Porting existing c# to java (lots of pinvoke/user32.dll) - Pitfalls? - c#

As states we are in the process of considering porting an existing app which uses the user32.dll quite extensively. Mainly for stuff like taking screenshots of the screen and moving the mouse around.
Using something like JNA we should be able to do the same thing, but are there limitations? Things we should be aware of
Thank you

As a first point I'd have to ask why? If you're using JNA you're losing all the cross platform nature of Java anyway, and losing the tight integration with Windows that C# gives you. There's no limitations that'd stop you doing it, but if you are just tying yourself to using native code anyway then I'd say there's not much point.
If on the other hand you're looking to gain cross platform capability and you're looking at porting, then you can take screenshots and move the mouse around using the Java Robot class easily, no native code required. If you are rewriting in Java, that's by far the approach I'd recommend.

Related

C#/GDI Converting to Universal Apps for Win10

I realize someone asked a similar question here but I think my question is from a different vantage point and I hope it might help people in the future understand their options.
Basic issue: I have written a C# App (obviously managed) as a normal application (non-universal) and in order to support a quick log window above an active typing area, I initially wrote some graphics using pure .net (GDI I believe). This worked initially but it became VERY slow sometimes and I want way better control over how it looks including some specialized graphics.
What I would like to do is rework the app AND find out which of the following makes the most sense for a Windows 10 Universal Application:
Call DirectX Directly from C# (from documentation this is either not-possible or not really ideal, base project for C# does not have directX like C++ does).
Call Managed Libraries from C++ and use DirectX as the Application Foundation.
Use C++/CLI and call managed code. (I believe this is not possible anymore).
Use SharpDX (I looked and it didn't seem available based on a question here).
Use another Third Party (such as Unity) but deal with game frameworks.
Other parameters: Since everything is C# already, would like to avoid total rework, however, if it's recommended I'm 100% ok with doing so. This is not a game application, though it needs performance like one on one of it's screens, it will mostly be driven by interfaces (could be XAML) and network data coming in on a separate thread. I also have no qualms about learning extensively (say a very different language from C# or C++) in order to complete this project because learning is value. If I am missing something glaringly obvious that would be great to know about it as well.

Java Swing vs C#.net desktop database application

So I'm planning on making a little desktop application that fetches TV-series information from an API and puts them in a database. I will then use this info to keep track of which episodes I watched etc etc (A bit like a desktop version of http://www.myepisodes.com)
Now I've done some research around stackoverflow but I'm still not sure which language will be the best.
Java swing using JavaDB or C#.net using MSSQL? (or even another alternative). What are the advantages / disadvantages of the two?
All else being equal, choose the language that maximizes your programming joy.
If you're targeting Windows, use C#. If you're targeting a cross platform product, Java's not the worst choice.
If you want easier cross platform compatibility go with Java, otherwise it seems like a personal preference. LINQ is very useful for queries in C# so you might want to check that out too.
If you find you really want to use C# but also want cross platform look at developing it using Mono: http://mono-project.com
I personally find Swing to be a total pain to work with, but XAML/C# is quite nice. C#'s native events make it much easier to program an event-based UI.
LINQ is also amazing for any sort of data queries you might need.

A project I'm assisting with is using freeglut for the client. Should I advise otherwise?

