What would you recommend for OS X development of a graphical application like those possible in WPF?
My specific background is in Smalltalk & Java, but I currently work mostly in DHTML/.NET (ASP.NET/C#).
Cocoa. Considered by many to be the best application framework ever. The language is Objective-C, SmallTalk-like language that inspired the creators of Java.
Really, there is no reasonable alternative to Cocoa for OS X development, unless you have specific needs like wanting to be cross-platform.
Aside from Interface Builder which is included as part of the Xcode tools, you can also use QT, GTK+, AWT & SWING (for your Java background), Tk, Squeak (for your Smalltalk background), Shoes (very cool little Ruby GUI toolkit), FXRuby (more Ruby), wxWidgets, XULRunner, and others I'm sure I've forgotten. For the most native-like apps, however, Interface Builder is your best bet.
Cocoa is the primary framework to use on Mac OS X. It's what Apple uses, it's what most new development uses, and it's where new features are principally added.
If you're coming from WPF, I think you might find quite a few of the concepts in Cocoa familiar. (Despite the fact that Cocoa is just a bit older.) It's built entirely around MVC, there are property-change notifications and bindings, there's animation support, there's a persistence and object-graph management framework, and so on.
(Also, you might want to add "mac" to the tags.)
With your Java background, don't get sidetracked by the now deprecated Cocoa-Java bridge. Early in OS X history, Apple provided a (laboriously hand-maintained) Java interface for the Cocoa libraries. Because of the semantic differences between Java and Objective-C, many of the most powerful features of Cocoa, including Key-value binding (upon which many other features are built) is very difficult, leading to divergence of Objetive-C and Java capabilities and the eventual deprecation of the bridge. All Cocoa development is best done with Objective-C or one of the many (automatically generated) bridges to dynamic languages such as Python or Ruby.
With your background in smalltalk, I would expect you could pick up Objective-C in a day or two.
I'm not sure what WPF is, but most development for the OSX platform is done in Objective-C with Cocoa. You can use the deprecated Carbon APIs with other languages like Java, but new applications for OSX really should be developed in Objective-C. You can start with Apple's guide with Xcode as your IDE.
To put it a different way than previous posters: if you are not designing your interface in InterfaceBuilder and manipulating it with Objective-C, then you are going to end up with an application that does not look, feel, act, or work the way a Macintosh application should, and it will stick out like a sore thumb to users. It will be an unpleasant experience for the user compared to other apps, and they will likely desire a different application because of it.
Toolkits like QT are acceptable if your application already uses QT and you want to port it fast, but if you're writing a new application (or a separate GUI) then write it in Cocoa using ObjC or ObjC++.
You might have a look at PyObjc which is a bridge between the Python programming language and Objective-C, including bindings for Mac OS X components, including Cocoa.
With a Smalltalk background, I'd recommend straight Cocoa and Objective-C. However, if you're leaning towards a dynamic language, RubyCocoa will let you use Ruby which I think you'll find easier to pick up than Python.
Related
I have a C# app that I've managed to get working with Mono and running on OS X. The application itself runs just fine, but it doesn't really look all that good when run on OS X. The button fonts look jagged, and many of the default features that are there for the Windows 7 version aren't present. To me it sort of looks like a Win98 application with an OS X top border taped to it.
I'm looking into possibly learning Objective-C so that I can write 'proper' OS X apps, but for the moment I'd like to be able to get my projects working on an Apple without having them look Frankensteined together.
Is it possible to convert a WinForms app over to Cocoa? Is Cocoa the correct interface to use?
If possible, what's the best way to go about it, and do any of you know of a good tutorial/writeup on the process to get me started? It'd be nice to see something that actually shows the process being done. I learn far more from example code along with a short explanation than I do from a generic article.
Thanks again!
If you want to create a native looking OSX application, you are correct that you want to use Cocoa. If you still want to use C#, you can use MonoMac.
