About Scientific Programming

[truelies]

I am a little curious about the software selection on Intel Fortran and Micorsoft Visual Studio. For making a Scientific Caculation Software with windows interface, which way is the best?

1. Write scientific caculation parts with Intel Fortran, interface with Visual Studio, then connect them together. I think Intel Fortran can't be used to write windows interface, is that right?

2. Write whole software with Visual Studio.

Thanks!

[bmchenry]

The answer to your question requires consideration of what and where you are now in the development process for each particular project and where you want to go.
You can get to wherever you need to go by so many paths it is nearly impossible to recommend the best path.
For example, do you already have scientific code? What language is it in? And you merely want to create a better interface for input and output presentation? Or are you starting from scratch?
Having worked with Fortran since the early 1970s I have a bias to continuing development with Fortran as the core of scientific programming. I then pick and choose how best to accomplish the interface tasks for the particular problem solution.
Sometimes I develop it all in Fortran. And INTEL Fortran can do it all.
Sometimes I use a 3rd party interface like Winteracter, etc (to streamline and simplify the calls to operating system and graphics interfaces, etc.)
Sometimes I feel the program/problem requires that I develop the interface with Visual Studio (C++,C# or Visual Basic) and link to Fortran.
INTEL Fortran has done a great job of keeping up with the modern advances in programming while maintaining its core purpose of FORmula TRANlation.
Your choice of how best to solve your particular scientific problems needs to consider many factors such as these I have just mentioned.

[tropfen]

Hello,

i am using the combination of VS2005 standard and IVF 10.1 for scientific programming. In combination with XEffort (http://www.xeffort.com/) i have build an interactive diagnostic and presentation tool for climate scenario simulations datasets. To build the windows interface i have used the included VS resource part.

In my point of view it is a very good combination (VS + IVF + XEffort) for programming scientific windows software.

Frank

[Paul_Curtis]

Quoting - truelies

... I think Intel Fortran can't be used to write windows interface, is that right?

No, that's wrong. There is no aspect of Windows programming, particularly GUI features, which cannot be easily and directly accessed with Fortran. IVF includes modules with Win32 constant defines as well as interfaces to Win32 API routines. Writing full-featured Win32 programs can be done with IVF as well as with any other language, and avoids mixed-language programming (which must not be easy since I estimate over half the posts in this user forum involve multi-language interface problems, rather than questions of pure-fortran programming technique) and also avoids dependence on 3rd-party toolkits or libraries which are likely to fail to provide the complete low-level controls you will need. This does require an investment of effort to learn the Windows message-pump architecture, which is very different from classical procedural code. There are many excellent programming books (ie, Petzold) to help with the Windows part.

[truelies]

Quoting - Paul Curtis

No, that's wrong. There is no aspect of Windows programming, particularly GUI features, which cannot be easily and directly accessed with Fortran. IVF includes modules with Win32 constant defines as well as interfaces to Win32 API routines. Writing full-featured Win32 programs can be done with IVF as well as with any other language, and avoids mixed-language programming (which must not be easy since I estimate over half the posts in this user forum involve multi-language interface problems, rather than questions of pure-fortran programming technique) and also avoids dependence on 3rd-party toolkits or libraries which are likely to fail to provide the complete low-level controls you will need. This does require an investment of effort to learn the Windows message-pump architecture, which is very different from classical procedural code. There are many excellent programming books (ie, Petzold) to help with the Windows part.

I did some study today, my new code is in plan status. No old code available. I think Fortran90 is a great for scientific programming. Intel Fortran has three ways to do the interface,

1. Program interface useing VS C++, then connects with Intel Fortran using data transfer

2. Intel Fortran has Quickwindows, which can only make simple interface

3. Using Intel Fortran win32 API

I don't know how 3. compared with 1. I can write the whole in C++, but looks like C++ is not as good as fortran in Fortran in scientific programming.

[tropfen]

Quoting - truelies

I didn't some study today, my new code is in plan status. No old code available. I think Fortran90 is a great for scientific programming. Intel Fortran has three ways to do the interface,

1. Program interface useing VS C++, then connects with Intel Fortran using data transfer

2. Intel Fortran has Quickwindows, which can only make simple interface

3. Using Intel Fortran win32 API

I don't know how 3. compared with 1. I can write the whole in C++, but looks like C++ is not as good as fortran in Fortran in scientific programming.

Hello,

have a look into the book if possible:
Lawrence, Norman: "Compaq Visual Fortran - A Guide to Creating Windows Applications", Digital Press

• ISBN-10: 1555582494
• ISBN-13: 978-1555582494

It was written for Compaq Visual Fortran, but most of it is the same in IVF.

There are also few examples at the Xeffort Webpage.

Frank

[veitner]

Quoting - tropfen

Hello,

have a look into the book if possible:
Lawrence, Norman: "Compaq Visual Fortran - A Guide to Creating Windows Applications", Digital Press

• ISBN-10: 1555582494
• ISBN-13: 978-1555582494

It was written for Compaq Visual Fortran, but most of it is the same in IVF.

There are also few examples at the Xeffort Webpage.

Frank

Hi all,

Fortran + Java using JNI.

High performance during development because of IntelliJ.
High performance during runtime because of Fortran and JNI.

And OS-independent.

And the windows users? There are some different themes for Java to look almost native to windows users.

Veit