Sounds like a great way to create incentives that undermine githubs mission while simultaneously taking a huge amount of developer resources for a product that will have niche interest at best.
Building app stores is hard. Moderating app stores is hard. Both are significantly different from running GitHub.
While I appreciate the author's intent here, this just doesn't feel like it was deeply thought out.
I'm not convinced by any of the arguments the author points out which would actually address the problem of App Stores.
Essentially they are arguing for Github (Microsoft) to host and have an App Store on their platform and control it because they listen to their users, which is quite frankly a very poor reason.
Cydia and F-Droid packages are already 'App Stores' which you can install apps via pressing 'Add to Cydia' or 'Add to F-Droid' buttons on the publisher's website. You're free to add whatever app and install whatever app you want without Apple / Google stopping you, and its decentralised.
The same has been said for desktop apps (Native or Electron), which have been hosted on the publisher's websites for years, without the need of an App Store. I can imagine an 'App Store' equivalent for Desktop which works like Cydia and F-Droid which allows users to discover and install an app with a simple deep-link into the App Store. Linux has this already, equivalents for Windows and macOS are possible but it should not come from GitHub.
The above approaches are already there and are completely decentralised. Unlike the author who is going for another centralised App Store in the hands of GitHub / Microsoft (again).
I don't know though, Microsoft already runs multiple app stores (Xbox & Windows, Edge). I don't see why they couldn't build it out for another platform or at least figure out some kind of arms-reach integration with Github -> Microsoft Store.
This could be a nice come back into the mobile sphere they lost out on. As a developer, I completely agree that Microsoft earned my trust over the years.
> The pendulum has shifted, and it's swinging towards PWA.
It has for some things, and for others native is king. Native just delivers experiences that PWA can’t (yet).
To take a simple example - let’s consider a ‘timer app’. On iPhone I ask Siri to set me a timer for 5 minutes, it sets the timer and I walk away from my phone. 15 minutes the timer pings on my watch where I can accept it. There are multiple steps in this user story that a PWA currently struggles with, including:
These are all artificial problems created by Apple, though. The first three are all doable on Android PWAs, and you can do everything with a TWA in the Play Store.
I don’t disagree, although I think these reasons do mean that it’s far too early to say “App stores are dying” as in the post I was replying to - at least in Apple’s case they are still central, and even seem to be becoming more core with deeper integration into MacOS.
Also, in this context, a TWA isn’t useful as the post is about avoiding an App Store / central control using PWA.
FF only removed the SSB (site specific browser) feature on desktop which was in an alpha state. There was no announcement that Mozilla was going to remove the entire PWA features (such as service workers).
PWA is a set of modern web standards. "Supporting PWA" can mean many things but "Removing unfinished SSB" doesn't mean "Removing PWA".
GitHub has been expending its services, diverging from its original intent (e.g. remote visual studio code).
I disagree. I think providing that a way to distribute applications through GitHub, with builds certified as coming from the original authors, would be beneficial.
Agreed, I doubt GitHub will do this but you know what, anyone could start an App Store linked to GitHub (or GitX, or all of them, doesn't matter).
This has been tried in the past - Freshmeat, SourceForge, but that was a different era and also none of them tried to upgrade and modernize. So now there seems to be a niche opening for this in 2021.
Make an App Store platform everyone would love to be on. Get developers on board by giving them first 2 years for free, then a moderate membership fee. Take a reasonable cut from IAP's against providing customer support for the payments/chargebacks/fraud. Just make it awesome, reliable and fair.
Building app stores is hard. Moderating app stores is hard. Both are significantly different from running GitHub.
While I appreciate the author's intent here, this just doesn't feel like it was deeply thought out.