MonoMac allows you to build your GUI natively on OSX (in fact, using the Interface Builder shipped by Apple), but allows you to write your app in .Net/C#. This way you can continue to use your existing application logic and only have to change the GUI code.
It's pretty much universally true that a ported app will always look like a ported app. Even large companies with huge budgets can't pull off anything better (I cite Adobe - what a mess).
Slapping a Mac face on a Windows app port will show its seams. If you want the application to behave like a native application (and take advantage of the performance-related goodies the platform offers), the absolute best approach is to use the architectural documentation, specs, and requirements I know you have - conscientious developer that you are - to adapt the design to the native platform. That's the Cocoa Frameworks (the API), which are written in Objective-C (the language).
There's quite simply no other path that doesn't end with a crappy-looking port that's riddled with bugs and behavioral problems born of the porter's insufficient familiarity with the target platform. This isn't just limited to Mono/C#-to-Cocoa/Obj-C. The opposite is just as true. Even Java-for-PlatformA-to-Java-for-PlatformB ports of desktop apps suffer these problems. Start with a solid architecture and build for the platform if you want the best user experience.
That said, you're already a step ahead by realizing this and wanting to do better. Bravo! :-) Though I don't know of any tutorials for this path, I'd suggest even that's not the correct approach since you indicated you're looking for quality. Avail yourself of the many Cocoa books and many more online communities (like this one) and learn the platform before committing to your Cocoa-adapted architecture and code base.
Update based on comment debate
To be clear: I'm not saying there's no way or that there's no tool out there that makes it possible. There're actually plenty I've seen but don't recall and won't bother googling. My point remains: the OP is concerned with quality of native appearance (and I assume behavior and possibly performance) and porting tools / translation layers don't achieve this due to inherent differences in the platform's architecture and user experience idioms. The OP suspects it might be best to learn the platform and build specifically for it and I'm agreeing. Your opinion may vary. Have at it.
I am used to programming in Java EE but right now I am being assigned to Java ME project. From what I am reading all those ME restrictions are quite painful. Do you have any tips on how to deal with such a crude language? Are there are any good libraries which might give a higher level of abstraction?
I am also looking for tips about developing tools or strategies e.g. tools like maven (I mean tools without which you can't imagine working). Are there are tools/approaches which you must know about in Java ME?
There are so many things to discuss here, you could write a book or two about the topic!
First of all, realize that "Java ME" is an umbrella term that encompasses a variety of Java environments, some very limited compared to Java SE, some nearly as capable. However, when most people use the term they're really referring to the limited side of things -- the CLDC (Connected Limited Device Configuration) and the MIDP (Mobile Information Device Profile). So let's assume that's where you're going.
The biggest problem you're going to face as a J2EE developer is the sheer LACK of APIs in a MIDP environment. Also missing language features -- no enums, no generics. Basically you're dealing with a subset of Java 1.3, which is a huge step back for you. The collections classes aren't there, for example, so you end up using the old-style Vector and Hashtable classes a lot. Or you end up using device-specific API extensions at the cost of portability.
The good news is that the NetBeans tools is pretty good for MIDP development, they've put in a lot of good features. You can also use Eclipse, there are some good plugins for that. The tool support is there, including ant tasks and so on.
There's lot of material on the web to get started with Java ME. Read some of my Java ME overview material (which still refers to Java ME as J2ME). Then figure out what platforms you're targeting. If you're doing BlackBerry stuff, for example, you'll want to learn and use the BlackBerry APIs and not MIDP.
There's definitely a learning curve, but you'll adapt quickly enough.
Let's suppose you're an IT student with a basic knowledge of C++ and C#. Let's suppose that you want to design apps that:
need to deliver some performance like archivers, cryptographic algorithms, codecs
make use of some system calls
have a gui
and you want to learn an Api that will enable you to write apps like those described earlier and:
is mainstream
is future proof
entitles you to find a decent job
is easy enough - I mean easy like VCL, not easy like winapi
So, making these assumptions, what Api will you choose? MFC, WPF, other? I really like VCL and QT, but they're not mainstream and I think few employers will want you to write apps in QT or Visual C++ Builder...
