?feasibility of numerical lib on the top of glib

["Sergey I. Panov" <sipan mit edu>]

glib has collection of nicely implemented container objects.
What are advantages and disadvantages of assembling numerical
library on the top of glib?



 Explanation(why I asked that question):

Through the years I was writing short C utilities for my personal
use. (I was piping data through them in gnuplot). I'd like to get
organized and redo everything a sa nice numerical library. I played
with it a bit and for now I use my own vector and matix objects. 

I get across gsl recently, but their container objects are not exactly
what I need --- e.g. I need resizeable vectors and gsl vectors can not
be resized. Before going to gsl with suggestion to change their object
model, I'd like to have some code. That is why I am asking that
question. What if there some disadvantages, that I am not aware of?

So far I see no problems -- GArray can be used to hold doubles 
and it took me a few minutes to make typesafe GDoubleArray on top of
the GArray (analogous to GByteArray). 


Sergy