[antlr-interest] are ides a good thing? (was: Serious doubts onusage of incrementalparsinginides)

[Eduard Ralph]

> And what percentage of your time do spend actually *working* in your
> IDE?

Obviously more then you.

> I don't know what a good answer should be, but it should certainly be
> pretty low!

Do you actually *implement* what you think up? I mean unless you do academic
type of things, you end up spending a lot of time implementing. Most you do
in academic field is writing some sort of paper or the other and then making
a sort of prototype of proof of concept, which is a totally different kettle
of fish. And even then I spent 30% of my time coding.
I know every book on software engineering writes something to the effect
that you should spend at least 60% on planning. But unfortunately only very
few customers have had a course on software engineering and only limited
imagination concerning a finished software product. That does tend to skew
things.

> If I only spend 20% of my time in my IDE, then even an 80% improvement
> in my productivity within the IDE is not going to make much difference
> to my overall productivity, is it?

As a matter of calculation: If you increase your productivity around 80%,
you save 16% of your overall time you spend in an IDE. An improvement of
around 16% is a good thing, isn't it?

> Your point about a good tool and a good programmer is well made.
> However, in my (limited) experience, even with good programmers I find
> that the more powerful the tool is, the more it is misused to the
> detriment of the program... 

But that still is a very poor argument against an IDE. An analogy would be
something like: I don't use any electronic correction systems in my car,
because that makes me drive more reckless. The factor is still: use your
brain first.

You've recited a study from 1968 (which unfortunately I didn't have the
luxury of reading) which probably is a bad way to make your point, because
technology has changed markedly since then. You've also indicated that you
spend very little time doing implementation, so I would like to ask you:
have you ever implemented a piece of software that took 3 months or more to
complete? Even if you only spend 20% of your time on it? I think you would
find that your approach is slightly off synch with coding reality.

Greets,
Eddie

P.S.: I don't mean to get insulting, but your arguments to seem to imply
that you have a certain white tower syndrome. With that I mean that you
probably are very good at whatever abstract thing you do, but you seem to be
a bit high browed about those ones of us, who actually make *things*. As a
suggestion: try a project in the industry (best 4-10 people, 6-12 months of
coding), you might find it an interesting different way to software
development.