Thanks for answers.
Note: The following answer was written several years ago with Desktop application development in mind. Today (in 2018), you'd probably just build a web application to end up with something reasonably cross-platform and device-independent. (For example, using ASP.NET Core on the server side, coupled with a UI framework/library such as React, Vue.js, or Angular on the client side).
Win32 API -- I'd forget about it, if I were you. Programming a Windows application directly via the Win32 API only makes sense if you're programming in pure C, or if you really need to do a lot of system calls, or if you're concerned about the additional overhead introduced by more "comfortable" platforms or frameworks (such as the ones named below). Programming UIs directly through the Win32 API is tiresome, messy, and you need to deal with lots of details. It's also not platform-independent at all, but you may or may not be concerned about that.
MFC -- Perhaps an option if you're programming in C++ and fixed on the Windows platform. I never understood what's so great about it, other than it makes the Win32 API much more comfortable (AFAIK it's basically a collection of object-oriented wrappers around the Win32 API that take away some of it's complexity / messyness). Also, it's also not very platform-independent.
Qt, wxWidgets -- Fairly widespread UI frameworks. Might be good options where platform independence plays a role. AFAIK both frameworks are targeted at the C++ language.
WinForms (.NET) -- Similar to MFC, this is also based on the Win32 API (USER32 and GDI+). AFAIK the WinForms framework is now being ported to Mono, and therefore somewhat cross-platform. However, it's not exactly the most up-to-date technology. For complex UIs it can also be somewhat sluggish sometimes. If I had to decide today which framework to use, I'd rather choose...:
WPF (.NET) -- More modern than WinForms, with more graphical capabilities and, apparently, faster rendering, as it is no longer based on the Win32 API (GDI). (And it runs on .NET, which I find a great platform to develop for. Programming in C# is so much easier than programming in C++ IMHO, which is also an argument against Win32 API, MFC, Qt, and wxWidgets.) Note that WPF is not cross-platform, it exists only on the Windows platform so far.
Then of course there's Java, including the UI frameworks that come with it. I can't say much about that since I'm not a Java person, but I could imagine that Java would be the best choice for platform independence; and it's the dominant platform (over .NET) in certain industries (e.g. mobile phones, banking, due to the very solid JVM and security considerations).
So my recommendation would be the .NET framework, and WPF for the UI, if you're planning to stay mostly in the Microsoft world. Remember that you can still use the Win32 API (you won't come closer to "system calls" than that) via P/Invoke.
If you enjoy coding in C# and working with the .Net framework I'd recommend that you have a look at WPF. WPF is a great GUI framework where you can do just about anything - and also make it shine! WinForms might be easier to get a grasp on, but I'd say WPF is more "future proof". Another positive thing is that WPF is really similar to Silverlight, so if you handle WPF well you should be able to write Silverlight applications too - if that's of interest. Please don't bother to learn MFC.. I can't believe there are many that is using MFC today for other reasons than that they were using it before, and didn't get an opportunity to change..
There are a lot of good jobs out there for .Net programmers, so being able to handle some GUI framework in addition to C# and general knowledge about the .Net framework will be worthwhile.
When it comes to your points about being able to "deliver some performance like archivers, cryptographic algorithms, codecs" this really shouldn't depend on your choice of GUI framework. This kind of code will be writen in layers outside of the GUI layer, and will typically be bound to the GUI. With WPF you'd write your e.g. cryptographic algorithms in C# in some class independent of the GUI layer, and then the View written in WPF would bind to C# code and get its answer from here. However, if you used WinForms you'd still do the same thing, and the performance relies on the algorithms - not the GUI.
When it comes to getting started with WPF there are a lot of questions on SO helping with this. So you should find good help with a quick search.
