Hacker News new | comments | ask | show | jobs | submit login
Open Source Messenger App for Android – Real-Time Messaging, Voice and Video (github.com)
76 points by yumoji 8 days ago | hide | past | web | favorite | 59 comments

Please no. Not another messenger nobody will use. We need to improve the existing, safe solutions like Signal instead of pushing new ones out there.

Signal intentionally breaks interoperability and disallows forks and alternative implementations. Let's not standardize on something that backs us into a corner.

So general public will always use fb messanger / whatsup forever. Awesome.

Signal is only atleast somehow secure app that regular people use because it has great unified ux and just works.

Just use Signal. Please

Yeah no way. WhatsApp already does e2e. Signal doesn't bring anything else that's significant for me as a techie, much less the general population.

Matrix can replace Slack and the various mobile messengers, meaning there's a hook for companies as well. I don't think it's ready yet, but I'd rather be a condescending preachy asshole to my friends and family only when there's a very compelling reason.

"The perfect is the enemy of the good."

Onboarding friends and family to Matrix or XMPP is almost impossible. Signal is dead simple and automatically discovers your contacts, while being secure, open, and developed by a non-profit.

Ideally I'd like it to be a federated system, but I'm not going to use untrustworthy systems like Messenger/WhatsApp because I'm not getting every feature I want.

Signal is not open. That's the problem. You are forced to use the official client.

Thats also probably the main reason why it is sucessful. Everybody gets the same. They focus on one thing and don't delute it with funky in a year maintainer lost interest rust clients.

I don't care. Give me ethical, privacy minded software not some phd experiments.

You don't care about freedom, got it.

>Yeah no way. WhatsApp already does e2e. Signal doesn't bring anything else that's significant for me as a techie, much less the general population.

You mean besides freeing them from the Facebook ecosystem? That alone is worth it to me.

Sure, WhatsApp does e2e encryption.

Guess what company is also on both of those ends? If people could stop parroting this as if fucking Facebook was trustworthy, that'd be great.

Facebook ads: Coming soon to your WhatsApp conversations!

I might use it once it allows federated access, otherwise only WhisperSystem can control the fate of the protocol. I mean there are much worse company out there, but I don't want to put my eggs in the same proverbial basket.

I refuse to use a messenger that is spearheaded by a man who adamantly refuses to participate in the wider FOSS community. Moxie only wants others to play by his rules.

"Great unified UX" does not include their Mac app unless there have been big changes since last time I tried it.

You have to pardon me. By unified i meant mobile. I am not sure about state of desktop. I from time to time use mac app with no issues but Signal in my eyes is sms alternative.

It works OK, but it's no iMessage.

I miss using Adium as my native Mac chat client for everything, alas AIM is dead and Google killed their Jabber interoperation.

I'm not a heavy multimedia user, but Signal's Mac and iOS apps have been working wonderfully well for me and my friends.

You mean whatsapp ?

This. Matrix/Riot is federated and works. Lets please use and support systems with open standards that have the perspective to interoperate in that space (irc, telegram bridge etc)

Considering the fact that Signal devs refuse to improve several glaring user experience holes (things like losing all group chats when switching phones, which are deal breakers for many people), we might need another messenger.

Use the built in backup and restore and you dont lose your groups. Signal's primary design direction is encryption and privacy, you're going to have to compromise somewhere.

Signal forces you to use a phone number(unlike facetime for example)

Which requires to use apple id.

And what's your point? An apple id can be anonymous.

You probably missed the point here. Messenger is just a use case. The license is pretty liberal and you can reuse the code in pretty much any app, not necessarily Messanger. social, dating, customer support.

Moxie personally rejected to allow to use it's code in mobile apps (since they are GPL) even when we was also open source (MIT) project.

Are you seriously objecting to someone just asking for the license the code is distributed under to be followed?

i am saying that he enforces that no actual fork can be done and don't allow even simplest exceptions for AppStores.

Never heard of Mesibo but I've been waiting for a solution like this for a long time.

My dream of creating a new kind of messenger just got a little closer.

In what way is this better then other similar apps, such as Signal, which is already an open source messaging app?

I have been regular, everyday user of Signal for over four years. Myself and my friends use it as our primary communication method. In the last two or so years, it has been what I can only describe as "complete and utter shite".

When I send an image to a friend, it can take hours to send a half-megabyte image despite my internet connection being fine. When this happens, none of the consequent messages send. If I close my phone, messages stop sending. So I must violate my privacy by keeping my phone open for sometimes, hours at a time, just to send a stupid little image.

