|
|
|
|
|
|
by ChuckMcM
411 days ago
|
|
|
This article and the comments here are kind of scary. It feels like Google's only supporting apps that "drive engagement". That sounds like they want to force developers into producing stuff they can show ads in because they need more ads, not like they need more apps. It also feels a bit like how software people STILL haven't figured out how to deal with a product that has a finite development cycle. Which is to say, a piece of code that is done and doesn't need any changes. You don't have Hardware stores forcing supply companies to come out with a new version of shovel every year right? A shovel is a shovel. There are probably 8 different types for various uses and within those perhaps two or three variants. So 24 or 30 variant of 'shovel' and your done. Some software can be like that too. The subtext though that Google is actively hurting their developers for unspecified goals which look like they are desperate to make more money but it certainly could be some other thing. It reminds me of all the wailing about people whose web pages fell in the rankings because they hadn't been "updated" but when you've got the most useful description of say the scientific method on the web, why should you need to update that? It hasn't changed. And yet the 'older' your page got, the lower and lower it ranked. |
|
|
The problem is that platforms these days are in a constant state of slow rug pull. Even if you have absolutely no bugs to fix and no new features to add, you still need to keep things updated just to make it work on the most recent version of the platform (which users are going to be on because that's the only one that receives security fixes). A slightly less damning case is when the app works but doesn't integrate well with the new parts of the platform, or even just its changing look and feel. E.g. old Windows apps often work fine but don't support hi-DPI properly, meaning that they look very ugly on that 4K display.
I don't think it's a problem that can be fully solved, but the impact would be much less severe if platforms stopped churn for the sake of churn. For example, we don't need a "fresh new" UI redesign every 3 years. And when it comes to API stability, Win32 should be considered the exemplary model of that - yes, it is a lot of effort to keep things working 30 years after they first shipped, but that's the only way if we don't want to be an industry that's constantly building castles on sand.