Python vs. C++ Builder - speed of development

[Laura Creighton]

> Laura Creighton wrote:
> >> Patrick:
> >>
> >> Is, or under which circumstances is the time which can be gained by
> >> using a RAD-IDE equivalent to the time gained by using Python
> >> directly?
> >
> > My guess is 'never'.
> 
> That's an extreme viewpoint.
> 
> > The problem with developing in C++ is not that
> > you don't have an IDE that makes development faster, but that in C++
> > there are so many ways to shoot yourself in the foot that you have
> > to spend forever re-reading your code to make sure that it doesn't
> > interact in nasty, horrible ways with other people's code.
> 
> Then don't interact with other people's code.

Not an option.  My job was interacting with other people's code.

>  Polymorphize your own tree
> where you know what you're doing, and keep the interface that everyone else
> sees very very simple.  This is no different a problem of API speccing than
> in any other language.  C++ simply provides you more ways to be stupid.
> Nothing's twisting your arm to be stupid however.  You can count on other
> people doing dumb things in their code, but if you're doing dumb things in
> your own code, that's your fault.  It's not that hard to ignore most of the
> "features" of C++ and just use basic class hierarchies to get simple things
> done.
> 
> > Or you spend your life debugging other people's badly designed templates.
> 
> Then don't use other people's templates.

Also not an option.  Debugging other peoples templates was the job.

> 
> > Or on the phone to Borland saying that you have found another bug in
> > the destructors ... or ....
> 
> Then don't use Borland products.  

Also wasn't my decision, but I have found Borland to be the most
responsive of all C++ vendors.  That is, it was _worth it_ to be
on the phone explaining bugs to Borland developers.  They were 
listening, anbd sending me test cases, and we were fixing things.

> Granted I spend my time doing 3D graphics,
> and I deliberately keep the OO aspect "lean and mean."  But I've never
> called Microsoft VC++ tech support for anything.  

I've never got them to fix anything myself.

> Also I've restricted
> myself to mature compiler products, I will *not* jump on the bandwagon of
> whatever crap MS has just put out.  Wait for a few service packs, let others
> take the lumps before you switch over.
> 
> I think, to answer the original poster's question honestly, you have to
> consider the complexity of what you're undertaking.  You seem to be talking
> about ridiculous, corporate-dysfunctional style complexity.  Sure, you'll
> have problems there with C++.  But you will also have them with Python, for
> different reasons.  For instance, no type checking.  The underlying problem
> is the dysfunctional corporation.  Change instead to a small team of
> programmers who know how to "fly in formation" and the problems go away for
> both languages.
> 
> > I don't think that it is possible to write an IDE that makes me as
> > productive in C++ as I am in Python with ed(1).
> 
> I can never take people seriously who say things like this.  You like ed
> because you've invested lotsa time in ed.  If you invested similar time in
> any other tool, you would like it also.  Why?  Because if you like ed,
> clearly you like what you know, not what is inherently a better paradigm for
> text entry.

You misunderstand.  Ed is the most primative tool there is in which
you can get useful work done.  It is as far away from an IDE as you
will get, which is why it is useful for his comparison.  He wants to
know if a speed up in IDE will match the speedup you get in Python for
C++.  And the answer is no.  Because the time you spend in C++ is
not mostly spent inside your development environment.  It is spent
outside, as you do something other than write code.

> 
> If you had said Emacs, I would take you half-seriously.  

This is what I develop in normally. 

>  If you said you had
> a mouse-driven GUI for your Emacs, because you're not the kind of genius who
> can remember every keystroke, you'd have my attention on issues of
> productivity.  There's nothing wrong with keyboard shortcuts, but users
> should not be required to know them.

Typing 400 words a minute helps as well.  I cannot do that yet, but
one person here at Strakt can.  Utterly awesome.  I thought that
200 words was good....

> 
> > That is because almost
> > all my time spent in C++ come from some other place than actually
> > writing code from scratch.  Other people's milage may vary.
> 
> Find a different job then.  Are you getting paid well for your C++
> headaches?  I wouldn't put up with it if it's as onerous as you make it
> sound.  Unless the money is really really good.

Last  C++ job I had paid roughly $400 Euros an hour.  And the reason 
they paid me this was because indeed, it was as onerous as I made it
sound.  This is what I was paid to do all day long.

You need to get over your notion that there cannot exist a person who
is excellent at something and who hates it at the same time.  

> 
> --
> Cheers,                         www.3DProgrammer.com
> Brandon Van Every               Seattle, WA
> 
> 20% of the world is real.
> 80% is gobbledygook we make up inside our own heads.
> 
> -- 
> http://mail.python.org/mailman/listinfo/python-list

Laura Creighton