Despite the fact that sending an image will cause consequent messages to be delayed by hours, messages have started, in the last 6 to 12 months, to be sent out of order. This can be disasterous, as a fair number of my friends have depression and other problems, which means (as an example) my "haha" responses can sometimes be seen to be responding to /other/ messages (And yes, I can use the quote function, but sometimes that interrupts the flow of messaging).

This, plus stupid little problems like, backups not being tested to be forward-compatible (something that is ridiculously easy to test, but that they could not be bothered to do), mean that relying on signal is a fool's game at this point.

Signal has some serious problems and I'm glad other people out there are talking about this, too.

I ended up stopping using Signal (except for a few holdouts) because of the message delivery problems; it caused way too many social problems.

I mostly use Telegram now, and it's been great. The worst part about it is everyone saying "but it's russian spyware!!" whenever I suggest it. It isn't, folks, and I'm well apprised of the theoretical vulnerabilities. They're just a compromise I'm more than happy to make, when the alternative is a chat tool that fails at chatting even 5% of the time.

I've had the same issues and a similar experience after using Signal the past few years. It took so much effort to convince my closest friends to install Signal and the single most frustrating thing it would do is force us back to SMS because sending of encrypted messages would continuously fail.

What really sucks is that there's no hope of me convincing them to switch again if a better messenger were to come along.

Using Signal for years also. I don't have any of these problems except during group messages. Group messages for whatever reason are inconsistent, don't send, or sometimes break apart into single messages and then join again into the group. There was one group message between me and my two cousins that broke for hours. I didn't get any texts or couldn't send anything and signal was constantly trying to download images.

I must contradict your Signal bashing really.

I'm a heavy signal user for about 5 years, never had any of the issues you're describing. Regularly sending 3-4 minute videos it all goes through fine. Exported/imported my chat history 4 times now when changing phones, never had an issue. Never seen messages to be sent out of order either. I really wonder if we're using the same app!

This is amusing. I am not "Signal bashing", I am talking about problems I (and many other people) have experienced.

The fact that you have it good, and have never had problems, does not contradict the fact that I and other people have had problems. Talking about these problems is not 'Signal bashing', just like talking about traffic problems is not 'car-bashing'.

1:1 chats have worked fairly well for me, but groups have been utter chaos with messages being delivered only to random subsets of the group and things like that.

I havent had these issues. Actually i started to use desktop app with groups for light colaboration - sending audio notes, pdfs and images.

I am thinking if this is happening could google be throttling signal somehow? That is the thing i have always worried about.

> google be throttling signal somehow

What does Google have to do with it?

As far as i know signal uses play services for transport.

I totally agree. Signal's UX is awful (regular user here).

Yes, or using this: TDLib – Build Your Own Telegram https://telegram.org/blog/tdlib

I don't think that post is accurate, perhaps because (based on the RSS feed) it is from 2015.

> Telegram’s source code is not an SDK or a library. [--] But to be fair, the messaging app doesn’t state they have an SDK. All they did is put their source code in the open.

TDLib is that SDK.

Are we able to run our own servers with your lib?

Good question, refer to on-premise deployment section @https://mesibo.com,

Made by the guys behind tringme? https://github.com/mesibo/tringme.github.io

As of now the backend code repository is empty. Let's wait for a few days and see where this goes.

"getting started is as simple as downloading SDK and copy-pasting sample code"....

uhhhh yeah, I'm sure that will go well.

Getting started to modify the software; {app,play}store links are given.

Not a single test in the repository makes me a bit suspicious, quality-wise.

The code is quite messy in general.

Hope to improve it with your help. Pls fork and contribute.

Some of it could be cleaned up fairly easily - like the use of Yoda conditions and other anti-patterns, numerous spelling mistakes, dead (commented out) code, redundant casts, unused variables etc. But the major problem is no separation of concerns (business logic is tightly coupled with UI logic and Android framework dependencies), which makes writing any automated tests for the thing next to impossible right now. Resolving this issue would require major refactoring.

Have you seen Signal-Android's codebase.

What The Fuck ?

There is no point talking about an app that implements a messaging protocol without talking about how that protocol compares to existing standard messaging protocols. I couldn't even find a comparison to XMPP.

It works, I honestly didn't expect it to be so easy. It complied without any errors and running. However it is not WhatsApp yet at least the UI is not as good as WhatsApp but a great start and one of the best open source implemention.

Native Desktop apps would be a nice addition and you have something that could take on the Facetime/iMessage ecosystem.

Applications are open for YC Summer 2019

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