Re: Resolution of IMs 201, 202, 216

Hi John,

> Re: IM 201
>
> [...]
>
> With the committee's agreement I would like to propose that
> we rewrite at least the SystemVerilog portion of section 3.1.3
> to make it look more like the formal text that would go
> into a specification - i.e. clean up by taking words like
> "recommended", "proposed", etc. and consolidating the tables
> on page 11 and 12.

I don't see a reason why you couldn't do this.  Even if we did
not have 100% consensus on this, it is still worth doing as it
will be closer to the final text regardless of level of
agreement.  Doing this does not imply that no further changes
can be made, after all.

> Re: IM 202
>
> Basically the consensus was to not allow exported functions
> to call imported functions but to allow imported functions
> to call exported functions. This would retain, at a minimum,
> the basic functionality provided by SCE-MI 1 in terms of
> call recursion. At this point I don't see a reason to
> go beyond this for SCE-MI 2.
>
> Would it make sense to add a section to the working document
> stating this so we can think about closing it out ?

Yes, definitely!  I agree with you that we reached consensus
on the resolution of this IM.  Please add it to the document
and lets discuss from there.

> Re: IM 216
>
> In addition, I noticed in the last version of the working
> document, sections 5.2 and 6.1.2 can be combined. The IM
> in 5.2 is falsely labeled as IM 203 when it is really IM 216.
> I would like to propose that we re-work the document to
> combine these sections and also to add some text clarifying
> when it makes sense to use ratio based clocks and when
> to use period based clocks along the lines of the discussion
> we had next week.

Please do what you propose.  On the topic of ration based clocks
versus period based clocks I would prefer to have no restrictions
on the mixing of the two types in a given testbench.  I think
having restrictions will end up complicating implementation and
they are not necessary.  I propose that the following rules be
defined to determine the relationship between period based and
ratio based clocks:

  1) If there are no period based clocks, the testbench behaves
     as in SCE-MI 1.1.

  2) If there is at least one period based clock, the clock with
     the shortest period will have a period/frequency identical
     to the implicit 1/1 cclock.  All other periods can be
     related to the ratios from this basic information.

As you can see, all that is required is to find the fastest
period.  Since the infratstructure linker has to scan the
testbench for all cclocks anyway, this is not a big deal.

Thanks,
Per

-- 
Per Bojsen                                Email: <bojsen@zaiqtech.com>
Zaiq Technologies, Inc.                   WWW:   http://www.zaiqtech.com
78 Dragon Ct.                             Tel:   781 721 8229
Woburn, MA 01801                          Fax:   781 932 7488