Subject | Re: [IB-Architect] enhancement to procedure/trigger language |
---|---|
Author | Jim Starkey |
Post date | 2001-06-07T14:19Z |
At 08:51 AM 6/7/01 +0500, Dimitry Sibiryakov wrote:
is needed. Java, as a design goal, inherited C syntax to the
degree appropriate and hence the unlabelled break. The Firebird
language is undo no such constraint.
It may indeed make sense to include both a labelled and unlabelled
break because simpler version is all that is necessary in 85% of
the cases. But I can't accept the argument, "hell, toss in everything
and let somebody else figure it out." Bad thinking, bad methodology.
Decide first on requirements then the minimal language elements
that meets the requirements.
Jim Starkey
>>The essence of bad language design is two constructs where one
>>Thoughts?
>
> Usually, when I ask my users such kind of question, they
>answer: "make them both".
> I think, "break" is a good feature and it allows avoid "goto"
>in most cases. But "goto" gives more powerful execution control.
>Probably, that is my opinion because I am a Delphi programmer.
> So, my twopenny is: "Make them both and let people use
>what they like."
>
is needed. Java, as a design goal, inherited C syntax to the
degree appropriate and hence the unlabelled break. The Firebird
language is undo no such constraint.
It may indeed make sense to include both a labelled and unlabelled
break because simpler version is all that is necessary in 85% of
the cases. But I can't accept the argument, "hell, toss in everything
and let somebody else figure it out." Bad thinking, bad methodology.
Decide first on requirements then the minimal language elements
that meets the requirements.
Jim Starkey