The Case Against Marzipan
uluroospeaks.com
uluroospeaks.com
The four Marzipan apps Apple themselves have shipped in Mojave are garbage by the past standards of the Mac and I worry this telegraphs that this kind of shovelware is fine. This can establish a kind of "If it's good enough for Apple..." precedent.
I also find it pretty hypocritical for Apple to constantly tell us they won't ship a touch screen Mac because you can't make an OS that is good for both touch and indirect input, but here they are providing a framework to dump touch screen UI apps on the Mac for use with a mouse. Instead of a poor unoptimized touch experience, we get left with a poor unoptimized mouse and keyboard experience, precisely the kind of thing Apple have spent years telling us that not putting a touch screen on the Mac would prevent.
That said, I can sympathize with the counter-arguments in favor of this. Mac app development is nothing close to what it once was, perhaps this is the best we can hope for.
Turns out even though "you can't get your hand around it", people did buy that.
I think the real reason Macs won't have touchscreens is that Jony knows it'll result in disgusting monitors. Can't have that now can we?
Reassess in a couple of years when app porters have matured with the technology. Gestures and pinch and zoom won't work with a mouse but the key point is interfaces designed for a form factor.
UWP on Windows mightn't be a great success but there's no reason a universal toolkit can't serve both. c.f. Qt - a first class citizen on mobile (Sailfish) and on desktop (KDE), while Plasma Active has real potential on postmarketOS.
On Android, where I am typing this message, a good number of apps designed for tablet work surprisingly well on a desktop (Outlook, Apple Flinger, Hot Death, Taskbar, iview, SBS on Demand, Supertuxkart to name a few). A bit more polish for mouse in the form of tooltips, contextual menus, scroll wheels and menu bars and it'd be a nice interface.
I thought Qt Widgets were deprecated by Plasma 5, which uses QML internally.
Frankly, there are many cross platform services I use where I would prefer a straight Marzipan port from the iPad app over the electron psuedoapp approach that's so pervasive these days.
Seeing touch sized click targets on HiDPI non-touch displays as one does with the Marzipan approach is not great.
That's always going to be a thing, as long as desktop environments lack a simple, foolproof way of switching system-wide between a "mouse/pointer input preferred" and a "touch/swipe input preferred" mode. (Windows 10 actually does this fairly well; I don't know of any other systems that offer the same thing, though.) It doesn't have to be a huge issue if the touch-friendly interface is well-designed. See GTK3, which has essentially switched to "touch-friendly" widgets and layouts with the new major release. Some people gripe about it, but at the end of the day it works fairly well - and at the same time, it lays the groundwork for a wholly touch-based interface to the Linux desktop.
Except that Apple spent a decade telling us it won't be a thing, which is why they never put a touch screen on a Mac. Until 2018 it had never been a thing. Supposedly this is iOS's remit, and designing the entire OS for solely touch (iOS) or solely mouse and keyboard (MacOS), they argue, is the better way forward.
Your arguments are perfectly valid for many other computer platforms, where mixed direct/indirect input exists. It has never existed on the Mac, supposedly by design, until now when Marzipan has provided a path for these touch UIs to ship on them. This is why it feels especially wrong on the Mac.
This is not the case at all. 3rd parties can't (officially) use the new APIs at all until next summer. I suspect (and very much hope) that the APIs and these apps will look quite different in 6 months
Conversely, it’s pretty likely an Electron version of XCode would run like garbage, a native version of VSCode would run faster and consume much less resources (memory, battery etc).
I don’t think it’s appropriate at all to argue VSCode somehow compares favourably.
But the comparison should be about how mac-like the applications are. if vscode or xcode are outliers, some other electron or Mac apps could be compared.
Ahh, found the answer: Home, News, Stocks and Voice Memos
The benefit could potentially be thousands of very useful apps that are well designed for desktop. Overtime the API and SDK's and libraries and best practices, etc, will mature.
Responsive design has been a thing forever in web design. I don't see why MacOS development can't adapt similarly. Ultimately two exclusive 'native' apps might still be ideal, but not always for everyone who can't afford to. So it's a lesser of evils type of situation.