Huawei ban and alternative OS for mobile

[quadrant]

The recent news that Huawei can no longer use Android has got me thinking about diversity. A lot of the operating systems I have seen on this site target personal computers. Are there any hobby operating systems targeting mobile devices (smart phones), and of them are there any that are at a usable state?

What are the main roadblocks to developing operating systems geared at smart phones? I'm under the impression that most phones use the ARM architecture, and that the ARM architecture is not that much different than the x86 with regards to writing an operating system for it...

I previously thought Android was the equivalent of Linux in mobile (in the sense that everything you need is in a public repository). But it seems that is not the case.

[zaval]

oh, I don't even know if this is a really bad news for Huawei, that they can no longer use andorid. I hate it so much. So many frustration and dissatisfaction as a user, the dumbest OS ever existed. google.

[nullplan]

The main road blocks are going to be the exotic hardware. Most phones come with a proprietary GSM chipset, and a SIM card interface. So now you have to get these two to talk to each other to facilitate mobile communication, and there's hardly any documentation out there.

Next roadblock is that there is far more diversity and far less standardization between phones. There is no ACPI that tells you how many CPUs there are and how to reach them. No MP specification to tell you how to start up the other CPUs. Hell, most of the time you don't even know what memory is available. And where to find the hardware. ATM this is all solved with magic information in the kernel (or with a device tree).

But there is some movement. Osmocom (https://osmocom.org) is trying to make Open Source GSM a reality, both in the phone and in the backend.

If Huawei had any sense they'd do their best to hurt the ecosystem while they still can. Open up the boot loaders on the remaining phones, etc.

[Schol-R-LEA]

oh, I don't even know if this is a really bad news for Huawei, that they can no longer use andorid. I hate it so much. So many frustration and dissatisfaction as a user, the dumbest OS ever existed. google.

I'm guessing you'd never used a system along the lines of Pr1mOS or VMS, then. :-p

(Actually, VMS had some really nice features for an OS of its day, such as automatic file versioning. But man, was that CLI painful for many purposes.)

More seriously, given your comments not just about Android (which I have no intention of defending, TBH, but I am guessing my beefs with it are not the same as yours), but about a number of other systems as well, I am curious as to what you consider a good operating system, or perhaps, what you would want an OS to be like if no current ones are acceptable to you.

I do see the reference to an NT-like OS in your .sig, but I'm not sure if that is really indicative of your real goals.

This probably deserves a separate thread, where we can all try to explain what we want most in an OS; in fact, I'll make one right now, and invite you and everyone else to it.

[quadrant]

Next roadblock is that there is far more diversity and far less standardization between phones. There is no ACPI that tells you how many CPUs there are and how to reach them. No MP specification to tell you how to start up the other CPUs. Hell, most of the time you don't even know what memory is available. And where to find the hardware. ATM this is all solved with magic information in the kernel (or with a device tree).

For PCs, was this standardization driven solely by the desire to be IBM-compatible? Would it be in the interest of the manufacturers of smart phones and their components to form a working group to establish some standard interfaces? Or would this stunt growth/innovation?

For example, back in the day each phone had its own unique charger. At some point, most switched to the "standard" USB charger. This seems a win for both the manufacturer (easy to procure from a 3rd party) and customer (easy to get a replacement, borrow a charger).

Another example, consider graphics cards in PCs. While a company can really customize a graphics card to work "amazingly well" with a specific processor, it is in their interest to make it work "modestly well" with a wider variety of processors so that they can sell more units. Does this dynamic not exist in the smart phone market?

As an aside, is there a Novena equivalent for phones? Where all the hardware (and related documentation) is as open as possible? Because if so, a "hobby" effort to make an OS for such a phone would be more worthwhile (focus is on the OS development rather than trying to reverse engineer undocumented protocols).

[Schol-R-LEA]

The thing is, the problem Huawei faces isn't about the OS itself; they could transition to a different one, either one they developed themselves, one which they license, or even one of the handful of existing open-source ones. While they OS performance, stability, and UX would have to be competitive, that doesn't actually represent a major hurdle for a company their size.