Good luck!
I'm looking for some arguments to pitch to my boss and fellow developers.
We're currently finishing up the preliminary UI mockups and getting ready to move on to the next phases of development. In the meantime, I've been digging through the depths of the Carbon, Win32, and wxWidgets APIs attempting to make some of the controls have a more native look and feel on the Mac and Windows platforms.
The more I dig into the Win32 and Carbon APIs to implement the things we want in our project's UI, the more antiquated they feel, and the more I'm beginning to think that we should be implementing the project as described in the last paragraph here.
We're using wxWidgets for our current projects. wxWidgets is coming along on the wxCocoa port, but it doesn't look like it's going to be ready for prime-time before we start major development efforts on our new application. On the Windows side of things, it wraps the Win32 API rather than WinForms or WPF (likely due to native vs. managed code).
We're already designing the system with the MVC pattern in mind, thus aside from having to write two native UIs, it should be very doable, and, IMHO, easier to get the desired UI effects using modern APIs such as Cocoa and WPF.
I've been trying to push these points subtly, but the start of major development is coming soon. Does anyone have any suggestions on how to pitch using native UI toolkits in our next application vs sticking with wxWidgets?
Thanks in advance.
Create your core code in Standard C++ and use Objective-C++ with Cocoa to create your user experience on the Mac and C++/CLI plus C# with WPF to create your user experience on Windows. Follow the platform guidelines for the Mac in your Mac version, for Windows in your Windows version, and don't even bother thinking about trying to share user interface code.
One good way to manage this is, instead of just Model-View-Controller, following a Model-Model Controller-View Controller-View architecture. Your Model Controllers are platform-independent and manage the higher-level functionality of your application. (For example, its entire concept of documents, file format, job queues, and so on.) Your View Controllers are platform-dependent and mediate between your Model Controllers and your user experience.
Of course you'll probably also want some platform-dependent code at the model level too; for example to use NSOperation on the Mac and thread pools on Windows to implement job queues. Just create your own lightweight abstractions for that sort of thing.
Actually I think using Qt has become very interesting since it's now LGPL
Every time you add a layer of abstraction, you trade control over the details for more rapid development. You'll be able to get a lot done up-front using some cross-platform framework. On the flip side, when you want to do something that the framework doesn't support—and lets face it: it isn't going to implement all possible things those native APIs can do—you either have to implement it (for all platforms) using the native API, or do some other wierd hackery to get a "good enough" solution. And of course, when things go terribly wrong, having that extra layer of code you don't own makes it harder to debug. There really is something to be said for owning your entire stack as much as possible.
Writing two font ends is a lot of work, maintaining two front ends is a huge amount of work, if you need your program to run on multiple platforms go with a multi-platform toolkit.
If you write platform specific front-ends, each using the state of the art tools for that particular platform, you will get a much better user experience - but the cost of developing and maintaining those will be on the same order of magnitude as developing the entire application from scratch for each platform (yes, even with MVC).
Personally, I'd rather stick multiplatform and don't give a damn for that eyecandy, but if I wanted to pitch the use of those native APIs, I'd work out the (end-user-visible) differences between how things are done in different GUIs. If you can convince them that, in order to feel native, the program's user interface has to look and feel very differently on Windows and OSX (because of different design guides/philosophies/whatever), they should understand that, even with wx, you would still have to implement it twice, to accommodate those different requirements, so you might just as well use the real thing, i.e. the native API.
Also see this thread on the Google Chrome mailing list discussing the same choice of UI for Chrome on different platforms.
Win32 is definitely waay to old, but you might want to look into something like Microsoft Foundation Classes which is designed to do native development with C++. I assume that a similar thing exist for MAC.
Personally if I was in your situation I would properly also go for QT or WX.
Does anyone have any suggestions on how to pitch using native UI toolkits in our next application vs sticking with wxWidgets?
No one likes a corridor-wiseass.
