Boden cross-platform framework: Native C++11, native widgets, no JavaScript

Deferring destructors is a suspicious pattern in general - in C++, I generally expect them to not be async and unpredictable like that. What guarantees do you make wrt destruction order? I hope it's not as complicated as finalizers in Java and C#...

The more logical model, to me, would be to have child views owned by parent views, and to only allow ownership-changing calls (i.e. adding or removing a child) from the main thread. That way all you really need is unique_ptr (from parent to children) and raw pointers (from children to parent, and from any observers). Although it would probably still be better to use shared_ptr just so that observers can use weak_ptr, since untangling lifetimes in callbacks can be tricky, and often it's easier to just check if the object is still there.

To give a specific code example, with unique_ptr, the same snippet would be:

    _window = std::make_unique();
    _window->setTitle("AwesomeApp");

    auto button = std::make_unique();
    button->setLabel("Hello World");

    // The following *moves* button, such that _window takes ownership over it.
    // It can only be called from the main thread.
    _window->setContentView(button); 

    _window->requestAutoSize();
    _window->requestCenter();
    _window->setVisible(true);

And furthermore, _window wouldn't have to be a pointer at all - it can just be a member of MainViewController.

Flutter, out of Google is a solution for iOS and Android cross-platform development that includes very fast UI primitives, a edit and continue workflow, intuitive and reactive UI framework, and a nicer high-level language to work with than C++ (Dart).

I had to install Java and Eclipse to give it a try just to reply to this comment.

I never liked Eclipse, it always felt extremely non-native to me. Nothing much seems to have changed.

Right from the start, the Eclipse Installer (by Oomph, apparently) is a web app that VoiceOver has difficulty with meaning I have to interact with it using web navigation controls which is a pain when I'm not expecting it.

The rest of the interface was a bit hit and miss. Either VO could read what was on screen, or it thought I was looking at a table that apparently had no elements in it.

Turning VO off and clicking around, still definitely not a native experience, though better than I remember it. Native-handling of text, access to macOS Services… but only in certain parts of the program!

After poking around a bit more, it turns out the only places I could get native handling of text were web views (of which there are many); most text presents non-native controls to deal with things like copying and pasting, precluding the ability to use built-in Services.

Besides that, the way the app is designed was just foreign. Non-native paradigms for presenting information like those tear-off palettes (that turn into weird, full windows when torn off, rather than inspector palette windows; if you click the internal (non-macOS) minimize button on one of these windows, the window itself stays the same size, but the UI inside gets smaller, leaving this huge empty window with nothing in it.

I was curious to see what VO would say about this window. It was just as confused as I was, thinking one of the two remaining on-screen buttons was a checkbox. That's probably how it was implemented internally, but it was certainly not a checkbox. Once I clicked on the internal restore button, the rest of the UI came back, but VO couldn't tell me what it was looking at. I had to Tab around blindly to get to a usable control, and even then, VO couldn't tell me where it was or in what context.

I couldn't actually create a new project because the "Finish" button, when creating a project, didn't seem to do anything, either through VO or using the mouse pointer directly. The button was lit up and coloured as though it were selectable, it just didn't do anything. So I couldn't actually play with the IDE itself, but I knew I was done with it.

The design looks and feels like Windows in the late 90s/early 00s, from the layout of various windows to those weird tear-off sidebar things. I especially despised the tiny icons that all practically looked the same, conceptually blurred together, and I couldn't find a clear way to make them bigger. VO saw each draggable toolbar as separate, making the toolbar at the top unnecessarily difficulty to navigate. Then those same icons infiltrated the global menus, making those a mess to wade through.

I'm a great believer that there is absolutely nothing stopping cross-platform desktop software being a first class citizen on, at the very least, the three main players: Windows, macOS, and Linux. Sadly, Eclipse, like most Java software, thinks it can get away with the last common denominator stuff and force it on other systems. It just makes for a rather unpleasant time, whether VO is on or not. Native widgets don't make up for non-native design patterns.

Sorry for the novel!

As to NVDA, I expect NVDA uses some heuristics to figure out how to read out all the myriad types of interfaces there are on Windows. I remember one app for vision impaired people, I can't remember it's name because I wound up passing on actually buying it, that basically constantly scanned what was on screen and made informed guesses. I'd say VoiceOver probably relies more on well-made apps conforming to system guidelines; if so, it's somewhere between naïve and brilliant, because I personally favour a system where accessibility is a first-class citizen, not an afterthought for which we need to call upon the powers of magic to figure things out.

Basically that, yes. I don't think people are necessarily conscious of the association, though.