For that matter, as the article mentioned in Quadrant's opening post points out, they are still free to use the FOSS parts of Android itself - it's the system-level integration which Google provides for the major apps which they can't take advantage of any more. And that's a much bigger issue when it comes to UX.

The real problem they face if they move to a different OS is the apps ecosystem, mostly regarding getting enough developers on board with the platform, but also in having to curate it to avoid the sort of problems Google Play Store is infamous for. Regardless of what you think of either the Play Store or the IOS App Store, they both have a well-stocked selection of browsers, e-stores, utilities, applications, and games - and as strange as it might be to say, the games are the most important part of a mobile app suite, followed by the web browser, the e-store apps, and the social media apps. Huawei would have to not only get a huge pile of games and messaging apps ported to their new platform (I expect TenCent would be fast to support it, others not so much maybe - especially in light of Chinese import restrictions and censorship regs), they would also need to convince most of the major retailers in each of the different national and regional markets they would be selling the devices in to port their e-stores to it.

This would be no mean feat, especially for units sold outside of PRC itself - while I can see them easily getting all of the various Alibaba apps ported (which also touches on games and social media, as AG is a huge conglomerate), I doubt that US firms such as Amazon would be willing to, especially since the ban would mean that those devices won't be used in their home markets anyway.

The OS itself is the smallest part of that, really.

[quadrant]

The real problem they face if they move to a different OS is the apps ecosystem, mostly regarding getting enough developers on board with the platform, but also in having to curate it to avoid the sort of problems Google Play Store is infamous for.

Ah I see. If the apps are written for Android, then it seems technically there is nothing stopping developers from publishing the same apps they publish in the Google Play store to alternate marketplaces (for OSs based on Android). All the alternative marketplace would need to sort out is how to pay the developers and (as you point out) moderate the content... If it mirrors the approach taken by Linux package managers (ex. apt-get), then the pay part could be ignored, and the focus be on community moderation of content...
Google's apps are awesome, but it isn't unimaginable for viable alternatives for each of them to emerge... Perharps an alternative OS could make more of these "system-level integration" features accessible to developers...

they would also need to convince most of the major retailers in each of the different national and regional markets they would be selling the devices in to port their e-stores to it.

Not sure I understand what you mean here by e-store... Do you mean websites like walmart.com?

[Schol-R-LEA]

The real problem they face if they move to a different OS is the apps ecosystem, mostly regarding getting enough developers on board with the platform, but also in having to curate it to avoid the sort of problems Google Play Store is infamous for.

Ah I see. If the apps are written for Android, then it seems technically there is nothing stopping developers from publishing the same apps they publish in the Google Play store to alternate marketplaces (for OSs based on Android). All the alternative marketplace would need to sort out is how to pay the developers and (as you point out) moderate the content... If it mirrors the approach taken by Linux package managers (ex. apt-get), then the pay part could be ignored, and the focus be on community moderation of content...
Google's apps are awesome, but it isn't unimaginable for viable alternatives for each of them to emerge... Perharps an alternative OS could make more of these "system-level integration" features accessible to developers...

I think you're underestimating just how much a successful consumer mobile platform - indeed, any successful consumer (as opposed to hobbyist or server) platform - depends on commercial developers. By a lot.

they would also need to convince most of the major retailers in each of the different national and regional markets they would be selling the devices in to port their e-stores to it.

Not sure I understand what you mean here by e-store... Do you mean websites like walmart.com?

I was thinking of things such as the Amazon or Walmart apps for Android and iOS, but more or less similar, yes. The apps generally have features the websites don't, and are (most significantly) integrated into the platform's messaging and announcements systems, as well as being more tuned for the host than the websites are (in theory, anyway... some of them aren't very tuned at all, as it happens).

[iansjack]

Slightly off-topic, but I think that this is going to harm US businesses, and consumers, more than China. A poor businessman, who has bankrupted huge corporations, is in the process of doing the same to a nation.

