Discourse as Android Application

(Tarak'ha (Sara)) #57

If it goes to Pushbullet and others, then all the better. :+1: Good to cover more ground.

(Michele) #58

I agree it’s most important to get this done with at least one, whichever works best.
If it can then be worked out with multiple services such as pushbullet and others, sure.

I believe it would be a life-changing addition towards increasing mobile Discourse usage, at least from what I’m seeing with my friends.

Not an android app, there’s no need. Just push notifications working someway, somehow.

I unfortunately have no experience with this so I have no idea how to make it work.

How can we proceed, assuming there’s a fair number of people interested?


I use email for that. :blush:


We can start by creating some sort of features spec that the plugin should follow. Then some dev can try and hack it together. I’ll try writing up something the next days.

(Michele) #61

So do I, but I cannot ask people who only check their e-mails once a month to check if they’ve been invited to a beta to change their wrong habits.

My priority would have to be to make it so it is a seamless experience for them just as it is for me.

@Moter8 Alright, sounds cool. I don’t think it has to be unnecessarily complicated

(Alessio Fattorini) #62

Totally agree with you, we can’t undervalue this aspect


And you are sure that they will check the discourse push notifications rather than email push notifications? But this is just bikeshedding, sry… I’m out.

(Michele) #64

Actually, yes. That’s the current standard for smartphone interaction, mostly through apps like Whatsapp, Telegram, Facebook both notifications and messenger.

Youngsters on mobile are used to being notified that way, so it is only normal that’s what they expect.

Not everyone uses e-mail, I know for a fact many of my friends don’t, as I am still in my twenties and I see on a daily basis how people use their mobile. It differs from my use, but I’m a minority and they’re the high percentage.

Your points may be correct when considering your point of view, but one has to understand there’s plenty different, and sometimes it is worth seconding others’ views and habits.

(Tarak'ha (Sara)) #65

To add what you already mentioned: Emails add a single ‘receipt’ for every notification. That is overkill for some power users. In fact, I currently have Meta here to send me no emails. I would fully prefer push notifications so I am still alerted without having Discourse in a tab on my browser. I have Discourse in a tab when possible but I’d honestly rather not because with push notifs that’s one less tab eating my RAM. Also, as of right now, all I have for a real-time papertrail-less new activity indicator is the favicon numeral-count. It tells me nothing else, so when I see a number I go to check and find out the new topics and replies are something that A. is not relevant to me or B. something I could honestly check later. I have ADHD so I personally need more of ‘what’ the new posts are so I don’t break my workflow to constantly check the open Discourse tab. Push notifs allow me to have more ‘mental freedom’ and less anxiety wondering if I’ll miss something as it happens. I use push notifs this exact way for gmail and I can say that it has reduced my stress, anxiety and I no longer worry exactly what is new. So I check less.

Furthermore, with services like Pushbullet, I get everything from my mobile onto my laptop. With what I mentioned above, I can integrate both my Nexus and my laptop together with Pushbullet.

Coupled with the fact that I plan on opening at least two Discourse instances, all those extra tabs add up, plus the checking anxiety.

(Michael Downey) #66

QFT; this is really important. Using Pushbullet or any proprietary service platform is a Bad Idea for an open source project.

Said proprietary service could go away overnight, change its terms of service to exclude parts of the world, or decide to fork a copy of all its data to the NSA.

Standards exist for mobile notifications on the various OS options. Compatible open source solutions should be made to work with them, to ensure long term sustainability and access to everyone.

(Bcguy) #67

Folks - I’m getting close to moving forward on this. My plan is to do this as a private effort that I will keep private for a year or so - then ultimately push it out to the public domain.

The plan is to initially do an Android app, then an Iphone app immediately following that.

I’d like to keep it as open source as possible - so if there is a good open source notification approach, please let me know. Otherwise we may go with the proprietary solutions for expediency sake.

Also - if we go with the webview approach for the forum which seems like the best immediate approach, are there good approaches that are recommended for caching, or pre-fetching (scheduled fetching) of messages so as to minimize the delay / time lag of seeing the more current messages at app launch time?

Any feedback appreciated. I’m going to be working with a contract developer on this project - I’m not the developer myself, but I’m busy developing the app specification.

(Tom Newsom) #68

Plenty of options availlable. This Stack Overflow thread is a good place to start, but Pushd seem to be the best at a casual glance

(Erlend Sogge Heggen) #69

There’s a new “magic app creator” in town called Manifold.js


My basic understanding of it after a quick read is that it assists you in the process of wrapping up your web app for native platforms. The more compliant your app is with the Manifest for Web Apps the better. It will attempt to fill any missing gaps by using Cordova.

There’s an on-site generator that you can feed live web apps to, and Discourse fared pretty well. It just wants a bunch of extra icons added.

(Kane York) #70

Note that Discourse doesn’t use an appcache manifest at all, though it might do well with one - would just need to make sure to timestamp the PreloadStore values, I think.

(Michel Alexandre Salim) #71

That would be great. Right now adding Discourse sites to my homescreen still launches them as a normal browser window, with URL bar and everything.

Adding manifests would definitely improve the experience.

(Nilesh Trivedi) #72

I am working with Government of India and considering using Discourse for civic discussion and open governance in a rural district in India. Discourse as a website is NOT good enough. Moreover, the Android app that was linked a couple of times above, is NOT available anymore.

Villagers in India whose only network access is their 100$ smartphone, do NOT have or do NOT understand e-mail. They understand signing up using their phone number, which isn’t easy to enable in a Discourse install. When Discourse sends notifications on email, it will open their E-mail app, not the Discourse app, causing further friction. Adding a website to homescreen is also something has some friction and therefore not many people do that, even though it is supposed to be easy. These users are also very price-sensitive and would want to minimize data usage for this app, but for now, I am assuming that Discourse website already handles it well using caching.

So, a Discourse app, even based on webview, would be amazing.

(Michael Downey) #73

Some useful data from the 2nd quarter of 2015 that is related to mobile web demand vs. mobile app demand:

For example, Asia is 90% mobile web usage and 10% app usage. Africa leans toward web even higher, and Middle East falls just behind Asia.

The US and Australia are the only places where app use exceeds mobile web use.

Full report: http://operamediaworks.com/innovation-and-insights/state-of-mobile-advertising-2015-q2

(Erlend Sogge Heggen) #74

Did anyone try “Chrome custom tabs” yet? Hopefully they’ll get some more attention with the introduction of Android 6.0 “Marshmallow” (they are backwards compatible with 4.x and up though).

This could be a very appealing workaround to Discourse’s “JavaScript on Android” performance woes. No, it doesn’t attack the heart of the matter, but it could do wonders for that 90% of your browsing time spent consuming your five favorite sites that you can actually be bothered to install a dedicated app for.

Combined with ServiceWorkers, the distinction between a custom chrome tab app and a native app would be fadingly slim.

(Sam Saffron) #75

Closing this in favor of

(Sam Saffron) #76