I'm involved in a development project that is using freeglut (based on the long defunct glut) for it's client.
The client will eventually allow full interaction with a large-scale 3d environment. Should I let the development continue with freeglut (is it even possible) or should I advise they use another alternative such as libsdl, opentk or even axiom.
I'm not a graphics person but I get the feeling freeglut might potentially be a limited choice.
The most convincing answer (for or against) will be accredited.
EDIT: A few points to make...
The project is already using the Tao Framework.
DirectX and XNA are not options (ie: something like freeglut or libsdl has to be used).
I did do my research and found that freeglut was once again under active development and that they have a release pending. That doesn't change my feeling that it may still be a potentially limited choice.
My question isn't on how it's done but on whether or not freeglut is still a viable choice for something that could potentially get big and whether or not there are more "modern" solutions that might ease development a bit.
EDIT: It would help if other alternatives have better mutli-threading support (not for rendering objects but for processing data and so forth).
EDIT: To elaborate a bit more... the client must work in at least linux and windows.
Thanks!
I don't think there's anything wrong with freeglut. It is being actively developed, and there is an active board for support on Nabble.
That said, it is not difficult to create an OpenGL context in Windows, so why not just use it directly? I always thought people used Glut/FreeGlut because that was what the Red Book used. (though the callbacks for mouse and keyboard do simplify things) I never timed it, but Glut always felt a little slower than using OpenGL through interop.
The Tao project has a good OpenGL wrapper if you want to use OpenGL directly. It also has nice .Net bindings for FreeGlut and GLFW, a Glut alternative which offers mouse and keyboard callbacks too.
I'd go for libsdl, its multimedia capabilities make it easier to work with audio hardware as well, it has native bindings to C# and a variety of other languages, and can also be of use if you ever decide to integrate a mobile interface for your project.
OpenGL with C# is done via wrappers. Most of them are not actively developed for anymore.
The canonical choice for 3d development with c# is either using managed directX, or using the XNA libraries.
R

Implementing a fully-functional prototype for a web-based application

I am building a prototype for a web-based application and was considering building the front-end in HTML, which can then be reused later for the actual application. I had done a Flash-based prototype earlier, which embedded the .swf into a C# executable. Flash made for rapid turnaround time while the Windows application provided unlimited access to fancy API's for DB access and sound.
I want to consider something similar for this one too. Does this approach make sense? I am particularly concerned about the way the HTML would communicate with the container app. From what I understand out of preliminary research, it would be only through JavaScript, which might quickly get unwieldy. This is especially so because unlike the Flash-based prototype which implemented a lot of its functionality in the .swf, the HTML UI will depend entirely upon the shell to maintain state. Also, I don't need anything more than access to a database. So a desktop application might be overkill.
Another alternative that comes to mind is to build the prototype using PHP and deploy it with a portable server stack such as Server2Go or XAMPP. But I've never done something like this before. Anybody here shed some light on drawbacks of this approach?
The key requirement is rapid iterations of the UI, reusable front-end code and simplified deployment without any installations or configuration.
Some of the best programming advice I've seen came from Code Complete, and was along the lines of, "evolutionary prototypes are fine things, and throwaway prototypes are fine things, but you run into trouble when you try to make one from the other." That is, know which type of prototype you're developing, and respect it. If you're developing a throwaway prototype, don't permit yourself to use any of it, however tempting it may be, in the production system. And if you're developing an evolutionary prototype - one intended to become the production system - don't compromise quality in any way.
It sounds like you're trying to get both, the rapid development of a throwaway and the reusability of an evolutionary prototype - and you can't. Make up your mind, and stand by it. You can't have your cake and eat it, too.
I think you off to the wrong start, here. Why would you want your prototype to be fully functional? A prototype is intended to be throw-away and to help flesh out requirements and UI. If you need full functionality, why not just skip to the final product? If prototyping is really something you want to do, I suggest looking into a specialized prototyping tool.
Are you prototyping the user interface for a customer? If you are, consider something less unwieldy like paper prototypes or presentation software (like PowerPoint) until you get the UI nailed down. If you can establish the UI and are clear about the customer's requirements, you can then develop the application in whatever the actual platform is going to be with a clear model in mind.
In my current project, I prototyped the UI in PowerPoint first. In a subsequent iteration, I used static web pages and some jQuery plugins to simulate actual user interaction. That proved to be very effective in demonstrating the interface, and I didn't have to build the application first.
I would join in on folks suggesting paper prototyping as the "idea", but not necessarily the implementation. The biggest point here is that tools such as HTML or Flash let you get "bogged down" in the details - what does this color look like? What's the text on this thing? Lots of time can pass by that way. Instead, what you should be focusing on is user flows.
One tool that keeps the spirit of paper prototyping without all the "paper" drawbacks is Balsamiq: http://www.balsamiq.com/demos/mockups/Mockups.html. It was covered by Jeff and Joel in one of the Stack Overflow podcasts; I've been using it for my own projects for a while. It's freeware, and it does its job magnificently.
If you know C# then another option you can look at is Silverlight. You can then leverage your knowledge of C# and/or JavaScript and interact with a rich object model.
Would that do what you are looking for? The installation would be minimal on the part of the client - download and install the Silverlight plugin
If prototyping is something you truly wish to accomplish here, paper and pencil will be your best friends. You can draw out as many iterations as necessary. While none of this is ultimately useful later on once you begin coding, it is as quick and rapid is it goes.
As mentioned previously, there are many prototyping tools which have a bit of a learning curve, but an alternative to consider would be using a framework such as CakePHP or Ruby on Rails which make for fast application logic and leave customizing the front end being the primary hard work left. And plus, you're left with a mostly functional application when you're done with your prototyping which can be tweaked as needed.
In either scenario, you're paying with your time either upfront (in the case with learning a new framework), of over time in payments (with the case of prototyping on paper or coding by hand).

