Re: [vhdl-200x] Corrections to Minutes for VHDL-200X-FT meeting, San Jose Dec 4, 2003

Subject: Re: [vhdl-200x] Corrections to Minutes for VHDL-200X-FT meeting, San Jose Dec 4, 2003
From: Andy D Jones (andy.d.jones@lmco.com)
Date: Tue Dec 16 2003 - 07:39:37 PST

Well said!  We need more functionality, not shortcuts!

Am I the only one that is concerned with defining '1' in std_logic to be TRUE? (and conversely, '0' to be FALSE)?  VHDL is supposed to be technology independent! Its application extends far beyond the IC, to boards where active low signals are still preferred.  This definition would make vhdl unusable at other than the IC and FPGA level.

The point is, 1164 was created to allow more accurate modelling of actual hardware signal levels, etc., not to enhance behavioral (which most of RTL has become) semantics.  It explicitly divorces signal levels from boolean interpretations for just that reason.  If a circuit needs to do something in response to a high level on a signal, then it should be so stated. It should not be confused with whether a boolean condition indicated by that signal level is TRUE (whatever TRUE means).

Numeric_std.unsigned was kept separate from std_logic_vector for the same reason. An implied conversion from vector to number space is ambiguous, the same way an implied coversion from signal level to boolean is ambigous.  And VHDL is all about avoiding ambiguity!!!

The example given about unsigned addition adds nothing to the argument.  The mear usage of an unsigned data type gives explicitness to the definition of the operation to be performed.  In the Unsigned data type, a '1' means a given power of two if set to one.  This is part and parcel of the definition of unsigned.

If we allow implied conversions of std_logic to boolean in conditionals today, what will the whining masses want us to allow next?  And then they will have precedence for getting their way! 

What happens when someone writes "elsif clk then..."?  Do we assume that he wants the rising edge?  What if he writes "elsif not clk then..."?  Do we assume it is not a clock event, or that he wants the falling edge?  This is exactly what I've been talking about:  where will designers want  to use this that is completely ambiguous and confusing?

I say we keep std_logic at the signal level, and not have an implicit conversion to boolean under any circumstances.  If you want to be able to avoid conversions in conditional expressions, use boolean data types in the first place, and leave std_logic as it is!

Andy Jones
Lockheed Martin
Missiles & Fire Control
Dallas TX


Munden Rick wrote:

(snip...)
Sometimes it is appropriate to build in stucco and other times in stone.  We don't need to merge VHDL and Verilog to come out with something in the middle.  Each language has its place.  I would rather see us concentrate on adding actual capability than just adding shortcuts.

This archive was generated by hypermail 2b28 : Tue Dec 16 2003 - 07:56:12 PST