Re: [gtk-list] Re: Nevermind (was: Glib BTree & GTree wrapper)

[Andrae Muys <a muys mailbox uq edu au>]

On Thu, 10 Feb 2000, Antonio Campos wrote:

> Derek Simkowiak wrote:
> 
> I don't know much (not to say nothing), about data structures, but glib should
> be a very little library (that's the intended use, isn't it?), so I don't see
> the point of adding many complex structures to the library.
> If you want this structure added to glib, shouldn't you make a library apart
> from glib and deliver it separately from glib?
> 
> I hope I haven't been very rude...

No you haven't been rude, but I will disagree with you.

I personally would like to see glib expanded.  Good data structures are
critical to good code, and every program needs them.  I know I have a
handful of files containing my personal linked list, hash table, tree
implementations, but since I've found glib I have been avoiding
them.  Why?  For the same reason the STL has been so successful.  We all
use the same datastructures, and when a good library like glib, or the
STL, is provided we all get to use the same API's to those datastructures
too.  This makes code much more readable and far more maintainable by
other programmers as both the syntax and symantics of the various
operations on a programs data are already understood.  It simplifies
documentation, and makes it much easier for multiple people to work on the
same code base (a huge win for anyone using the Bazzar model).

In light of this, I would love to see *more* not less in glib.  I would
also like to see some of the Gtk+ specific stuff moved out into Gtk/Gdk so
glib can become a true datastructure and utility library independant of
problem domain.  One thing I have on my TODO list is to compare the
current state of glib with the STL and see where glib win's, loses, and
where there are holes that can be filled.  In fact I joined this mailing
list solely because I am interested in glib and couldn't find a glib-devel
list.

Andrae Muys

--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Andrae Muys <andrae@humbug.org.au> "Never ascribe to malice that which is
Senior IT Officer                   adequately explained by incompetence."
The Centre for Magnetic Resonance                   -Napoleon Bonaparte
The University of Queensland