Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

yeah the base open sources Android is barely usable. Even simple things like getting a local notification at the right time are not a given.

If they do the same with Fuschia, they'd better make it closed sources then.



Same with Darwin and MacOs.

Is there even a functioning darwin build for Big Sur??


Sorry, not familiar with Android core. For your example, how would it work without proprietary technologies? AFAIK, Android uses GCM for notifications. Adding integrations for it don't seem to make much sense in core open source Android as that will make you tied to Google's ecosystem.


Because I don't have Google's services on my phone, I have two always-visible notifications that never go away:

"Telegram" / "Push service: tap to learn more" (tapping opens https://github.com/Telegram-FOSS-Team/Telegram-FOSS/blob/mas...)

"Signal" / "Background connection enabled"

Both are serving the same purpose, for different apps. Even before we get to network connections and servers (the things that might make sense to be proprietary), without Google's ecosystem the local side has a bad UX and requires hacks.


That is just because they are not using something like https://bubu1.eu/openpush/ and instead implementing their own push services. This is not a problem with Android but with the apps.


I was speaking about _local_ notifications, not push: some clock / reminder apps never fire notifications or way too late. I think it's because the OS aggressively shutdown apps running in background.


IMAP Idle works fine for e-mail so long as Android doesn't kill the background process. No need for proprietary service.


An open platform would have a framework that supports multiple back ends.


This would be definitely nice to have but why does core Android itself need to have support for every feature?

We can have open source OSs on top of Android that build with a specific service set. Sort of how Ubuntu builds over Debian.


It doesn’t have to do anything, but if it doesn’t support what are considered standard features of a phone today in an open manner, it’s hard to take AOSP seriously as an open platform.

Also worth noting that when someone does build on top of AOSP, they are at the mercy of Google’s roadmap for the closed versions of Android.


To clarify - I should have said: “it’s hard to take AOSP seriously as an open phone platform”


It already does. https://bubu1.eu/openpush/

The fact that this is possible is something I have commented about on this very forum for years. It is trivial for the Amazons and Huaweis of the world to build an API like this that supports Google's FCM, Amazon's ADM, etc. and provide a library for developers to use that will let them easily deploy push-enabled apps on devices across all ecosystems.


Already does?

“The OpenPush project aims to create ... Development is still ongoing.”

Doesn’t sound like it from the website.


It is already usable. Nobody would claim that a Linux distro isn't open just because it doesn't have a built-in push API. You can build one on top for other apps to use, and you can do the same on Android. This is notably different from iOS.


Is it already usable?

The repo says it is unfinished, and commits seem to have stopped a year ago. It looks like it has been abandoned.

Has it been used in production somewhere?


Here's yet another option that does essentially the same thing but doesn't provide a reference implementation for self-hosted push server. https://github.com/onepf/OPFPush/tree/master/samples/pushcha...

I wouldn't be surprised if Amazon has their own internal implementation of the same thing to use in their Android apps.


So, no. It’s not already usable and there are no complete alternatives you can point to.

You speculate that Amazon may have an internal framework for this, but so what? Even if your guess was correct, that’s not AOSP, and it’s not open.

Your statement that AOSP ‘already does’ this simply isn’t true.

It seems like you don’t distinguish between things that are possible in principle, and things that are actually true.


You're essentially suggesting that Android should have Tor built in. The point is that it's possible to use other push services on Android and has been done, just like it is possible to use Tor on Android. There is no reason that AOSP itself needs to implement Tor any more than Windows or AOSP needs to provide a push API. Both are possible to implement on top of the platform and have been implemented on top of the platform.

This is unlike iOS, which is the real broken platform.


Why are you talking about iOS?

Also who said anything about Tor? What relevance does it have?

Of course any open source project can have features added in private. Those are not part of the platform.

It’s certainly not true to say those features already exist in the platform, as you falsely tried to claim upthread.

It also doesn’t mean you can add them to the platform if the maintainers don’t accept them.

As I say, you are confusing what is possible in some other reality with what is actually true.


Push messaging is a service, just like anonymous proxies are a service. It doesn't make sense to put them into a platform if they can be built on top of the platform. It would make sense to put something like this into iOS because it can't be built on iOS. It doesn't make sense to put it into AOSP.


Earlier you said AOSP had a framework for push notifications, a statement found to be false when checking the links you provided.

Now you say it doesn’t make even sense to put this feature in the platform, confirming that you were lying when you said it already was part of the platform.

Why are you talking about iOS?


AOSP has a framework to support multiple backends. It's called the Service framework. With it, you can build a library to simplify dealing with backends, and I pointed you to one such library.

> Now you say it doesn’t make even sense to put this feature in the platform

Of course it doesn't make sense to build that library into the platform.

> Why are you talking about iOS?

iOS doesn't provide the ability to make this possible, which is probably why you are so confused.


“I pointed you to one such library”

You pointed me to an abandoned codebase. When I pointed out that it seemed incomplete you said:

“It is already usable.” ( https://news.ycombinator.com/item?id=25361587 )

Do you still claim it’s already usable? If not, why did you say it is?

You say: “Of course it doesn't make sense to build that library into the platform.”

I agree, but nobody is talking about building that library into the platform. The comment you are replying to is talking about the feature, not that library:

“Now you say it doesn’t make even sense to put this feature in the platform”

Do you think it makes sense to put this feature into the platform?


> Do you think it makes sense to put this feature into the platform?

How many times do I have to tell you that it doesn't?

Do you think AOSP should have a framework specifically for uploading photos to a photo sharing service in the background? That is another standard feature for smartphones. Of course not, it just needs the ability to support uploading photos to multiple services, and people can build libraries on top of the platform to make it easier. This is notably another thing that is not possible on iOS.

> You pointed me to an abandoned codebase.

And another one that isn't abandoned.

> Do you still claim it’s already usable? If not, why did you say it is?

No. I assumed it was because they had already announced it. Any developer who would like it can simply finish the library themselves.

We seem to be in agreement except for your wacky ideas of how much should be in the platform vs. how much should be built on top of the platform.


> No. I assumed it was because they had already announced it.

You said it was already usable after I said that the link you posted showed that it was incomplete.

I.e: You made a claim that you didn’t know was true, but stated it as fact, even when you were asked to confirm it.

This is known as lying - saying something is true, when you don’t know it is true.

In this case I didn’t even call you on it - I just asked whether you were sure it was usable.

Your response was to immediately reply with a false statement.

A true answer would have been: “I don’t know, but I assume so because it has been announced.”

Because you do this, what you say simply cannot be taken at face value.


Says the guy who has been caught lying many times after he finally found a single time that I lied. Sure, buddy.


> Says the guy who has been caught lying many times

If this were true, you’d be able to link to an example of me intentionally stating a falsehood.

> after he finally found a single time that I lied

This is the single time you have confessed to lying.




Consider applying for YC's Summer 2026 batch! Applications are open till May 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: