceptance
For servers, Linux is unbeatable. On supercomputers (the Top500 list for example), it has 100% market share.

That should never be a mark of "best".

Even looking at Netflix: The bulk of Netflix runs on AWS, and the a lot of uses Linux.

On the business end, not serving their bread and butter video content delivery where FreeBSD rules.
 
My personal opinion as a LineageOS user - it is a nice try, but in practise it sucks. It's hacky, only geeks would do it. Updating is a major hassle. And all interesting Apps are on Google Play only, so you are kind of stuck with a very limited set of apps.
I used to have Lineage OS on my phone some time ago. It's very hard to use it without google services: no glovo, uber and bunch of other apps work.
Recently I found another project -https://lineage.microg.org/ It's the same(almost) LineageOS + it has microG - opensource(?) implementation of google services. You can enable location plugin of choice and there you go - all apps(that I tested) will work again: uber, glovo etc. The big plus from this is you may not login to use this services which grants some privacy. On the other hand you are getting fully(almost) functional Android phone.
Ans you can use F-Droid for opensource apps and Aurora for non-open-source. Aurora itself is opensource and can be installed through F-Droid.
From my experience it's the most balanced way to keep some privacy and do not struggle a lot in day to day use.
 
I used to have Lineage OS on my phone some time ago. It's very hard to use it without google services: no glovo, uber and bunch of other apps work.
Recently I found another project -https://lineage.microg.org/ It's the same(almost) LineageOS + it has microG - opensource(?) implementation of google services. You can enable location plugin of choice and there you go - all apps(that I tested) will work again: uber, glovo etc. The big plus from this is you may not login to use this services which grants some privacy. On the other hand you are getting fully(almost) functional Android phone.
Ans you can use F-Droid for opensource apps and Aurora for non-open-source. Aurora itself is opensource and can be installed through F-Droid.
From my experience it's the most balanced way to keep some privacy and do not struggle a lot in day to day use.
Good stuff! I did not know about microG and Replicant. However, I believe these are just hacks and only for IT specialists. The truly free experience will come soon when we have the complete Linux phone for the masses. I believe this will be coming soon. And then it cannot be turned back. Once you have the Linux phone porting FreeBSD on it should be straight forward.
 
and postmarketOS, and ... there are so many of these projects, yet they all share the same challenge: supported devices are very limited - you can't install it on *your* device (unless you happen to have a (old) device which is supported laying around).

Think about the business model for "consumer devices" - what are these companies making money on? Well, they sell you a new phone every two or three years! So if everybody keeps their phone for five or ten years, these companies will not survive.

So, the suggestion of the OP is "make money by competing and providing support services.". Let's look at it: who makes money by competing? Professional athletes and teams, nobody else. (Of course there is a broadcasting industry that makes money from broadcasting rights and selling advertising, but that is something else).

That leaves us with companies who makes money from "providing support services". Let's see, the big one is Red Hat and there are several smaller ones in different countries around the world doing the same thing as Red Hat on a local scale, this seems to be a working business model so far. But these companies provide *software* support, not hardware support.

Anyone who knows about a company that makes money from (open source) *hardware* support services, raise your hand.

And remember the key point of a consumer device: it has to have a low enough price that lots of people will buy it...
 
and postmarketOS, and ... there are so many of these projects, yet they all share the same challenge: supported devices are very limited - you can't install it on *your* device (unless you happen to have a (old) device which is supported laying around).

Think about the business model for "consumer devices" - what are these companies making money on? Well, they sell you a new phone every two or three years! So if everybody keeps their phone for five or ten years, these companies will not survive.

So, the suggestion of the OP is "make money by competing and providing support services.". Let's look at it: who makes money by competing? Professional athletes and teams, nobody else. (Of course there is a broadcasting industry that makes money from broadcasting rights and selling advertising, but that is something else).

That leaves us with companies who makes money from "providing support services". Let's see, the big one is Red Hat and there are several smaller ones in different countries around the world doing the same thing as Red Hat on a local scale, this seems to be a working business model so far. But these companies provide *software* support, not hardware support.

Anyone who knows about a company that makes money from (open source) *hardware* support services, raise your hand.

And remember the key point of a consumer device: it has to have a low enough price that lots of people will buy it...
Well companies like Dell, HP, and IBM, Oracle (in the past with sparc) do make a lot from hardware support, and will continue to do so, I can definitely see it being a viable business model.
 