The US is fast becoming the parriah of the world.

[Schol-R-LEA]

I missed this earlier this past week, but ARM is also forbidding sales of licensed ARM-based CPUs to Huawei, which is a much bigger matter for that company. First off, they are a UK firm owned by Japanese investors, not a US one, and AFAIK neither the UK nor Japan have officially joined in the US action against Huawei (though I may be mistaken); in other words, it could be seen as a reflection of how fabricators and IP licensors worldwide see the situation regarding the industrial espionage allegations (which honestly is the bigger issue for most corporations, as I doubt that either the illicit data mining or the allegations of violating trade sanctions against Iran and North Korea are much of concern to non-US companies). It also shows that, declining power or not, the US has a lot of influence elsewhere (perhaps even undue influence), if only because so many companies license American IP covered by the ban in their products (which ARM Holdings claims to be the main impetus in this move).

IIUC, pretty much every computing device Huawei makes is ARM-based, and without ARM processors, they are is serious trouble.

They were also forced out of JEDEC (technically they have 'voluntarily withdrawn' but it was under pressure from other members), the SD Association, and the Wi-Fi Alliance, each of which is a major blow.

While it is possible that they could shift to, say, MIPS (probably Loongson, the current versions of which are indeed licensed from MIPS Tech), or RISC-V, it seems rather unlikely to succeed. It is almost certain that MIPS Technologies (now owned by Tallwood Ventures, meaning they are entirely under US management) will follow suit, and none of the RISC-V processors are quite at the level of development to match the performance of the major ARM implementations (SiFive's FU540 is almost there, but not quite, and no full-featured SoC version is ready at the moment, while the RISC-V SoCs that do exist are decidedly underpowered) even if they could find a fabber who would sell to them. Huawei would have to pour a lot of money not only into changing to a new architecture and new CPUs, but also into either licensing and/or purchasing a whole new range of CPUs. They may end up needing to invest in bespoke production of the CPUs and SoCs themselves, or even creating their own fabs, which would be a huge expense even for a company their size (hell, even their government backing might not be sufficient to cover something that big).

This isn't even addressing what the inability to use SD cards and develop Wi-Fi enabled hardware would mean to them.

[alexfru]

While it is possible that they could shift to, say, MIPS (probably Loongson, the current versions of which are indeed licensed from MIPS
Tech), or RISC-V, it seems rather unlikely to succeed. It is almost certain that MIPS Technologies

Wave Computing, Inc.

(now owned by Tallwood Ventures, meaning they are entirely under US management) will follow suit

IANAL, but I don't know to what extent that is possible since MIPS Open happened.

[Schol-R-LEA]

OK, I missed that entirely. Thank you for pointing that out, I don't know how I didn't hear about this until now.

(It looks as if someone needs to update the Wikipedia entries for MIPS Technologies and MIPS Architecture, and maybe add one for Wave Computing - I checked, there is no page on them at all right now - but I personally don't know enough to do it accurately.)

That still leaves Huawei with contracting with a silicon manufacturer with no strong connections to the US (such as T-Platforms or - more likely - ICT) to produce suitable implementations of the architecture, as well as changing their production systems to use them.

This ties into what iansjack said, potentially; with enough backing, a move by Huawei to use domestic or friendly-partner production would backfire for the US, as it would boost the development of those domestic sources. Potentially. Whether Huawei (and their backers in the PRC government) have enough clout to pull it off before the company implodes is an open question.

[quadrant]

It is sad that the decisions of those at the "top" do not reflect the wishes of the majority they have been elected to represent. Even sadder that they will not be the ones to bear the brunt of the consequences.

[nullplan]

It is sad that the decisions of those at the "top" do not reflect the wishes of the majority they have been elected to represent. Even sadder that they will not be the ones to bear the brunt of the consequences.

