Re: Why RPG IV Sucks

[Alan C <steelville@xxxxxxxxxxx>]

Raul,

Your comments do not disagree with Bob's at all, in fact in another 
comment of his while responding to another, Bob agreed with this point.

The comments he made point to the way nulls were implemented inside RPG, 
something he can obviously handle very well, but that there are probably 
better ways to handle it.

In fact we owe him for some of the designs of new functions in RPG over 
the years that make our life easier, due to his advocacy with the IBM 
compiler team (and one or two that he now regrets himself :-) ).

And the " = *on" is certainly redundant, but it was probably used for 
emphasis in his point.

As to the implementation, it must have been difficult to think about, 
because in SQL (which foisted "nulls" on us in the first place), there 
is a three-way logic matrix used to calculate logical variables, and 
sometimes it seems to me like a square peg in a round hole. I've read 
that some "respected" computer science voices have even objected to the 
whole idea. However, we can handle it, we are of the initiated elite RPG 
fraternity, to paraphrase Booth!

-- Alan


Raul Jager wrote:

> I disagree with you comment, any serious programmer does understand 
> the call to a funtion, and the manual explains clearly what the 
> function does.  I find the " = *on " redundant.
>
> NULL values are a relatively basic feature in data bases, even if our 
> S/36 baground didn't have it.
>
> I don't care if a beginner programmer finds something new for him.  
> Keep up to date with manuals.
> _______________________________________________________________________________________________________ 
>
> Bob Cozzi wrote:
>
> > If %NULLIND(Date) = *ON
> >  // Your date is meaningless!
> > /endif
> >
> > One issue with this approach is that nobody understands it, it isn't 
> > well known, and it is virtually never used.  Also even if the null 
> > indicator for the DATE field is *ON, you can still use that field 
> > without restriction. That is no notification is sent if you now move 
> > or multiply that null date with another field.  Why? Because the null 
> > indicatory is separate from the data.  Not a good solution at all in 
> > my view.
> > -Bob