What is a good desktop programming language to learn for a web developer? [closed]

As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, visit the help center for guidance.
Closed 10 years ago.
I'm want to learn a desktop programming language, preferably C, C++ or C#. I'm a PHP/HTML/CSS programmer and I would like to get into desktop applications. I need something pretty powerful and I would like to be able to create applications with Windows GUI's.
What would the Stack Overflow community recommend? Is there any knowledge I should have before diving into these languages?
edit:
A web programmer wants to create Windows applications and you recommend C? What's wrong with you people?!
/edit
Obviously C#.
C# will be easier to get into and will let you build Windows applications using WinForms or WPF and all the new Microsoft toys in .NET. If you know your way around PHP, you should already be familiar with the syntax, object oriented concepts, exception handling, etc.
I suggest you don't complicate your life with C and definitely not with C++ if all you want is to create Windows GUIs. They do provide a good educational experience and they are useful for more advanced things (cross platform development using other toolkits for instance) but at the price of a steeper learning curve and reduced productivity.
Also, if you are a web developer, C# is the only language among the 3 options that you can (realistically, heh) use for the web. ASP.NET is not a bad framework and might be worth investigating too.
"I need something pretty powerful and I would like to be able to create apps with Windows GUI's."
For rich windows GUIs, I'd be tempted to look at the "xaml" variants: WPF (full desktop) and Silverlight (like flash) - this will give you the latest tools etc, and a lot of styling power.
For this, C# would be the primary choice. Managed C++ is an option, but it's more work.
Take a brief look at C first, to the level of writing a few tutorial programs. C is small, well defined and the definite language definition (Kernighan & Ritchie - The C Programming Language) is only a couple of hundred pages long. Although getting your head around pointers and recursion can be a challenge the language itself is sufficiently small that it's possible to know the language completely. One of the pleasures of writing C is it's not too hard to reach the level where one can just code without having to continually refer to documentation to see how to use little know feature x or library function y. Think of learning C as putting the foundations in.
However to actually write useful applications you'll want to use C++ or C#. Which depends on what you wish to write. C# will probably score in most business-orientated cases because of garbage collection and general ease of use, but C++ would score for high performance, low level, or game domains. There's also vast amounts of C++ library code available which can be leveraged relatively easily (and pure C code too which is easily adapted). If your aim is to write something with a particular problem in mind then the availability of such could most likely offset the higher overhead C++ has in just getting things done.
By way of example on this I did a filtered search on 'image' on Sourceforge and returned the following project count
C# : 139
C++: 569
And even with something purposely more business-orientated C++ still scores higher - 'finance'
C# : 12
C++: 32
With anything mathematical or technical the score is heavily weighted against C# - 'fractal'
C# : 4
C++: 44
C : 24 <-- (C++ explicitly excluded)
This is a pretty crude analysis of course, but for example if my aim was to write a desktop fractal generator program then I'd pick C++ over C# as the extra effort involved in handling the UI in C++ would be easily offset by the plethora of example code available in C++. OTOH if I was writing a business application where there was no great advantage to having example code and lots of UI to handle then I'd go for C#.
Afterthought edit: Another reason for at least familiarising yourself with C first is that there's also Objective-C as the main development environment on the Mac and iPhone. Objective-C is C with a very thin wrapper and only really a first cousin to C++, so if you envisage that you might ever be in the position where you want to port code to or develop directly on the Mac then I think you'd find it less confusing if you at least have some feel for how C++ (or C#) builds on C rather than having to 'step sideways' across the tree.
C# vs the others really solve quite different things but given your web background and it's rapid development approach, I'd definitely go with C#
My suggestion is to learn C++. C# is good, but it will prevent you from writing portable code. If you learn C++, you got many choices. In windows you can use Windows Forms application or WPF with .NET framework. MFC is another good choice. QT is the other choice which works on multiple platforms. So by learning one language, you can write code for multiple platforms and you got a variety of GUI tookits. After learning C++, learning C# would be trivial.
