Fuchsia is now half dead. After cutting back its staff, Google has also cut back its size to microfuchsia. It will probly end up fully dead, or just merged into Android code. What a shame, But it’s typical of Google, who promotes new projects to keep its name in the media, only to later kill them. They once had a “project vault secure OS on an SD card” - but it was never heard about again.
1 Like
It’s probably better this way.
My intuition tells me that,on Fuchsia phones,there would have been even less options left to get rid of the Google spyware crap,compared to Android.
I don’t like Android,but due to a lack of alternatives I have to use it,and you can still make it somewhat usable and private if you have USB Debugging/ADB and know the right adb shell commands.
I doubt that would have still worked with Fuchsia.
3 Likes
Why? the OS kernel has little to do with what is build ontop.
Windows NT as such is also not a terrible design, it’s what you do with your tools.
In any case; Fuchsia has already indirectly contributed many things to other projects. Including a bit to Haiku. : )
https://cgit.haiku-os.org/haiku/commit/?id=485b5cf8bc00ab5aaee5db81a91d135dc61d9cc3
2 Likes
Speaking of kernels,well,Android using the Linux kernel with its GPL license that requires hardware vendors to publish their driver source is what makes Custom ROMs possible.
With a kernel under a more permissive license,they will keep their driver sources closed,now forget about any Custom ROMs.
Today there are already lots of manufacturers who just don’t care and keep the drivers proprietary,I own such devices and it sucks.
But now that’s illegal and there are other manufacturers who do better.
With the Fuchsia kernel that would be absolutely legal.
2 Likes
Not really, no.
Custm ROMs often keep the kernel as built by the manufacturer and run a new userspace on top of it. The monolithic Linux kernel prevents to do things in any other way, whereas with a microkernel, in theory, you could upgrade the microkernel while keeping the (binary) drivers provided by the manufacturer, and possibly you could easily recompiler the drivers (if you have the source) for a newer version if needed.
With Linux, manufacturers just take a version of the kernel, make a lot of dirty hacks on it, and give a source that no one else can use, except maybe as a reference to write clean drivers that can be upstreamed.
Unless you are talking about something like Replicant, which is what you get when you actually build everything from source. The results are not great: only very few phones are supported, often without 3D acceleration, sometimes without modem support either. The Replicant devs do a great work with what they have, but it is not easy at all. And the help they get from Linux GPL license is counterbalanced by the monolithic approach which prevents extracting just one driver from the manufacturer sourcecode.
Anyway, as far as I know, there weren’t really any plans for Fuchsia on Android phones, only some rumors? Like with many things at Google, it started as a solution looking for a problem. And maybe it didn’t find enough problems to solve?
4 Likes
There is an old discussion about embedded mobile devices, Google wants Fuchsia to be new iOS:
2 Likes
It won’t happen unless Google phones become as popular as iPhones which I doubt. People are more concerned now by privacy and will stay with Android phones. Those who are less are already happy apple clients.
They probably knew it was a wishful thinking. Why would you name a project after a flower if you don’t want it to fade?
Android has a niche market in the USA, worldwide adoption is greater than iOS:
Google is a corporation, it is interested in profit, it is not possible to take Google’s word for it, in one country they say one thing in another country the opposite, corporations are on the stock exchange, they do not have stable capital in other words (stock assets) , only speculative “money”, which does not exist, corporations know this, hence this wave of mass layoffs in several corporations.
Ordinary people don’t understand privacy, they think they do, they use bad privacy software, which is advertised as safe, people want convenience, using their favorite programs on their mobile phone, Fuchsia can run Android apps, Harmony OS too, that’s enough for common user.
The issue is that Google fired people from the Fuchsia team, already trained, training a new team takes time, Fuchsia has not yet matured to use or replace Android, Google does not have the hands and skills for this, like several projects it abandoned.
Fuchsia had a lot of hype as the “next big thing” from Google, but I reckon most of its technical qualities (micro kernel, “capability based” safety and other embodiments of OS theory of the 2010s) are to be found in Genode. This, and particularly it’s Sculpt desktop system, are making nice progress and updated twice a year.
Okay so you won’t get it in your local phone shop anytime soon, but Genode follows the same paradigms and is available to play with now.
It’s not named after a flower, it’s named after a color.
