It's definitely a big problem with software. This "need to constantly pay developers" is part of the justification for moving towards subscription payment instead of one-time payment. The logic goes: One-time payment won't pay for future changes, and developers are always making changes, and they need to be paid for. But, what if I don't want future changes? For almost all software I've ever paid for, as a user, I just want the goddamn software exactly how it is today, forever. I'd maybe be willing to pay for security updates, but that's it.
But, noooooo, the company decides developers must perpetually change the software and designers need to perpetually change the UI, and therefore you must continue to pay for it, regardless of whether you want any of those changes.
Funny, I was just thinking to myself that maybe a paid subscription Windows could be a solution to undesired UI change. Because it would take away the lure of the upgrade sale licence. 64 bit w2k? Yeah, I'd totally pay for that.
If I were in charge of Windows product strategy I think I would create competing subdivisions that replicate the model of Linux distributions but on the NT kernel. Sworn to binary compatibility but inspired to try different monetization schemes.
8 too? I was going to write something about this, particularly in how that's the reason why I don't really disagree with GP despite pondering the exact opposite "what if". But the sudden appearance of an 11 that nobody asked for gives me the impression that the culture that was fostered on the paid upgrade is still very much alive.
Meanwhile my mind has taken a deeper dive into the rabbit hole fantasy of competing Windows distribution divisions. I'd expect it to seem like a total failure : the surface. E.g. with Steam users refusing to get lured into the "gamer Windows" (something with an X in the name), the "crippled version that is entirely ad-funded" (10, basically) drawing too much hate for the entire brand while still not being crippled enough to push people to paid options and the one you can get with the Office subscription being a little too incompatible for employee lockout reasons with the plethora of in-house essentials required to run at customers. But Microsoft could still come out stronger, having a well defined place for bad ideas to go to die. And who knows, one of the niche distributions might be essentially sysinternals with a kernel and a package manager.
But aren't those usually on a subscription anyways? Could still be a factor, in theory: they might pull an Oracle pushing all their subscribers onto 11 unless they upgrade to the golden account tier that still receives updates. They actually did this, I think, with 7, but I assume not at a scale that could bring the pattern anywhere close to being a central motivation.
They are mostly on subscription; but what is the rationale for subscription? The right to upgrade, and the ability to have all machines running the same version. Remember, that in order for subscription license to be valid, it is usable only for machine that came with OEM license in the first place.
I once worked for a corp that didn't have subscription. The rationale was, that is was not important who was running which version at their computer, and the subscription was thus a cost on top of OEM license that didn't bring any value. Since the computers were refreshed every 4-5 years anyway, they always had a supported Windows version.
> But, what if I don't want future changes? For almost all software I've ever paid for, as a user, I just want the goddamn software exactly how it is today, forever. I'd maybe be willing to pay for security updates, but that's it.
Problem is, most software doesn't exist in a vacuum; it is part of an ecosystem. Security updates that you might pay for are not enough, because the ecosystem is constantly changing around the program you bought.
I'm author of HashBackup. It uses remote object storage services so if there are incompatible changes made on the storage side, the software has to adapt. What if IP4 is no longer supported? Adapt. A library like Boto2 gets rewritten and the old version loses support. Adapt. Google Sites decides to force you to migrate to a new version that looks like shit. This just happened to me, and it's so bad I'm having to do a whole new documentation site.
Just keeping software working has become a big challenge because of ecosystem changes.
Great point! It's an ecosystem-wide problem. I'd go even further to argue it's a problem baked into the very psyches of software developers. Your remote storage service should not be changing, particularly in ways that break backward compatibility. Nobody should be dropping IPv4. Nobody should be rewriting a library in such a way the new one is incompatible unless the old one retains support. Nobody should be forcing you to migrate to some crappier version that looks like shit.
Stop changing things just to change them. But if you're actually changing for a good reason, consider the downstream effects of your changes.
I can take a bash script from 20, maybe 30 years ago, and know that it will still do what it's supposed to do, and all the standard commands it calls, like sed, awk, and grep will also do what they are supposed to do. Bits don't rot by themselves. Careless developers in the ecosystem make changes and don't consider backward compatibility, and we euphemistically call it "bit rot."
That's a bit unfair. The iPhone 13 just came out and one of the memes and jokes going around is that it looks the same as the previous phone and users are just buying the same thing again. Even tech people I know are joking about it.
So the UI people are always going to be busy because even people in the industry equate looks to changes. If the looks don't change they will all be scratching their heads confused why this is Windows 11 when it looks identical to 10.
I think the main reason for memes is because Apple always markets everything they do as some kind of massive innovation despite either bringing something from the past or just copying what competitors already had long ago. They at least used to pretend they make something different with design changes, but now even this facade fell down.
That is why consulting and outsourcing exist, you only pay for developers while you actually need them.
There are even companies outside software business, that pay contractors per tickets, basically a budget is assigned and the contractor does as much tickets as they are capable of for the given budget.
Now the quality of the code is not for the faint of heart.
This ignores the changing nature of consumer demand and competitive pressure. Often time, implementations should be treated as a product that needs nurturing rather than a one off project where you get to the finish line and you're done.
apple podcasts is a fine example, among others, of this conundrum. on every iteration an "improvement" makes the app less usable. on the last iteration they decided that i should now "follow" shows and not "subscribe" to them. what's worse is that some screens had the old and the new names resulting in extreme confusion. on that same iteration, someone decided that the way i was deleting episodes wasn't cool enough. so they replaced a working feature with a half-working one!
Interesting. The Podcasts app was a disaster when it was released. I honestly can’t imagine how it could have gotten worse.
I’ve been using Podcruncher since 2010. Manual downloads, no hiding episodes, custom playlists. All in a straightforward interface. It is amazing that Apple with its resources can’t deliver anything comparable.
Damn... the idea that maybe we should be banding together as a democracy and cutting deals to subsidize the engineering work of companies that are willing to "lay fallow" their design teams is fascinating.
Bonuses paid to design teams that are a direct function of the length between unprompted UI adjustments and redesigns?
Iteration on designs is of course permitted, but there’s more pressure on “get it right” rather than “dirty MVP”. Once launched, changes can’t go to prod-lest they lose their bonus-and emphasis is more on more in-depth and longer-term user research and refining the design.
This way design teams are still doing valuable work, and there’s no crunch to both design something and then get it out ASAP. The next time front-end teams come back to work on any kind of UI refresh everything is already sorted and ready to go.
