KaiOS Technologies and Mozilla partner to enhance Gecko engine for KaiOS
kaiostech.comYoutuber 'TechAltar' made a video[1] on March 11 giving his perspective on this.
While I wouldn't normally link to a YouTube video in this context, KaiOS Technologies specifically reached out to TechAltar to provide him with details so he could cover the story.
It's likely that one of the reasons KaiOS decided to reach out was because of his July 2018 video[2] covering KaiOS, which has over a million views.
[1] https://www.youtube.com/watch?v=_UPk3mpcDP4 [2] https://www.youtube.com/watch?v=OA_g2bQgOXY
Linked it yesterday (2020-03-11) with very little interest
His videos actually got me into operating systems, from his covering of more obscure alternatives to his videos on previous attempts that failed.
For those who don't know, KaiOS is a closed-source proprietary fork of Mozilla's abandoned FirefoxOS for mobile devices. It's become extremely popular for low-end feature phones in the developing world.
Not sure about closed source part: https://github.com/kaiostech
Unless I’m missing something, none of KaiOS’s UI is there. I see Gecko, the kernel, some hardware enablement bits, and sample apps.
it's partially open source, so you don't get the stuff you mentioned.
How can it be closed source? What licence was FirefoxOS originally released as? Or you mean that they added new bits to it and kept them closed source? In that case, what's their rationale?
I think FirefoxOS was under the MPL: https://en.wikipedia.org/wiki/Mozilla_Public_License
> As such, it allows re-licensing. MPL software can thus be converted into a copyleft license such as the GPL, or to a proprietary license (example:KaiOS).
Even most recent version of KaiOS still runs on ancient Linux kernel 3.10.49 (2014) with Firefox 48.0a (2016). Not only lacking in features, but in security updates too. With updated Gecko engine and proper PWA support KaiOS might have a chance to compete with other mobile operating systems. And if somehow this new engine will be released as open source, other projects, like postmarketOS, might be able to use it too.
That explains why I was unable to find any secure way to store keys or sensitive data in their API[0] docs.
Which means user authentication apps will not come any time soon, like Authy (Authenticators), Bitwarden (Password Managers) or BankID (Swedish) (Identity Verifiers).
Yeah, until they can get their act together on the security side this has no chance to become as widely used like androidOS/iOS. For some places this is less important than others. But as countries seem to be looking into digital-only currency this will become more and more of a problem.
I could understand version 45 (somewhat) as it was ESR. FF48 came out around the same time that Mozilla dropped support. My assumption is that KaiOS hasn't really poured many resources into that part of the OS. An update to the kernel and moving to Firefox 68 ESR seems like an amazing upgrade for users.
I can't imagine Mozilla putting in the resources without also open-sourcing all the things.
Please, KaiOS -- make a high end phone and hit the market that Mozilla never did with FFOS -- wealthy developers who want a platform completely unburdened with apple or google.
I'm not sure what's actually interesting about this partnership, it seems obvious, but one thing I think is a huge problem is that they're mentioning Gecko so much -- they should be trying to move to Servo as fast as possible.
There is already a project called servonk[0] that shows this might be very possible.
Why does it have to be a web-browser-based platform?
I've bought a ZTE Open[0] back when FFOS was made since i am in general a fan of Mozilla (despite their missteps now and then) since before Firefox was a thing. The phone was a disaster - applications were very slow, everything was sluggish and in some cases i lost calls because the UI had frozen due to swapping or whatever it was doing.
Now, you might say that it was a low end device, but here is the thing. Years before that i had a Nokia 6600 [1] which has literally less than 10% of the resources ZTE Open has, yet it was able to run multiple applications without a breeze (it was the first time i ran an IRC client on a phone) and even had several 3D games (which, imagine that, used software rendering despite the phone's limited CPU power).
(and of course there were PalmOS devices that were running on even weaker hardware yet they provided UIs so responsive that put even the fastest Android to shame -- but i have very limited exposure to those to judge properly)
Nokia 6600 puts in perspective how awful Android is nowadays, let alone FFOS that couldn't even manage to remain literally usable with more than ten times the resources.
Web-based seems to make a lot of sense. Something like 90% of all apps in both the Apple and Google store are already web apps. WASM makes this idea even more appealing.
webOS in 2009 isn't even a comparison for a few important reasons.
* They were primarily RAM restricted at only 256mb.
* The then new v8 engine was about 4x slower than the current iteration (not to mention using much more RAM). Webkit is also much, much faster today
* They used QtWebkit -- a slow and grossly outdated version of webkit that offered bad performance even back then.
* There team seemed to have zero experience in actually using and optimizing a Linux OS.
