What doesn't "work"?
What libraries do you use: std_logic_arith or numeric_std?

Sorry, I think I wasn't clear enough. I'm using numeric_std to convert 
std_logic_vector to unsigned.  For example: When I use "to_unsigned()" 
it doesn't work. But "unsigned()" works.

Thank you

> For example: When I use "to_unsigned()" it doesn't work.
> But "unsigned()" works.
Of course: these are different functions with different operands. The 
to_unsigned() expects an integer as first operand and a vector length as 
second operand. Then it converts the integer to a unsigned vector of the 
desired length.
The unsigned() cast converts nothing, it just takes the std_logic vector 
and interprets it as an unsigned vector.

Try this:
http://www.lothar-miller.de/s9y/categories/16-Numeric_Std
Its German. But a picture says more than 1000 words...   ;-)

Hi Lothar, thank you for your answer. Can you explain how

1

unsigned(...)

 and

1

signed(...)

 works? They are not defined as functions in numeric_std. Thank you.

See the definition of those types:

1

  type UNSIGNED is array (NATURAL range <>) of STD_LOGIC;

2

  type SIGNED is array (NATURAL range <>) of STD_LOGIC;

  (from numeric_std)
And the one for this here:

1

  TYPE std_logic_vector IS ARRAY ( NATURAL RANGE <>) OF std_logic;

  (from Std_Logic_1164)

Then you will see, that the cast only tells the synthesizer to look at 
the very same bits in another way.

This also works the other way:
If you write "110001110001" then you don't know whether its signed, 
unsigned or a std_logic_vector (it also could be a bit_vector or 
somethin else). Then you must tell the synthesizer what you mean. This 
is done with a qualifier
http://www.lothar-miller.de/s9y/archives/82-Qualifier.html