24

What is the recommended approach for building line of business apps for Windows 8? e.g. complex bespoke finance application

They clearly don't fit into the Metro style so presumably will be a regular desktop app

So is the recommendation to use WPF? Is there a vNext for this?

The type of apps I build as a developer are desktop style apps. They are not ones that would fit into the Metro style. We haven't heard much from \Build about this style of app.

So if you were going to launch a desktop app to coincide with the release of Windows 8 - what would be the prefered use of technology. (I get the impression WPF is on the way out)

Is there anything in Win RT that can be leveraged?

thanks a lot

ChrisCa
  • 10,876
  • 22
  • 81
  • 118
  • There ARE new features in wpf 4.5 (if `Is there a vNext for this?` meant this) – xanatos Sep 15 '11 at 20:45
  • 1
    Specifically, "What's new in WPF 4.5": http://msdn.microsoft.com/en-us/library/bb613588(v=VS.110).aspx. There are a lot of data binding improvements for more smoothless declarative coding. My favorite is that binding to observable collections (`INotifyCollectionChanged`) is now thread-safe just like normal properties always were - no more `Dispatcher.Invoke` in models. – Pavel Minaev Sep 15 '11 at 21:22
  • I cover the new features of WPF 4.5 in a serie of post on my blog too: http://www.jonathanantoine.com/2011/09/20/wpf-4-5-is-here-check-out-the-new-features/ – Jonathan ANTOINE Sep 22 '11 at 10:23
  • If only a WPF app could run on a ARM based window tablet… – Ian Ringrose Feb 05 '14 at 10:08

6 Answers6

7

WinRT is for building Metro-style apps. If the app you're building doesn't fit the Metro style, then just build it as a traditional desktop application. Not all applications will fit the Metro style, and they don't have to.

Chris Pietschmann
  • 29,502
  • 35
  • 121
  • 166
  • 1
    But here http://arstechnica.com/microsoft/news/2011/06/windows-8-for-software-developers-the-longhorn-dream-reborn.ars?comments=1&start=40#comments-bar they are even talking about a WinRT based Office. And from the vision i can't think why WinRT should not cover it - even if it does not now in the beta phase. – Lothar Sep 16 '11 at 06:48
  • @Lothar If you want to produce an application that will run on an ARM machine then you need to create a WinRT / Metro application, hence it may well make sense for Microsoft to create a cut down version of office designed to work on ARM based tablets. (although I can't see it being a replacement for the desktop version) – Justin Sep 16 '11 at 09:34
  • Yes, I know all apps shouldn't be Metro and I say as much in my question. I am trying to find out what the prefered desktop technology is. I get the impression that WPF is on the way out. I'd like to know what the Windows 8 way of doing desktop apps is. – ChrisCa Sep 16 '11 at 09:40
  • 1
    @Roonooir the answer would be your preference of wpf or winforms for desktop apps as winrt is only for metro apps. – Chris Pietschmann Sep 16 '11 at 16:58
  • Imagine I want to break the rules and make a desktop-style application in Metro. Could I not write my own drop-down-menus? Surely they can't stop me? (the API's MUST support 'rectangles with text'). – Jeff Oct 19 '11 at 21:40
3

Well, this is a bit old question now, but since situation is more clear nowadays (after release of Windows 8 and its tablet version - Windows RT with some new hardware by Microsoft) you may reconsider targeting WinRT instead of WPF for your new projects. This will allow your apps to run on tablets with Windows RT OS in addition to work on desktop in "Metro mode" and in the observable future (taking into account MS plans for unifying development for desktop and mobile by targeting project to WinRT) this also could simplify porting part of main desktop functionality of your financial app to its mobile counterpart version running on Windows Phone and allowing the user to do something useful even when they are on the go.

As for how much complex financial application could fit into "Metro" design and concept - you should have a look into several existing financial application in the Windows Store and compare their complexity with your possible requirements.

Another way to determine if WinRT will suites your desktop app is thinking about how complex will be breaking all functionality of your app into multiple relatively separated screens with input elements large enough for touch input by fingers - if that's looks doable, then you probable should go with WinRT.