Well companies like Dell, HP, and IBM, Oracle (in the past with sparc) do make a lot from hardware support, and will continue to do so, I can definitely see it being a viable business model.
Not enough to sustain the company (or that part of the company). Look at IBM for example: they have only kept the mainframe and the server hardware, the rest have been sold. And it's not because of hardware support that they keep the mainframe and server hardware business, it is because of the software (and to a certain degree the software consulting services) that they sell on that hardware.
 
Not enough to sustain the company (or that part of the company). Look at IBM for example: they have only kept the mainframe and the server hardware, the rest have been sold. And it's not because of hardware support that they keep the mainframe and server hardware business, it is because of the software (and to a certain degree the software consulting services) that they sell on that hardware.
IBM and Sun (well Oracle now) have always been a little more complicated in that sense. Speaking of IBM I think they now own Red Hat. Dell and HP are probably a better example (although HP does do software too, I don't know about Dell), but there are definitely companies that make money from hardware repairs and support.
 
Look at IBM for example: they have only kept the mainframe and the server hardware, ...
Mainframe and Unix servers are indeed a portion of IBM's business, but a small one. Admittedly quite profitable, but not big.

IBM is a lot of things, but it is not primarily a systems company. It is much more a solutions company. Customers go to IBM saying "make my IT work". Parts of it, or all of it. At times, that used to include payroll and human resources processing (that part of IBM was eventually sold, I think it forms the core of Fidelity's human resources business), printing (I think that is now Ricoh), phone systems (IBM bought Rolm for that reason), programming, systems operation, and today cloud-based services. Spanning from desktop machines (including Thinkpads and support, today part of Lenovo) through data centers to supercomputers. Even today, there are large companies and government agencies where all of the data centers (including real estate for the data center and all staffing) are done by IBM. Some of those companies agencies have a CIO, but they are the only IT employee of the company: all they do is negotiate the contract with IBM. For the parts that IBM has sold off (like laptops), they have trusted supplier relationships. For example, there are companies where IBM provides all IT services, and IBM buys MacBooks from Apple and distributes them to the employees it's supposed to serve.

It's hard to understand that business when one comes from the hobbyist mindset of "I'll buy one piece of computing hardware at a time, and then do all the rest myself". In particular, IBM's hardware doesn't have to be the best or the cheapest one. Mainframes are heinously expensive, and are not all that fast for their prices. PowerPC servers were good and reliable, but they were always a bit pricey, and not that fast compared to the competition. But that doesn't matter: A big bank, insurance company, manufacturing company or government agency doesn't buy Whetstones or SPECmarks or Linpack. They buy solving their IT problems.

IBM is not primarily a supplier of hardware (like servers or printers). It's not primarily a consulting company that tells customers how to do IT. It's not primarily a software company that sells OSes, databases, middleware, application software, and zillions of other software products. Supporting the systems and software that it sells is a part of its business, but not the core. From an employee headcount viewpoint, the biggest part of IBM is actually business services, where IBM acts to some extent as a temp agency for IT staff - but renting out sys admins is not what makes IBM unique. None of those pieces are IBM; the totality is.

Or perhaps it used to be; I have not worked for IBM for ~4 years now, and I can see from a distance that it is shrinking further.
 
I'd really like to see some investment in RISC-V, perhaps a kind of 'Linux' approach to the hardware world where device manufacturers all produce open source specifications which make money by competing and providing support services.
Trust me, there are heavy investments in RISC-V going on right now, they just are not put so much into the broad day light.

Communist China is investing heavily in that technology to cut off technological ties to the west. This is the reason why nowadays the fastest RISC-V CPUs with the most cores originate China.

There are also working ports of Android using that ISA. So I do expect the Chinese to push that chip quite much.
 
Communist China is investing heavily in that technology to cut off technological ties to the west. This is the reason why nowadays the fastest RISC-V CPUs with the most cores originate China.
The same is true with ARM. Matter-of-fact, there is a big scandal brewing right now about China having seized the ARM IP. The question is: Will they be able to turn those efforts into a functioning chip/CPU/computer/IT industry?
 
How come nobody mentions Raspberry Pi or PineBook in this thread?

Also, the Chinese efforts are called LoongSon. Linux maintainers did complain that their ISA is a blind copy of MIPS code. Phoronix also noted that benchmarking available LoongSon processors showed poor performance...