A single-core 1GHz A5 chip was a crazy idea. The ZTE Open C came out 6 months later *at the same price8 and had a better screen (480x800 instead of 320x480), twice the RAM (512mb instead of 256mb), and probably close to 4x the CPU power (2x A7@1.2GHz instead of 1x A5@1GHz -- for reference, AMD runs an A5 as their security coprocessor on their x86 chips).
The ZTE Open was simply a mistake. It would have been a mistake no matter the OS or native vs web apps.
ZTE Open ran Firefox OS, not webOS. Though my point was that a much older an MUCH weaker ARM-based phone that didn't try to pretend the web is a platform but instead used native code managed to both outperform and be a much more usable device.
I think it should be a web-browser based platform for a few reasons:
- ease of development
- leveraging gains in the platform development
At this point, the web is already the biggest app store the world has ever seen. FFOS succeeded in one of the most difficult areas of building a new platform -- getting companies to write progressive web apps that ran on the phone. I even spent time writing little utilities, a tinder clone, and an instagram clone for the platform.
There are at least 3 big companies with huge dedicated teams to making that platform better and faster. Slowness can definitely be an issue but that's a lot more solvable than a dearth of applications or a lack of open interest in improving the platform. Optimization is a long-tail problem, and it's clear by android's dominance and the progressive slowing of apple software every year that you just have to be good enough, but hardware goes a long way to help.
Even more than this the biggest problem was the market FFOS was aimed for didn't give FFOS time to go down the optimization path. You can mask inefficient software with beefy hardware, and I actually had two firefox phones that worked wonderfully -- the Flame[0] and the Fx0[1]. Those two phones only scratched the surface of how powerful the hardware could be, and I didn't have any of the problems you mention. I still have 2 (!) FX0s in my closet somewhere.
[0]: https://wiki.mozilla.org/Firefox_OS/Flame
[1]: https://developer.mozilla.org/en-US/docs/Archive/B2G_OS/Phon...
I don't know about you, but i always see my ZTE Open as hardware that was horribly abused. I do not see it as weak or anything, the hardware was more than capable since - as i mentioned - much weaker hardware could do much more. It is the software that was garbage.
And my point is that the ZTE open is nothing compared to the hardware modern (even mid-range) phones are shipping. All software is shit (to varying degrees) and other platforms mask their inefficiency with beefy hardware, and continuous fucking updates that almost don't run on older phones.
If an analysis of FFOS being bad depends on (in my opinion) running good-enough software on terribly slow hardware, the wrong question is being answered.
My entire point is that the software isn't good enough, is that the software is garbage for the task of being used for anything else than browsing the web. The hardware isn't terribly slow, it is WAY beyond adequate for providing smartphone features, as proven by an actual smartphone that existed a decade before with hardware than 10% of the power that ZTE Open has yet provided a much better and snappier user experience.
The question on my original message was rhetorical - of course it doesn't have to be a platform built around a web browser. That is a massive waste of resources that could be used for a ton of other things.
Servo is essentially Mozilla's Rust browser test ground, and my understanding is that it is not meant to stand on its own as a browser engine.
I saw it more that servo is the codebase they want firefox to be, but changing out the engine in a humongous codebase like FF is extremely difficult so they're taking it a step at a time. They merge bits and pieces in from time to time[0] and I expect that to continue until firefox is servo in as many places as possible.
Looks like servo might only be a breeding ground for ideas after all.
[0]: https://en.wikipedia.org/wiki/Servo_(software)#Relationship_...
> who want a platform completely unburdened with apple or google
I don't think KaiOS will be. It seems that Google is one of the main investors in KaiOS Technologies[0][1].
[0]: https://www.kaiostech.com/google-leads-seriesa-investment-ro... [1]: https://www.crunchbase.com/organization/kai-technologies#sec...
Last time I heard about servo I read that it was a proof of concept and that the good ideas would be ported to Gecko. I could be wrong. I'm still quite confused about what servo is and will be.
Companies can't survive on high-end products only, unless there is a monopoly or external cash. (I do want new platform of course, like everyone else)
I'm not sure this is true, especially in the era of free-flowing VC cash and low interest loans.
Maybe KaiOS/Mozilla's (now defunct) mobile department in particular can't, but companies can definitely survive on high end products only.
Is there more to this story than a corporate press release? or should we wait until something substantive appears?
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...
It seems that if you want your app to be available on the KaiOS app store, your app must have ads, and must use KaiOS' platform for those ads:
"Currently, we are prioritizing apps for QA that use KaiAds for monetization. Visit the KaiAds website to learn more.
After you have integrated KaiAds, go through this checklist."
They can't really sell the OS, so that makes sense for everyone as devs can offer the app for free too.
I wonder if there's a way to pay to opt out of their ads though...