What is the D plan's to become a used language?

[bioinfornatics via Digitalmars-d]

Dear,

My topic is a kick in the anthill. I hope this will help D to 
think on his problems.

D exist since 1999, if we look behind, what is done?
We have :
- a huge cemetery of D project
- no D killer application
- miss the goal what to do to improve D experience
- each new D release your application is broken and often with
some D compiler bugs
With around 15 years of works we are at same state as the
beginning.
What to do:
   - Stop to add new feature in D (new annotation or whatever is
not an urgent needs)
   - Get a compiler and a language stable
   - When this is done work on some free (libre) framework in some
specific fields
      * grid computing (sub-field: data-computing)
      * web-server
      * graphics computing (sub-field: gui )
And not more, not to disperse …


 From the beginning D had a bad start (1999-200X):
- dmd compiler was not totally open.
- we got a fight on standard library phobos - tango
With this language started to get a bad publicity
- d2 is come with a new syntax which kill lot of project: -> 
http://www.dsource.org/
survivor are: phobos: derelict, gtkd all others are dead

We get for years deprecated module: xml, json, mmfile

When we introduce a new feature we release it with a partial use:
- @safe often is not used as is not yet implemented into phobos
- range even the basic io as ByLine which implement range can't 
work with std.range.takeOne as is not an Input Range. While a 
basic function such as takeOne do not really need to use save 
method to work a forward range is enough.

So we adding feature and going forward without consolidate the 
previous works.
That is like you give to a children a gift, barely opened you 
give another gift.

Currently D project are:
- not open project to make money (no code shared)
- personal code ( which fit only one point of view )

Really I hope to see D3 with a clean model and a stable language. 
Do not release it as a toy like was done for D1 and D2.
Release D3 with much ah possible no compiler bug and a well 
implemented standard library. We have get enough experience to do 
it.


Why we do not create a useful D programming environment as is 
done is java, python, ruby.