language growth

[sebastien]

>If every significant language improvement wereadded to the language, 
>it wouldn't be an improvement at all.  All those
>good things would add up to a bad thing.  The optimal size for a language
>is not enormous.
I think this is true for any produce in cumputer dvlpment.
The author start by delivering a framework to include the set of features 
which is, at this time, the features that this kind of tools must contain.
The framework succed because it provide a good abstraction and integrity
between the features. Then use of theses features become more common for
the users, some wich was important at the begining become obsolete because 
new way to use the tool are discovered.
Quite time larter need for new features are discovered, because new market 
emerge for the product or because users become more familar with the
tool and want to use it more intensively.

Then two roads are in front of the product:

1- The product don't evolve and let the place to a new and richest produce wich
will integrate this new needs. Think about, awk and perl for sample.

2- The product try to evolve and intergrate new features. But the problem is that
the initial framework hasn't been designed for theses features. So the product 
loose integrity and become more complex. Then a nich appear for a product that
provide a new framework with better integrity between the new set of essantial 
features.

But with some precautions, the road 2 is much longger than the first one for
the product.