All the best
C# is the highest-level language of these, which means that it's the easiest one to accomplish what you want.
Windows GUIs can be surely made with C++ and even with C, but it's harder.
definitely first go to the basics: learn C.
then go to your real goal. if you want any chance of going multiplatform C# is discarded. you'd want to go with C++ and Qt, or wxWindows.
In my opinion you should starting studiying C, to have a base knowledge about aspects like memory, file accessing. C is a low level languaje, this means that you will have more control over the way you access to operating system, but to create windows application you'll have to code a lot of lines of code. For this reason, after doing few things with C, i should start with C# with Visual Studio 2008. .Net Framework comes with a huge number of classes that makes your life easier.
Hope it helps.
Kind Regards.
Josema.
C/C++ both allow low level programming. That means that everything is possible, but even simple things can be hard. If that's what you mean when you say that you want something powerful, you should probably start with C++. It at least gives you some decent GUI libraries (gtk+ with gtkmm or qt), whereas GUI programming with C is a pain.
If, on the other hand, you want to develop applications quickly (and that's what you mean by powerful), then C# is probably the way to go.
I woudl recomend C#. It's not as powerfull as the other two, but for your needs I belive C# will be least painfull. C and C++ are low-level languages, so I think they will be a lot of work for GUI and stuff.
Save the C. To those people that advise learning C, please give arguments for that. I claim that it has absolutely no benefit to learn C rather than C++. I, unlike them, have arguments to back this claim up.
In a nutshell: C is already a subset of C++ for all intents and purposes, but it's far from the most interesting subset. It only offers a very limited insight into new proramming paradigsm. C++, on the other hand, defines a few very interesting new paradigms which, furthermore, are a generalization of all that C has to offer (i.e. iterators which are a generalization of pointers).
To reiterate: if the choice is between C and C++, choose C++.
Although I agree that C is "useful" to know it is certainly no longer essential for the majority of programmers. For me these days it's main function is that it has given me an understanding of how the underlying hardware affects program performance and behavior but that's the end of it. I haven't used it in development work for the last 9 years.
In regards to getting to grips with object-oriented development I feel that both C# and C++ will give you a firm foundation but C++ is far more flexible (and harder to use correctly) while C# is more consistent and way to go if you're going to get into .NET development. I would therefore recommend learning C# and one of the "new" .NET-presentation technologies like Windows Presentation Foundation or Silverlight. Alternatively, learn Windows Forms first but it's a bit lacking in presentation.
The only case where I would recommend C is if you are going into fairly low-level stuff like device drivers, embedded controllers or similar but in most of those cases you could still use C++.
Having come up through the languages, first C on Unix, then C++ with MFC and then C# (starting with 1.0) I think that going back in time to earlier incarnations of the language is a bit like learning Latin. Oh, probably just stolen something from Paul Graham there but I would agree with it.
C is very good if you have to have a minimal object code size, very good for embedded stuff where it's one step up on the assembly language. I'm thinking of things like a Microchip PIC etc where you may only have 1K word of program space.
C++ is very good if you can have the larger object size but still need to be able to hammer the metal directly - so things like device drivers, network stacks, etc etc. The kind of glue program that sits under everything.
An OS could end up in either of these - or more likely a mix of the two depending upon the programmer, the age of the code, what it needs to interface with, the depth of the snow outside (just one more run and then I'll write that method!)
C# is really good at the higher level business focussed user layer applications. They could be web based, client based, a mix of the two - it doesn't really matter. The nice thing about the .net family is that you have a rich library, for this layer that is important, and it is being heavily developed to be richer all the time.
As computers become more powerful the cost of the developer outweighs the cost of the processor or memory.
Go with C#, ignore C and C++. By the time you are happy with C# you won't have any need to program in a lower level language for the speed increase.