A bit of backstory on this name: it dates back to the 1990s, when Apple was struggling with their aging OS and trying to make a new one. At some point they made a big TODO list, and used blue post-it notes for features that should be implemented in the current OS, and pink post-its for things that should be in the big next major rewrite.
So the two versions of the system became known as “blue” and “pink”. Eventually, no one wanted to be working on a “blue” project, and “pink” became an over-featured monstrosity that never shipped. Some people were frustrated with that lack of progress, and left to create a new company, called Be. And implemented their own vision for “pink”.
They didn’t manage to sell it back to Apple, however, who instead eventually purchased NeXT.
There is a large overlap between the BeOS developers and the Fuchsia ones. And they tried to make another variation of pink, giving it a name that is very clear about it, if you know this backstory.
8 Likes
That’s good. I believe there’s some ambiguity about how the flower’s name should be pronounced - “fyu-sha” as English speakers often do, or “fu-ksia” as presumably intended when the genus was named after Leonhart Fuchs.
But since no botanists are involved in the name of the color, there isn’t anyone to stand up for Fuchs and the common pronunciation will provoke no controversy. Only a problem for poor spellers, since the pronunciation reflects an “sch” interpretation vs. actual spelling “chs”.
As a microkernel on top of Linux? I can imagine why the original team had to be removed.
It’s a version of Fuchsia with basic functions running on Android, reminiscent of L4 kernels:
The layoffs are to cut expenses and leave shareholders with a profit, after all, corporations do not generate wealth without real assets.
So the two versions of the system became known as “blue” and “pink”. Eventually, no one wanted to be working on a “blue” project, and “pink” became an over-featured monstrosity that never shipped. Some people were frustrated with that lack of progress, and left to create a new company, called Be. And implemented their own vision for “pink”.
Curiously, BeOS colors are red and blue…
1 Like
The fate of Fuchsia is, in my opinion, symptomatic of a declining civilisation that has switched to rent extraction rather than innovation. Take for example the Boeing 737, an airliner that first flew in 1967. This should have been replaced at least a decade ago with a new airframe, but they keep it in production long past its use by date with kludges as required to keep it (sort of) competitive. Most of these big name silicon valley outfits are just like boeing, the consumer products atrophy and become increasingly shoddy but that is okay because they are too big to fail since their lucrative government contracts mean that they will be bailed out.
The EU benefit from the newer A320 series that competes with the 737. China and (even!) Russia have even more modern Comac and Irkut equivalents
Unix dates from around the same era as the 737, but forms the basis of almost all modern computing. It is a real pity that google faiiled once again to replace it, but here we are.
3 Likes
A lot of the stagnation in computing is inertia that’s inherent in complex systems created by haphazard aggregation. Big changes have to really deliver something, to compensate for the effort and the damage. Maybe we’ll be able to climb out given the combination of 1) some kind of digital analysis of software gives us the tools, and 2) some new problem area where existing platforms don’t cut it. Obviously Google thinks there’s some need, in that their micro Fuchsia plan looks kind of desperate - an OS that doesn’t have anything ported to it, running on top of Linux? But with all their resources, they can’t make a clean break.
1 Like
I think that says it’s more about one specific country, rather than a whole civilization, at least for now. Let’s hope that this one country does not bring everything else down with it.
3 Likes
Wow even as an old BeOS developer and avid nerd history follower I never heard of this back story, thanks for the insights!
Some sources so that I don’t make things up:
https://asleson.org/public/mirrors/www-classic.be.com/aboutbe/benewsletter/Issue14.html
Google’s Fuchsia OS was one of the hardest hit by last week’s layoffs | Ars Technica (which also mentions that iOS was initially called “purple”, probably from the same lineage of naming projects after colors)
2 Likes
I don’t know what the benefits of microfuschsia are exactly, but I imagine there could be a lot of benefits for taking this approach with Haiku. Unlike Fuchsia, Haiku already has a reasonable amount of native software, and the VM sandbox can make up for the fact that Haiku does not have multi-user or much security.
I think we naturally assume Haiku has to be on bare metal to be a “serious” OS but IIRC read somewhere that because of so-called cloud, most linux installations are in VM rather than bare metal these days.
The Register’s Liam Proven talks about using Plan9 to run microVM for appropriately stripped down linux, perhaps Haiku might do the same? And maybe Google were paying attention at that talk?