What consequences? For me (average consumer) nothing much changes. One less choice for phone manufacturer. For Huawei, they now have a lot of motivation to find an independent supply chain. It will be interesting to see the consequences of that. China is a big enough market on its own to keep Huawei alive, even if they are forced out of every other market.

For Huawei customers, Google has already stated that they'll continue to support already-sold phones.

For all other phone/electronics manufacturers, this should be a wake-up call: Apparently, a campaign from the Orange Man can torpedo your entire supply chain. Best find US-free alternatives. And apparently UK-free alternatives as well, but then I don't remember a time when the Brits weren't considered a US colony.

[lmemsm]

The real problem they face if they move to a different OS is the apps ecosystem, mostly regarding getting enough developers on board with the platform, but also in having to curate it to avoid the sort of problems Google Play Store is infamous for.

Ah I see. If the apps are written for Android, then it seems technically there is nothing stopping developers from publishing the same apps they publish in the Google Play store to alternate marketplaces (for OSs based on Android). All the alternative marketplace would need to sort out is how to pay the developers and (as you point out) moderate the content... If it mirrors the approach taken by Linux package managers (ex. apt-get), then the pay part could be ignored, and the focus be on community moderation of content...
Google's apps are awesome, but it isn't unimaginable for viable alternatives for each of them to emerge... Perharps an alternative OS could make more of these "system-level integration" features accessible to developers...

There are already alternatives to Google Play Store, at least for apps that work on Android devices. F-Droid is the best Open Source ( https://en.wikipedia.org/wiki/F-Droid ) example I've been able to find so far. I was very disappointed with the lack of Open Source development of apps written in anything other than Java for Android. I use C and SDL for app development. I've found very few Open Source projects encouraging sharing of C code for Android devices. I even wrote to the FSF about the topic of mobile development since they have some posts at their site regarding Free Software for mobile and how important it is, but they never even bothered to reply to my message. There are also several closed source alternatives to Google Play Store. The Kindle site seems the most visible to me. They even have a tool to check if your app works on various devices and once you run the check, they ask if you'd like to add your app to their store. I've seen lists of several other sites trying to offer their alternative to the Google Play Store too.

Microsoft tried to copy the popularity of Apple's and Google's app stores and I don't think they've managed to make a real dent in that market. Microsoft also already has several loyal developers that work with their compilers and other development tools. Looking at their results, creating a popular alternative to Android and IPhone with a successful app store isn't an easy task. One of the issues with app development is that each system is so different and requires development time, resources, etc., it doesn't pay the app developers to try to support the smaller markets. I personally think SDL is a nice solution for developing since it works with Android, IPhone and most desktops. It's difficult to work with C in Android though. Even though the kernel is C, it's like working with a second class citizen on an Android device. You can only really use C/C++ for command line applications (which require something like a terminal emulator to interact with) or for libraries called from Java. Both Apple and Google are pushing alternative language development (Swift or Kotlin) which makes portability even more of an issue.

There are some projects for developing independent operating systems for smart phones. Most of them are Linux based (like Android). Several are starting with Android or Linux source code and marketing their product based on the idea that their version is more secure than other corporate solutions out there. Librem-5 is one of the few that says it has solutions for supporting some of the proprietary hardware needed in order to make phone calls successfully or connect apps to phone features. Canonical attempted to cross into the phone market, but seems to have given up on the idea. There are Open Source projects like UBports that are trying to keep going with what Canonical started. There are also a couple of projects like Android x86 and Remix OS that are attempting to use Android source for operating system development. I believe the Remix OS project is no longer active. So it seems like there are a lot of inactive and/or failed attempts in the market (some by rather large corporations).

It would be nice to have some alternative solutions in this area. I've tried contacting a few of the Open Source mobile projects out there to see if they need developer help. While many advertise they want volunteers, they don't seem very interested when I contact them. If they do really want help, it seems to be in very specific areas. Was also looking into some of the projects working on Open Source digital assistants. Same results. On that note, if you're interested in having some portable C applications ported to your operating system, I'm still looking for a project that really wants volunteers.