[antlr-interest] Programmer modalities (was: Serious doubts on usage of incremental parsinginides)

John D. Mitchell johnm-antlr at non.net
Tue Apr 26 09:14:17 PDT 2005


>>>>> "Anthony" == Anthony Youngman <Anthony.Youngman at eca-international.com> writes:
[...]

> Are ides even a good thing? :-)

Alas, that depends on way too many variables to have any definitive,
general answer.


> Working in Visual Studio (iirc) it won't even let me make "mistakes",
> which is real pain in the neck when I'm half-way through typing something
> and then need to go off and check something. As soon as I take focus away
> from the line, it is checked, and if there's an error the line grabs
> focus again. BLOODY ARTIFICIAL STUPIDITY :-)

I haven't played with VS recently but tools which heavily constrain the
developer so that they can only ever write valid constructs definitely
cause problems for serious programmers.  However, that approach can work
wonderfully for simple end-user level "programming" tasks.

> And, I think it was IBM's work, it has been shown that code that is
> written and checked WITHOUT the aid of a computer tends to work better
> than code that has been written interactively. I *much* prefer working on
> paper - it's far easier to have an overview over the entire project - and
> all these fancy tools get in the way.

Such "clean room" engineering approaches are quite expensive.

[...]
> It's my guess that people who do make this transition are far less
> reliant on visuals or, like me, actually tend to positively *dislike*
> visuals.

False.  Different people are wired differently in terms of the dominance
order of the various sense modalities in their thinking.  It sounds like
neither of the top two modalities are visual for you (where it is one of
them for the majority of people).  IMHO, the tools and approaches you use
should play to your strengths.

Take care,
	John


More information about the antlr-interest mailing list