glob2-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [glob2-devel] coding guidlines


From: Stephane Magnenat
Subject: Re: [glob2-devel] coding guidlines
Date: Tue, 21 Oct 2008 14:13:01 +0200 (CEST)
User-agent: SquirrelMail/1.4.9a

> Bradley Arsenault wrote:
>> I disagree with these hard coded guidelines. I feel that it should be up
>> to the programmers discretion what kind of organization they do. It
>
> for some reasons this "it's up to the programmer" lead to a
> - very flat folder structure with
> - tons of files in each folder with
> - barely any relation between most of the files in a folder
> - some folders that split src and lib others that don't

The original idea was to split into directories elements that could be
used separately (I have used libgag alone in personal projects).

> - files defining about 50 classes
> - files with 6000 lines of code

But what is the problem with this? If we have autodoc with correct
descriptions of classes and of logical modules that encompass several
classes (in the opposition to naive documentation of members), that is
good.

> how about we put all in one file? nowerdays pc's wouldn't complain and
> then everything is in one place and we can navigate in the code much
> faster!!1

I would love this, but C++ is crappy and does not support the use of
non-previously defined classes ;-)

The latter sentence is a half-joke, as I do not believe that excessive
split in code do result in better quality. Of course, I admit that files
should stay at a reasonable size, but again I would like to emphase that
understanding of the whole structure is much better served by descriptions
of the modules at different levels of details than by global and
context-independant syntaxic rules.

I know that software engineering scholars would object, but I have never
seen any software engineering scholar writing something of the complexity
of glob2; and for sure not in the time we did glob2.

Have a nice day,

Steph





reply via email to

[Prev in Thread] Current Thread [Next in Thread]