Attendance
  Charlie Guy 
  Peter Ashenden
  Dave Bishop
  John Ries
  Eric Marshner
  Jim Lewis

Patent Slides context presented

FT 05

Question are they implicit or explicit.
to_string are explicitly declared with all predefined types.

Discussion on the format character for time and real.
Will reference C standard for meaning of format standard.
Action Item: David need to consult ISO C standard for actual definition
  of format string format
Action Item: Peter contact Andy to see if we can get a copy of the
   ISO C standard.
Action Item: for David to make some example of using the format specification

Only implemented f, e, E, and g format types.  No other formats.

Issue about sim_resolution being a deferred constant.  Deferred constant
is computed at elaboration time, resolution isn't set to simulation time.

Action Item for David to change sim_resolution to a function from deferred
constant.

Eric interested in a more general solution.Allow declaration of function as the
same name of the type.  Issue is that types cannot be overloaded. No resolution
on this point.

Peter brought up that he thought the requirement was to allow for strings of any
type. Thought that we should have implicit function that is defined for all
types. There are some issues that like generate string vs aggregate. Do you have
the type qualification.

Discussion of 'image and 'value and if we generalized them.  Use is an issue here.
The attribute current are on types not objects.  Could allow the attribute on objects
like 'range, 'low, 'left and so forth.

Jim said that he thought we had looked at 'image/'value and thought we couldn't
do this in fast track, and we should spend effort on the to_string functions.

Peter suggested that we define an implicit to_string that has only one
parameter, the value, and have David's to_string be addition to_strings that do
pretty printing of values.

Charlie: How do we define this to work?
Peter: We have a small number of type constructs.  We can define the to_string
for each scalar class of value and define the composite types in terms of the scalar
components.

Action Item for Peter: update proposal FT5 to allow for
to_string for every type.


FT08

David presented proposal.
Action item Jim, Web site needs update.
Action item David proposal needs removal of dread from summary.

Otherwise proposal is done.

Code review of CP proposals.

Discussion of SLA and SLL behavior. David believes that 
SLA is defined incorrectly, it should be defined as the
same as SLL. Not the definition of replicate the right most
bit. David has made the SLA for std_logic_vector match the
LRM definition.

Discussion on the numeric standard package shift_left
and shift_right functions whose behavior is different from the SLL/SLA/SRL/SRA.

Very big discussion on the fact the functions are different.
Should be make the consistent? Does it have compatibility issue.

Resolution is
Std_logic is left alone.  
Action item: David to modify the SLA and SRA for unsigned to be the same as SLL, and SRL.
 For signed SLL and SRL will be modified to match SLA and SRA functions.

Peter proposes that SLA and SRA shouldn't be defined for unsigned.
David counters that the SLA and SRA should be defined as SLL and
SRL.

Peter and David disagree.

Resolution is David's view.