On 4 Mai, 12:55, Marco van de Voort <mar...@[EMAIL PROTECTED]
> wrote:
> On 2008-05-04, thomas.mer...@[EMAIL PROTECTED]
<thomas.mer...@[EMAIL PROTECTED]
> wrote:
>
> >> > programming languages), but instead some programs start to depend
on the
> >> > IDE. With Seed7 I want to be more or less IDE independent.
>
> >> IDEs are im****tant for productivity.
>
> > IMHO IDEs do improve productivity, but the bottleneck is
> > somewhere else. I think that most of the time spend
> > in programming is not spend writing the program. It
> > is the time used to find an algorithm, to decide over the
> > interfaces to other modules and and the time spend for
> > similar high level tasks.
>
> Not every programming task is a math problem. (G)UI, input validation,
data
> im****t/ex****t and businesslogic take a non insignificant part of the
part
> too.
I was speaking of a bigger project where I was part of
(and this is not the only example I was involved).
Dozens of people working for years. Some design decisions
(done by managers) lead to loss of "productivity" which
was much bigger than what an IDE could improve in
productivity:
- A mainframe had to be used for part of the software.
- The rest of the software had to run on PCs.
- The PCs had to use windows (Linux and Unix where
strictly forbidden because of political reasons).
- C++ was used for some part of the software which
should run at the mainframe and the PCs.
- The interface part of the software was written
in Java.
- The Java part used a homemade framework instead of
a standard framework for the interface (that hat
an influence at the turnover of Java developers).
- The C++ part had to be developed with Visual C++
while the Java developers used Eclipse.
- Since Visual was not available at the mainframe a
make based build process was used there.
- The PC only part of the C++ software was so
dependend on MS libraries and the Visual build
process that one employee failed to move the C++
part to Eclipse/make after months of work.
- Demands for performance without any performance
testing lead to premature optimization, which
according to D.E.Knuth is the root of all evil.
- Configuration values had to be taken from two
independent configuration libraries. The interface
to this libraries was just the opposite of a
key value logic (from a viewpoint of complexity).
- An external binary only C library had to be used
although an in house development of another
branch of the company was available and would have
fitted much better.
- The OO data model had some severe bugs.
- Although the OO data model was changed and
adjusted to customer requirements all the time
the severe bugs stayed in for political reasons.
- Therefore all developers had to code around this
bugs in almost all modules, instead of fixing this
bugs in the OO model.
- The managers admited the situation but did nothing
to improve it.
- If someone pointed out the problems and a way to
solve them he was classified as beeing negative.
- This list is by far not complete.
I consider all this decisions in the above example
design decisions which lead to a productivity loss.
Compared to these losses the gains by using an IDE are
negligible. If you think that you can get back all the
lost time by something the IDE can do for you, just go on.
And BTW this is not the only example I know of.
Greetings Thomas Mertes
Seed7 Homepage: http://seed7.sourceforge.net
Seed7 - The extensible programming language: User defined statements
and operators, abstract data types, templates without special
syntax, OO with interfaces and multiple dispatch, statically typed,
interpreted or compiled, ****table, runs under linux/unix/windows.
|