Some of that kerfuffle can be thought of as a result of Apple, Intel, AMD, NVidia, and other big players hogging the TSMC lines for smaller and smaller fab process sizes, so the older and bigger sizes are 'lefotvers' that the Chinese maker pounced on to make LoongSon.
 
How come nobody mentions Raspberry Pi or PineBook in this thread?
Don't know about PineBook.

The Raspberry Pi is not particular "open". The core IP in the CPU is stock ARM, manufactured by Broadcom. Not open at all. The SoC is only partially documented, which is one of the causes why support for operating systems other than the Pi foundation's own Debian version (used to be called Raspbian) is often behind or incomplete.

What I hear about the Chinese MIPS clones is that they are technologically about a decade behind the west.
 
Don't know about PineBook.

The Raspberry Pi is not particular "open". The core IP in the CPU is stock ARM, manufactured by Broadcom. Not open at all. The SoC is only partially documented, which is one of the causes why support for operating systems other than the Pi foundation's own Debian version (used to be called Raspbian) is often behind or incomplete.

What I hear about the Chinese MIPS clones is that they are technologically about a decade behind the west.
Dunno how eternal_noob would react to that...
 
I stopped worrying about things you can't control decades ago and i live a peaceful life now.
No more thoughts about CPU backdors, NSA wanting to bribe Torvalds to plant a backdoor in Linux, Black Hat hackers, etc.

If a CPU is either closed or open is far, far away. In the end it's all the same. All is evil.

I just don't care anymore.
 
I stopped worrying about things you can't control decades ago and i live a peaceful life now.
No more thoughts about CPU backdors, NSA wanting to bribe Torvalds to plant a backdoor in Linux, Black Hat hackers, etc.

If a CPU is either closed or open is far, far away. In the end it's all the same. All is evil.

I just don't care anymore.
For me I am actually less interested in the privacy or security concerns (although I think they have a point, if only in theory) I just like things to be open source because it's an interest of mine.
 
The same is true with ARM. Matter-of-fact, there is a big scandal brewing right now about China having seized the ARM IP. The question is: Will they be able to turn those efforts into a functioning chip/CPU/computer/IT industry?
Nope, it's not true for ARM. ARM's IP belongs to a Japanese company called Softbank, while RISC-V IP belongs on the ISA's to nobody. So I somewhat doubt it for ARM.

Regarding your second question: maybe you haven't been following the news lately, but Communist China is pretty far advanced technlogically. For example they were the first to land a probe and communicate with a probe on the dark side of the moon.

Right now they are building their own space station, because nobody wanted them to contribute to ISS. They got the engineers and the drive to achieve something, and if they want to achieve something, they usually do this very quickly.
 
For me, the important part of open/closed is not even so much privacy/security, but more along the lines of technical capacity and your own shot-calling. For example, if the boot sector is properly documented, it could be much easier to swap Android from your smart toaster, install NetBSD, and then tell it to open your garage door Alexa-style :p
 
Nope, it's not true for ARM. ARM's IP belongs to a Japanese company called Softbank, while RISC-V IP belongs on the ISA's to nobody. So I somewhat doubt it for ARM.

It's more complicated. ARM's IP is owned by a British company called ... ARM (and I forget whether it's ARM Ltd. or ARM Inc. or exactly what the legal title is). That company in turn is owned by SoftBank, which is a Japanese venture capital company, investor, conglomerate, holding company, hedge fund (all of those descriptions fit more or less).

The problem here is that ARM had a Chinese subsidiary, which had a complete copy of the IP. And that subsidiary has recently announced that they are unilaterally "seceding" or disconnecting from their British/Japanese parent. And obviously taking the ARM IP with them. Read the various press reports about it: the net result right now seems to be that ARM China is controlled by Beijing, and has forked the IP.

Regarding the second question: There is a huge gap between high-profile prestige government projects (such as particle accelerator or space exploration), and mass-produced and economically viable technology. One of these days, over a beer, I shall tell the stories when I helped work on BEPC and BES, pretty accurately 30 years ago. The technology gap between the particle physics research lab in Beijing (where we were operating a western-class particular accelerator and detector) and the suburbs (where nearly all transportation was via bicycle, heating was coal, and bigger companies had their own coal-fired power plants to make electricity) was just astonishing.
 
Speaking of the expense of designing and manufacturing processors, and the specialist knowledge involved. This is precisely why some companies have a model of selling licenses for designs and leaving the manufacturing to others (e.g ARM). The design itself can be fully open source and saves a lot of work.
 
Back
Top