But if the application is expected to show a large number of different tables simultaneously or, for example, should collect a lot of typed information by user (i.e. long descriptions, rich text formatting, lots of fields to fill etc.) then it probably will be usable in desktop mode only with hardware keyboard and large screen in front of user (i.e. on tablets/mobile it will be just painful for user to do all those things). In that case the WPF will suites just fine.

Sevenate
  • 6,221
  • 3
  • 49
  • 75
2

Suppose you stick with WPF as your target.

Then you could try to keep as much as possible of your code within the intersection of WinRT, Silverlight, and WPF. Maintain two (or three) projects and make sure that both (all) of them compile. Eventually you'll probably have to use something that's platform-specific to WPF. Simply stub it out for WinRT/Silverlight.

This gives you maximum flexibility no matter what Microsoft does with Windows 9, and allows you to keep track of how much of your code can be ported directly over. If they bring WinRT XAML to the desktop, you're covered. If they improve the Metro/desktop integration (a richer task switcher, say), you're also covered. If nothing happens, then you've only incurred a small development tax.

WPF will still be around after version 4.5. May not get many flashy new features that don't appear first in Silverlight or WinRT, but WPF is close to a mature technology anyway. Visual Studio, Expression Blend, and AutoCAD should be enough to secure its future support.

Tao Yue
  • 696
  • 5
  • 9
1

Yes the solution would be to ignore everything else and go to WinRT and metro interface because it's plain clear now , everything else is declared dead by Microsoft and there is one and only one in the future, the metro interface (modern UI). Creating a new project based on another technology it is investing on a dead technology, it could be a disaster for any developer. The problem is that no one want to use the one and only one technology that Microsoft is supporting, the so called Modern UI.

The situation is almost PARADOXICAL.

Everything is declared obsolete except the modern interface, so to create new project one HAS to absolutely use the modern interface as a target but all we know that there is the 'little' possibility that no one will want (as now) software written for the modern interface (metro).

We have some successful LOB applications for Mac (that sell well) and we wanted to bring them to Windows but considering how confused is the situation we at the moment prefer to wait and see because this could be the worst time in the history to develop a new app for Windows considering users DO NOT WANT TO USE TILES AT ALL because they are just glorified widgets and nothing more... and people for LOB want true applications not these jokes that only a non technical person that spent the most of his time on Facebook can think can substitute true applications

My idea is that Microsoft has to understand that tiles will never work for LOB but I think they will never do, and that will cause a true disaster to Microsoft and the entire echosystem that will be remembered for the next 50 years

RPT
  • 131
  • 1
  • 3
1

It's really early to have a good answer to that question. WPF apps will not go away and I think it will go on. What specifically will change is open. Some things might be said in one of the talks and others will show later in the process of Windows 8 and later. Personally I think there is no immediate need to change your current strategy on LOB. If it is WPF use that, if it is WinForms or WebForms it is good as well. Maybe an approach to have a special tile that sums up some key data and calls the app will put additional value.

This is just an opinion

Sascha
  • 10,231
  • 4
  • 41
  • 65
0

I think the same recommendations apply as they do now. If you want deep OS integration than you're probably better off with WPF rather than Silverlight. However, Silverlight 5 does a lot to plug the current holes in Silverlight 4 if you go for an out-of-browser application.

I think it really depends on the kind of deployment you're wanting.

Chris Shepherd
  • 1,436
  • 1
  • 11
  • 17
  • Choosing **Silverlight** for the new project will be not a wise decision in the long running perspective, since it will go away (as a browser plugin) close to 2023 or so (when MS officially discontinue any support of it as they already stated). – Sevenate Jul 23 '13 at 07:38
  • Silverlight is dead. No one create new things is Silverlight anymore. – RPT Dec 13 '13 at 14:16
  • My answer is over two years old. Also, plenty of companies are still using WPF and Silverlight even now. – Chris Shepherd Dec 14 '13 at 13:27