Hacker News Reader

Comment by dt3ft | Hacker News Reader

.NET 6? Hold up, WPF was never really finished. This is demonstrated by existence of companies which sell lacking components that should otherwise have been included with WPF: Telerik, Infragistics, Syncfusion etc. Heck, creating a Report with out-of-the-box WPF is an unforgettable nightmare.

I, for one, would prefer that Microsoft, instead of pumping all these half-finished "cool new toys", ships stable, performant, feature-complete software. We either have to buy 3rd party components, or write our own components from the ground up. Sure, they provide a Button or a Combobox, but good luck if you need anything slightly more complex, like a searchable GridView, autocomplete Combobox, Charts etc.

View on HN

The WPF grids are fine and autocompletion is trivial. Which is the proper way to write frameworks: make it easy to make the bits you don't have.

I love that it isn't created like the old RAD frameworks where you are supposed to easily drop a datagrid in, connect to a DB and basically have some Line of Business CRUD on your screen within minutes. The Telerik/Syncfusion et.al. components are massive bloated and geared exactly towards that, and in order to do that they need to make a ton of assumptions about the MVVM/MVP/whetever architecture of the app. There are a few small omissions from WPF but I wouldn't say that the huge suites from Telerik and similar show that it's lacking.

So when purchasing Visual Studio subscription in order to write software for windows, you expect nothing to work out of the box? I'd love to see you tackle a WPF monolith within a regulated environment where no 3rd party libs are allowed.

It's 2021, we're expected to deliver amazing, feature rich software in very little time, without buying 3rd party libraries.

Meanwhile, there's Microsoft, using the funds from our subscriptions and building stuff like Xamarin Forms and abandoning it half-way through. I wish they'd spend less time fooling around, and more time improving the tools that have become bread and butter.

> stuff like Xamarin Forms and abandoning it half-way through

Eh? I've worked on projects using Xamarin Forms, and they turned out just fine.

Yes, it would be nice if there were a few more controls in the main Xamarin Forms library, but I don't see a problem with the community picking up some of the slack (which they have).

"I've worked on projects using Xamarin Forms, and they turned out just fine."

Same here, but it is my experience with Xamarin that has led me to conclude that it's the wrong choice almost 100% of the time. Sure, it can be made to work but at the end of the day you're coercing C# to do something it was not designed to do. For mobile I am convinced that PWA/hybrid mobile is the correct choice 95% of the time with the remaining 5% being platform native apps written in Android Studio with Kotline or Java (or Xcode with Swift or Obj-C). Anything else is just an exercise in sticking non-round pegs into a round hole.

> at the end of the day you're coercing C# to do something it was not designed to do

Hmm, not sure I can agree with this statement, as I don't see what would make C# itself unsuitable here.

> For mobile I am convinced that PWA/hybrid mobile is the correct choice 95% of the time

But this I fully agree with :)

While the Xamarin Forms projects I've worked on turned out well, I feel like the whole experience of developing, experimenting, building, debugging and testing would all have been so much easier with a PWA. A PWA would also make it much easier to build what your UI designer really wants, instead of having to compromise all the time with what is readily available - having CSS at your fingertips makes all the difference.

> So when purchasing Visual Studio subscription in order to write software for windows, you expect nothing to work out of the box?

If you mean I expect to compose my own complex controls out of less complex building blocks, then yes that's what I want and expect.

> Meanwhile, there's Microsoft, using the funds from our subscriptions and building stuff like Xamarin Forms and abandoning it half-way through

I agree WPF and others have been backbenched for 10 years. Microsoft panicked and needed a strategy for cloud and mobile, while they weren't exactly being threatened on desktop. So from a business perspective it made perfect sense: leave desktop for 10 years. Now they are slowly coming back to it with .NET 5 which is nice, but it wasn't very pleasant to be a Windows desktop dev between 2010 and 2020.

WPF DataGrid causes memory leaks - this is still the case in .NET 4.8

Is it still the case in .NET 5 though?

I think that it was intentional. Microsoft often comes under pressure for competing with their suppliers and this was the one example where they didn't even try.

It's also why no-one used WPF: without a grid it was worse than Visual Basic 6 for Microsoft's bread-and-butter business customers. This was a huge strategic misstep, as instead of hiring "WPF" developer you had to hire a "Telerik" or (shudder) "Infragistics" developer or spend a coupe of months extra onboarding new hires.

Lack of understanding why Telerik, Infragistics, Syncfusion etc. exist.

A library with a rich ecosystem of third party vendors instead of the clunky GNU/Linux offerings.

It's much better to have the foundations and build the rest yourself (or purchase what you need) instead of having the lowest-common-denominator come from MS.

There are a 1000 different ways to implement autocomplete menus and charts; that would be a nightmare to produce or use.

All the 3rd party library hanger-ons is a real putoff. Asp.Net docs linking(with broken links) to a 3rd party, paid identity framework to make auth work for Blazor/Spa? Gross. It's 2021 and this stuff is table stakes for OSS web framework ecosystems.

It remains a framework at the end, same as React. The difference is there is less concurrence, and developers can price up their components.

Reply on news.ycombinator.com