Re: [vhdl-200x] Modular types

On Wed, 2014-07-09 at 22:08 +0200, whygee@f-cpu.org wrote:
> Le 2014-07-09 21:44, Brian Drummond a écrit :
> > On Wed, 2014-06-25 at 14:17 +0000, Martin.J Thompson wrote:
> >> So, are there any features you would like which are not mentioned?
> >> Constraints which do not make sense?
> > 
> > I would agree but I would suggest (as Whygee suggested following his
> > experiments) adding shift and rotate operators.
> > 
> > The meaning of these is easy to understand for the modulo-2**n case. 
> > but
> > may require some thought in the more general case (I would suggest 
> > shift
> > on a bit position basis, followed by a modulo operation like the and/or
> > operations.
> > 
> > - Brian
> 
> Hello,

> why are we spending time on other bases, since they
> don't bring speed or ease advantages ?

Other bases have uses as counters, minute hands, etc. For these, one
could argue that the boolean and shift operators have few if any uses,
and I would agree.

But if we are providing boolean and shift operators for the special case
of modulo 2**n, then we have to do one of:

1) restrict modular types to mod 2**n. Prohibits other legitimate uses.
IMO : Just No.
2) Define boolean and shift only for mod 2**n and prohibit for other
cases. Ugly irregular language feature. IMO : No.
3) Define boolean and shift only for mod 2**n and leave their semantics
undefined for other cases. IMO : that would not be appropriate in Ada.
4) Have well defined semantics for all cases, in such a way that
efficiency and implementation burden are tolerable, and that the common
2**n case has the desired properties. IMO : this is the only tolerable
option of the first four.
5) Split modular types into two  : power of two, and other. IMO : too
much complexity
6) ???

- Brian


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.