[m463]

As a purism owner, and a long-time linux user... I think that is a fantasy statement.

Theoretically, there could be an open platform like the PC (with usb + pcie + ATX case/power supply + etc..) with open interfaces. However the reason this came into existence was by microsoft's design to commoditize the hardware to drive software sales of its (closed) operating system.

With cellphones all of those interfaces are being subsumed so the trend is one chip + a display + a battery. The chip is IP of many vendors.

Additionally, the linux distributions have not had the highest performance. For example, frequently there is poor or no graphics acceleration.

I think Linux will always be behind commercial/proprietary platforms. One could arguably say that the iphone is a multi-billion dollar platform, with more careful engineering, development and tuning than any other device on the planet.

That said - I do believe linux based phones are nearing that "good enough" stage where dedicated users can make it work for them and people may at least have a choice.

[zozbot234]

> I think Linux will always be behind commercial/proprietary platforms. One could arguably say that the iphone is a multi-billion dollar platform

One could have expected the same wrt. proprietary *NIX workstation and server hardware in the 1980s and 1990s, and where are those today? Linux is dominating that market. Embedded brings more trouble because the hardware, far from being a "multi billion dollar" endeavor, is all-too-often entirely undocumented and sloppily hacked together, where a barely workable state is considered "good enough" for shipping. But even there, Linux is easily gaining ground over proprietary OS's. The underlying dynamic is clear enough.

[zepto]

> One could have expected the same wrt. proprietary *NIX workstation and server hardware in the 1980s and 1990s, and where are those today?

Server hardware is of course commodity and dominated by Linux.

The high end workstation proprietary workstation market seems pretty alive and dominated but a closed Unix still.

> Embedded brings more trouble because the hardware, far from being a "multi billion dollar" endeavor, is all-too-often entirely undocumented and sloppily hacked together, where a barely workable state is considered "good enough" for shipping. But even there, Linux is easily gaining ground over proprietary OS's.

This is true of ‘embedded’ but phones are not embedded.

> The underlying dynamic is clear enough.

I think this is wishful thinking. Embedded and servers are quite different from phones and workstations.

[canucker2016]

> ...However the reason this came into existence was by microsoft's design to commoditize the hardware to drive software sales of its (closed) operating system.

Don't make the tail wag the dog.

IBM was in control of the hardware design and the rest of the ecosystem, heavily influenced by their previous work on the IBM System/23 DataMaster.

Microsoft just jumped on board for the expected lucrative ride, supplying a CP/M-like OS that they purchased from a local shop (SCP's QDOS).

IBM PC history:

https://arstechnica.com/gadgets/2017/06/ibm-pc-history-part-...

https://arstechnica.com/gadgets/2017/07/ibm-pc-history-part-...

Byte Magazine, Sept. 1990, "The Creation of the IBM PC"

https://archive.org/details/byte-magazine-1990-09/page/n451/...

[m463]

IBM designed the original ISA bus PC. But with microsoft's non-exclusive agreement it was able to sell MS-DOS, and it kickstarted early clones like compaq. IBM tried to close things down with the PS/2 OS/2 and microchannel, but they couldn't close the barn door.

here's another fun article: https://www.joelonsoftware.com/2002/06/12/strategy-letter-v/

[canucker2016]

Besides a license for MSDOS, the PC clone makers also needed a compatible BIOS. Compaq was the first out of the gate with a mostly-compatible IBM PC BIOS.

The IBM PC-compatible BIOS from Phoenix Technologies allowed more companies to jump in the PC clone wars.

https://www.allaboutcircuits.com/news/how-compaqs-clone-comp...

https://www.quora.com/How-is-the-BIOS-for-a-computer-motherb...

[znpy]

> I think Linux will always be behind commercial/proprietary platforms.

the dumb thing of this whole situation is that 50% of the "commercial/proprietary platforms" in this market are android phones, still using Linux as kernel.

[new_realist]

That's because the kernel is a commodity. Whether it's Linux or Mach, the user can't tell the difference. It's just plumbing.

[dTal]

It's not a "commodity", it's nearly 30 years of engineering that would take a multi-billion dollar investment to even come close to replicating. If it's so replaceable, why does everyone (except Apple) use Linux? When was the last time you bought an IOT device that ran, say, QNX?

All software is "just plumbing". The shiny-shiny on top is the mere tip of the iceberg and not what makes everything, you know, actually work.

[goodpoint]

This can be also phrased as: software freedom does not reach the end user.

That's why GPL and AGPL exist.

[noduerme]

it certainly touches the end user the moment they try to use the device for productivity and realize where the walls of the garden were established. But at that point maybe they're not a "user" anymore.

[goodpoint]

A that point they are a "used".

