Re: Are subroutines codependent?

Admittedly, there are things I'd like to see in RDi - like the state of indenting even when there are blank lines - as in Visual Studio, IIRC.

Dieter, you mention that "...there is no prototype anyway, declaring the interface is sufficient..." - this is now the case in RPG, as well, for procedures in the same module.

There is also a wizard for new procedures - opinions vary on how good it is, but it IS there. You can select different kinds of procedures, and it might have an option that does not create a prototype - I haven't tried that.

I agree that it would be a nice feature if RDi would "notice" when a procedure is not defined yet - it actually might, with live outline - does anything show up in the left to indicate this?

So Dieter, I suggest that you create an RFE for this feature - I would vote for it. It does seem to line up nicely with the recent change that prototypes are not needed for local procedures.

Vern

On 7/9/2014 1:05 AM, D*B wrote:

<Birgitta>
IMHO for beginners using ILE concepts subroutines might still be helpful for
encapsulating and structuring the encapsulated source code.
But as soon as they have learned to think really modular, they'll find out
subroutines are not needed
</Birgitta>

Might be true for ILE newbies with some years RPG background. Real beginners should not hear about subroutines before they have incorporated to use procedures. It's better to learn first how to do and maybe afterwards(or maybe never), how people did write RPG programms with RPGII. Too many people are stuck at writing RPGII logic, but using free format.
One of the big problems are the tools!!! RDI (or how it might be named today) is looking very modern, but under the cover LPEX logic is working. For example: using Eclipse free Java Editor to write a new method (= procedure), you could just use the not existing method, the Editor would complain immediate "I don't know this method" and you could choose "Then implement it" and the editor writes for you all formal stuff, including needed parameters and return value (there is no prototype anyway, declaring the interface is sufficient).

D*B