there are discussions about less intrusive update messages, but I´m asking if it would be possible to handle it (optional) like discord or signal.
I really like Joplin, use it on multiple devices not every day. So I get many, many....many Update messages. I could disable updates at all, but I also had the message, that the server is different and I need an update.
Discord for example installs on start updates. Signal has a only a message at the top: "A new update is installed, please restart!". For Users who want full control should disable it.
- Would this also a way of handling updates for joplin?
- Does Joplin really need always a full update auf 150MB when maybe just some parts are changed?
Of course that would be better, but who's going to implement all that? In fact, come to think of it, perhaps that could make a good GSoC project as it's quite a bit of work to get all this in place.
I am curious about that. When you put Joplin in the Mac AppStore, would it give you benefits in den update process?
Same for the Windows Appstore. Is it too cumbersome, expensive or why are many developers not use them?
It's expensive and cubersome. The project you are comparing Joplin to is a commercial application and they have several teams to take care of the required infrastructure and software changes, as well as publishing to the various app stores. We don't have teams here so that's why it hasn't been done.
So we need a shout out for developers to commit and help with this!
The Windows store policies have been updated recently to allow publishing nearly any application and without having to use the .msix/appx packaging that was demanded historically. That's how you have a load of new applications being published there since the Windows 11 release. Joplin could likely now be submitted as is, but I wouldn't personally recommend it.
I think publishing to the Windows store and not trying to use .msix is a disservice, because you're otherwise wrapping around the usual .exes and .msi's (ignoring the legacy Appx format). Whereas actually using .msix gives you some nice technical advantages and actually ends up being fairly similar in approach to e.g the Linux Snap format. The updates get handled by the OS entirely, the updates are atomic, etc.
Using Msix might be a valid solution to silent updates by itself, but an .exe on the Windows store would offer no advantage over the current method (apart from visibility in the store itself)
I'm partially interested in looking into Msix purely on curiosity related to my existing packaging work, but I could try applying it specifically to Joplin to see the extent to which it's viable.
hihihi...my picture worked!
By the looks of it, making msix packages of Joplin would be fairly easy. Electron Builder already supports appx, and appx -> msix is a trivial repackage, in the future Electron Builder might end up just supporting msix directly too.
The more difficult part is managing the uploads and the relationship with Microsoft to be able to publish to the store. It also doesn't solve the problem with people not using the Microsoft Store version (which I'd honestly believe the majority of people would continue using the .exe even if hypothetically there was an MS Store entry). And I don't have a clue about MacOS neither.
Personally I think I'd be spreading myself thin by taking up the mantle there since I have enough packaging problems going on already, but I think Laurent has said before he'd be alright with somebody publishing Joplin to the store if they were so inclined.
Yeah, I understand. Maybe the benefit of "never think again about updates" would get more people to use the MS Store version. I would do it, how will the mass do?
Maybe a poll in the forum could give some infos. But I understand that you have many other construction sites at the moment.
Just a note about polls here:
Even though there are thousands of people using Joplin, here on discourse you have like a hundred people actively keeping up with the forum. Moreover, the forum folk is very different from average user (might I say non representative?).
So, running a largescale poll campaign about anything could be jokingly viewed as 10+ times worse in predictive power than flipping a coin
Not to be too negative, I believe polls here are actually useful when you want to get help, i.e. as a way to spark discussion, listen the community and list good options (in contrast to executing a plan relying on your personal hunch). It's just a cautionary joke about over reliance on statistics as a way to democratic decision making.
Thats a good point. The question is how to reach the users and get a clear picture?
Not really possible, most people will simply use an application without ever getting involved with the community or its development - I would wager many (most) people use Joplin simply because it is free, not because it is open source or the community behind it, most "normal" people don't get excited about application updating methods.
Some people come to the forums expecting it to be supported by some large company with dedicated members of staff assigned to particular roles rather than being community driven.
There are loads of applications I use frequently that where I have little to no knowledge of the development, new features etc. beyond skimming the change notes upon update and for those where I simply view them as a free tool I would be mildly irritated if I was suddenly prodded to participate in polls and discussions.
The other main public discussion areas are Discord and Reddit but both are far less active than the forums and both have the same issues.
User input is a way to contribute to a project.
Change my mind
Rant about Open source
Coming from the high moral stance, by its design the Free movement (FLOSS) is restraining itself in actively seeking contributions. When a project can, it goes without burdening the user or restricting his rights, i.e. doesn't do a vendor lock, doesn't implement telemetry, doesn't use advertising, etc
In commercial software, your question of user feedback would be easily solved by more telemetry and questionnaires pushed onto a user.
In Free software you get not only higher standard of code but code of conduct as well.
Of course, some concessions are necessary for a project to stay alive and prospering. Yet in those concessions you get much more transparency and choice in contrast to closed source software.
Personally, I would love to give my user input (and my data) to an open source project if it would help them even just a tiny bit. But in most cases, Free projects are surrounded by peers and users holding themselves and each other to the higher standard of morals. So much so, the "default settings" (what settings should be used as a default) are heavily debated on half of the all features, let alone implementing more controversial ones like opt-in telemetry or, god forbid, ads.
RIP Mozilla: even FOSS fans are leaving you. Good night, sweet prince
So, how to get more people contributing? Contribute yourself
The lack of telemetry definitely makes it harder to know what development should focus on. Sometimes you implement something and you never hear anything about it, which could mean:
- Nobody's using it, and it could be removed.
- Or everybody's using it but it works so well that nobody ever complains about it or ask for improvements
The best we can do I guess is to listen to user's feedback and create a polls about certain features now and then, and that way we get a sense of what should be improved first.
It's sad we cannot make a poll about implementing opt-in telemetry... or can we?
As a case for telemetry, checking the checkbox might be an only way to show support for those who want to help but just can't in any other way.
Say you live in Kosovo, you don't speak English, internet users there are minority, 10 EUR donation is about 10% of your hard earned income. Like what else can you do?
I could think of a creative, non-intrusive way to invite users to participate.
For example, if in the App, there would be a "Next features" link (unobtrusive in the corner), that even could be disabled in the settings, and lead to the polls page. More users could get curious, than a hidden link in the Help menu.
Or the poster I made (look above) could be one of the default empty pages, when you first install the the app and a request to please look whats the next hot feature in the polls.
I understand what you mean. It's true that tighter integration with the forum (and its polls) would be great as well as many other great features (or refactoring or documentation initiatives etc). Backlog is never ceasing.
In that regard, feedback is like infrastructure (bridge or road), you always need it in a specific moment but have to maintain it perpetually (especially when you're not using it). Therefore, infrastructure is expensive but with the effort of community we can make it less so.
For example, if you know a few things about design (based on your pic) you may create a mock up or a few of the feature you're suggesting. If it's good, that would reduce the cost of the feature development in the future.
Some skills are not so tangible but still very valuable, like marketing. You seem to be interested in discord: you may share around our discord server in note-taking adjacent communities or even help there with the support. That would surely increase the number of people involved in our polls and costs exactly zero in developers' time.
Like in business, the more you spend (time or resources) on your idea, the easier it would be to convince people to try it. Of course, you don't wanna go overboard (keep it cheap). So if you've got a skill (or interest in acquiring one), it's better to put exactly that skill to use, pitch a crude prototype and if successful get to work on better version.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.