[lmm]

The whole phone will be a commodity soon if it's not already. When was the last time something felt like a significant innovation/differentiator in a phone?

(Personally I haven't been excited about a new phone feature since the S7's notification LED - and that apparently wasn't important enough to keep in newer versions).

[npteljes]

Cameras and the accompanying image processing software have been getting exceptionally good. Not that I'm personally excited about these, but I think it deserves recognition.

[lmm]

Meh. I took some photos with a 10-year-old digicam a couple of weeks ago and it was much the same "marginally nicer than my phone" experience that it was when phones started getting cameras. Thank goodness I could get away from that horrible fake bokeh that recent phones do. I will admit that HDR support can be pretty nice.

[npteljes]

Low light performance is also very nice in some phones. But what I'd like to emphasize is how automatic it all is. Surely someone with an entry level DSLR and a bit of Lightroom know-how can outperform an iPhone 11, but it's work to carry it around and apply the postprocessing afterwards, while on the phone it's just a few taps. So what I'd like to say is that it's not an achievement in imaging, but a big leap forward in automation.

[disqard]

This is an excellent observation.

[fitzie]

totally agree. while I'm optimistic that arm chips will be more open in the data center and embedded space, looks like mobile arm id becoming more proprietary. Intel could shake things up if it starts to work on riscv and brings some of its graphics and wireless tech over to it.

while it is nice to have a powerful phone, it really doesn't need to be any more powerful than what is in a tv set.

[zepto]

> it really doesn't need to be any more powerful than what is in a tv set.

I’d like to believe this, but if it were true, the Pinephone would be super responsive, and nobody would be complaining about iOS browser performance.

[cptskippy]

> However the reason this came into existence was by microsoft's design to commoditize the hardware to drive software sales of its (closed) operating system.

Couldn't get past your rewriting of history just to paint MS in bad light because you don't like them.

[ghaff]

What rewriting of history? The PC (and then volume server) era very much saw previous vertical silos tiled into a horizontal structure with x86 at the processor layer and Windows at the operating system layer. In the 1990s, a lot of computer companies were ready to concede the OS crown to Windows.

(To the degree I'd argue with the statement, it's that it was as much Intel's doing as Microsoft's; Andy Grove even dedicated part of one of his books to this.)

[cptskippy]

Microsoft didn't design or intend the commoditization of PC hardware. It fell into their lap by happenstance and they fought to retain it.

[m463]

I was not putting down microsoft, I think they did the world a favor by bringing us commodity hardware.

fun article:

https://www.joelonsoftware.com/2002/06/12/strategy-letter-v/

[cptskippy]

Microsoft didn't set out to commoditize computer hardware, IBM did. IBM literally wrote the manual for IBM AT (later ISA) and then shared it with the world with the intent of commoditizing add-in cards for their PC. IBM however thought that by controlling the BIOS, they would retain control of the PC. COMPAQ reversed engineered the BIOS a year later and essentially dropped control of the PC market into the lap of the owner of DOS, Microsoft.

I think Joel is taking some liberties to push his narrative. Either that or he's regurgitating some nonsense barfed in his ear during his time at Microsoft in the 90s about how it was their plan all along. Microsoft wasn't being savvy with the way it licensed DOS to IBM, they didn't seen the coming reverse engineering of the BIOS. Quite simply, they didn't license DOS exclusively to IBM because they couldn't.

After licensing BASIC, IBM later approached Microsoft to source an OS in part because they had a deadline and because their failure to license CP/M was by virtue of the fact that they were IBM. Gary Kildall heard "IBM" and basically told them to get fucked. Microsoft became their Trojan horse to prevent another Kildall situation.

There's a lot of different stories about the origins of QDOS from Microsoft paying SCP to reverse engineer CP/M to QDOS containing CP/M source code. Regardless, Microsoft didn't have an exclusive license to QDOS when it re-licensed it to IBM. It wasn't a tactical move, Microsoft didn't provide IBM with an exclusive license of QDOS because at the time it simply wasn't there's to give. They managed to acquire ownership of QDOS from SCP prior to the PC launch but again that wasn't with the intent to commoditize computer hardware.

Post PC launch and COMPAQ clone, Microsoft wasn't fully at the helm either. IBM saw the error of their ways and sought to rectify it with PS/2 and MCA. If Microsoft had assumed control then IBM probably wouldn't have teamed up with them to build OS/2. Microsoft did however understand what OS/2 meant for their own bottom line and sought to subvert IBM. I would argue that up until the mid to late 90s, Microsoft was largely protecting it's own interest from IBM who was trying to wrangle back control of the PC market.

Again, Microsoft didn't design this. It wasn't their intent from the onset, the market fell into their lap and they fought off everyone else's attempts to take it away.