Again, another vote for C# and against C. Sure C is interesting from a technical point of view and will teach you all the details, but the truth is, you don't need to know all the details to make an application. If your goal is to make something work, which it sounds like it is, it's C# hands down. C# is the highest level of the three, which means less time fighting the language, more time making stuff work.
It sounds to me like most people are recommending you learn C so that you can eventually make better C# programs. Why wait, go straight for C# and learn it. I do think C++ is an absolutely excellent language and one day you should learn it, but there's no reason why you can't make a quality C# language if you've never learnt C or C++ - many people do that every day!
It doesn't really matter what kind of language you start with but if I were you I would do:
Start with the basics in C (It will probably be hard to understand)
Move to C++
And at last C#
I'd suggest you start with C#, but quickly progress to C++ as once you have learnt how to code generally, you should return to your roots. I'm currently studying electronics because I'm going back even further. The more you know about how and why, the better you can know about how best to get where you need to go. I think a C# programmer that doesn't know any lower level language is going to write rather sub-optimal code, but a low level programmer learning C# is going to be overly cautious of all the managed stuff. So, whatever you choose, you should consider doing both in the long run.
I would go for C#.
I've done my time (lots of it) with the other 2 and find C# 'cleaner'. You also get a rich library of all kinds of functions included in .Net. I assume that you're not interested in developing for other non-Windows platforms which might have been an argument for going for C or C++. I also assume that you don't need the squeeze the last drop of performance out of the system where C might have given you a slight advantage.
To those who say you ought to learn C in order to understand how programming really works I would say that you ought to learn assembler in order to understand how C works. I did and it does help but is it really worth the pain? To be a programmer you have to switch on your brain. Think about what you are doing and understand your tools!
My vote: C#
Learn C#.
Is good to begin with the basics, but better than begin with C I would begin with the basics of C#.
This will give you a solid foundation with the fine points of C# basics and after you could move to more advanced topics.
For example you can take a look to Charles Petzold ".NET Book Zero" here.
I recommend::
First get into C. spend some time get use to it(but not too much)
Then get into C#. that's the way you can deal with GUI things more easyily.
IDE like VisualStudio can help you a lot dealing with GUI.
If you wanna be a real good programmer of C/C++/C#, then it might be not the good way but I think you are not interested in being a Guru or something like that :)
There are a lot more jobs developing windows apps in C# than C or C++. Learn C#.
C# is a good place to start. You can get it for free and you can do some fun graphics using XNA studio (also free).
I would only learn C++ if you have a specific reason to.
If you are doing Windows programming I would recommend that you do the first few chapters of one of Petzold's Programming Windows books. It will give you a basic understanding of C and of how Windows works at the bare metal level. You don't really need to learn C. Charles Petzold holds your hand and gets you up and running in a basic C Windows program.
Learn C#...if you want to develop desktop apps.
With proper set of libraries memory in standard C++ can become as easy as managed (smart pointers). Some C++ widget libraries are as easy to use as C#.NET. However it will probably perform worse than C# due to several abstraction layers. And with C++ you need to learn non-standard libraries. With .NET most of the components that you need are standard.
C is ancient and does not have STL. It demands you to be smart, but I doubt the advantages of it.
I need something pretty powerful and I would like to be able to create apps with Windows GUI's.
Well C is pretty powerful so that would would handle that criteria but something like C# and WPF would make life so ,uch easier if you're aiming to GUI apps.
Without C/C++/C# experience start with C. Edit: Ok, start directly with C# or C++ don't lose time. You cannot seriously code Windows desktop applications with C anymore, it's just to painful.
Then move to Object oriented languages to get something more powerfull.
If you wan't to stay far from windows plateform, C# is not the better option, or check out the Mono plateform.
IMO, C# is far more advanced than C and C++, you will build big application faster. WPF librarues will give you nice UI without windows style (but only on windows systems...)
Are you planning to make gui based applications for windows? Depending on what, I recommend VB. However, given your choices, I would pick C#.

Categories