[11011001]

So, is it possible to get project builder to compile fortran code with the absoft f77 compiler?

I am slapping an interface on an old f77 program. I would like to use .nib's for the interface and pass off stuff gotten from the interface to the old fortran program. It's be nice to do the interface with IB instead of using fortran and absoft's MRWE library.

So, any way to do this?

[11011001]

Meh.. I'll just program the interface in Project Builder, and when it's done I'll copy the code to an Absoft project, so I can get at the fortran code. (have a prototype already)

It kinda sucks, I have to make this a Carbon app. That's okay, I guess.

Thanks anyways.

[Angus_D]

Or you could just use IB without PB?

And anyway... Fortran? Ewwww....

[Richard Edgar]

And anyway... Fortran? Ewwww....

What have you got against the later versions?

[11011001]

Originally posted by Angus_D:
Or you could just use IB without PB?

And anyway... Fortran? Ewwww....

hehe, ya, I could.. I don't know my way around Absoft's editor really well yet, so I feel more comfortable working in PB.

But right now, I am just laying down the interface, programming an editor for the input data file, I don't really need to stick the Fortran in until the end, and then all I have to do is copy the code, and resource files over to Absoft, and write a few functions to interface with the Fortran code.

Hehe.. and yup. Fortran, sigh. The fortran is an old program that does some chemical modeling. Fortran is still used a lot for scientific applications.. it's neat, but oh so picky. But one thing I hate about this fotran code is all the GOTO statements!!! Oh God! It's absolutely rediculous, it's so hard to follow. I now have a clear appreciation for why my old programming teachers told me to stay away from GOTOs. They are the .

[smeger]

When I was doing quantum computation simulations in grad school, I was forced to do 'em in Fortran (on Windows).

I found that Fortran promotes the least elegant code possible - it is one seriously fugly language. But it's fast. But it's also why I'm no longer in grad school.

I always planned to do some benchmarking against something like the Blitz++ numeric libs for C++, but never got around to it.

[Richard Edgar]

I found that Fortran promotes the least elegant code possible

In what way?

[invisibleX]

Originally posted by Richard Edgar:
In what way?

I'm a newbie here but I think he said it was the GOTO statements.

[smeger]

Originally posted by Richard Edgar:
In what way?

I found that the code I wrote in Fortran tended to be completely unsuitable for code reuse because it had absolutely no encapsulation of any kind, lots of globals, and tons of gotos. I also found the reading anybody else's Fortran code was a nightmare. This included my own old Fortran code. So again, I tended to have zero code-reuse.

I tend to think in an object-oriented, heavily factored way. Things as simple as calling functions seemed hackish to me in Fortran - the language seemed designed to just type out a bunch of commands and go.

I seriously detest this language.

[Richard Edgar]

because it had absolutely no encapsulation of any kind

In what way do modules and types with private components fail to provide encapsulation?

lots of globals

Strictly speaking, Fortran does not have any globals at all. It has things (a few things, actually) that can be made to act like globals. One uses them when convenient, but not for the sake of it.

and tons of gotos

Why is the fact that you used GOTOs a failure of Fortran? I've never felt the need for them. Fortran (like most languages) has a full set of flow control constructs. Perhaps you should have used them? As I mentioned in another thread, even C (that hallowed language) has setjump and longjump.

I tended to have zero code-reuse

I have not had that problem. Nor have the people who assembled netlib.

Things as simple as calling functions seemed hackish to me in Fortran

If the statment
y = myfunc(x)
looks hackish, might I ask what doesn't?