I think action speak louder than words...make a small prototype of how you think it could be done, and show it. Maybe you have to do this in your spare time.
Cocoa is really great so I think that with little code you can show an idea...well, this require that you know Cocoa enough.
I am in the beginning stages of creating software for a mISV-to-be. The program is a desktop application and in the long run I want to have a native version for both Windows and OS X (I have a looked at various cross-platform APIs, and none of them meet my needs). Initially though, I don't think it makes sense to develop for two platforms at once. With that in mind, I have been looking at WPF for Windows and Cocoa for OS X, and they seem similar.
Has anyone had experience porting one to the other? Are there particular techniques/paradigms to follow that will make porting easier? Ignoring business considerations, would you recommend developing on one of them first?
Well. Once you've written an app for Cocoa, it is possible to port it to Windows. This could be done using gnustep or Cocotron.
If you do it the other way, WINE is meant to make porting easier.
I would rather write the OSX version first. This is because Windows users have no clear idea what they want an application to look like. In my experience, they are quite able to suffer through all kinds of user interfaces. Consistency has little value to them. Since there is no common agreement, what a Windows app should look like, nothing stops Windows users from actually liking OSX designs, and they even frequently do. iTunes for windows looks like a very typical OSX app and you hear very few complaints that it would not be enough Windows-ish.
Going the other direction, this is not true. OSX users have a clear preference for Cocoa apps and very little tolerance for, as an example, things like GIMP or Inkscape which work under OSX just as well as anywhere else, but look plainly ugly to the OSX trained eye.
I think that you're on the right path by choosing windowing environments that are specific to each platform. This approach allows you to create a user experience on each platform that isn't restricted by the compromises inherent to cross-platform windowing toolkits.
A good first step is to break your design down into two parts: platform specific and platform neutral elements. You can already put any UI code into the platform specific column, but maybe your app will need some data persistence that can be written in platform-neutral C++. What you may find with this approach is that there is quite a bit of logic and infrastructure that you can write in a platform-neutral way, leaving just the UI and glue code as platform-specific.
There was a recent episode of Late Night Cocoa titled Porting Large Applications to the Mac platform. Your app may not qualify as "large" but this podcast gives quite a bit of great porting knowledge from someone that's done it a few times.
I'm currently working on porting tools in this space and have many years of oft-painful experience in either using or writing cross-platform frameworks on Mac and Windows.
One of the biggest problems in the past has been Apple's refusal to open up the nib format for Cocoa (Carbon nibs were open XML files years ago). That changed with XCode 3 and the .xib format, as well explained by Frasier Speirs.
At the basic layout level, at least, there is now an opportunity to automate porting from one XML format to another. I regard WPF (XAML) as cleaner and so I'm using that as my base format and migrating to Cocoa.
When it comes to the code behind, whilst you can use C# under Mono, the CocoaSharp project seems either stalled or very slow and I wouldn't recommend it.
If you are comfortable with C++, consider having as much logic as possible in C++ with a thin platform-specific layer in C# and Objective-C.
Another approach worth investigating is using a dynamic language like Python or Ruby. I'm not sure which is more mature at present between IronPython and IronRuby but both are now supported by Microsoft people. On the Cocoa side, I think the flexibility of Ruby syntax will triumph and RubyCocoa is probably overtaking PyObjC.
Otherwise, work in C# and Objective-C and maintain two completely independent code bases with identical designs. Fortunately the frameworks have comparable semantics for most things, especially if you make use of bindings.
Well, there is not a straightforward path. The best method is to use something like Model-View-Controller pattern or some other architecture to separate business logic and so forth from the presentation. However, unless you are using Mono, there will be very little code for you to share, I think.If you are developing WPF then you surely doing .NET and, other than Mono, Objective-C is the standard programming tool under Mac OS X.
Keep a good design and you can have most of your code simply be an Objective-C version of your .NET code and vice versa rather than trying to find a migration path.