Subject: Re: Summary of Four.
From: Kevin Cameron x3251 (dkc@galaxy.nsc.com)
Date: Wed Aug 15 2001 - 20:54:30 PDT
> From owner-verilog-ams@eda.org Wed Aug 15 19:42:40 2001
>
> I understand that the goal of the voting was to established the relative importance of the
> various issues. However because a "dont care" ranking is not handled consistently I fear
> the ranking could be somewhat skewed.
>
> The Avanti ranking of "dont care" items is a blank field which is interpreted by Excel as
> a zero and this consequently pushes these items towards the top of the list which I'm
> sure is not the desired effect.
>
> I wonder if National doesn't care about items 13 to 84a because the
> rankings for these items are exactly sequential.
>
> Kevin can you confirm this?
>
> If this is the case then the later items are being unfairly pushed towards the bottom of
> list.
>
> Thanks,
> --Martin
I ranked everything, but the list I gave to Vassilios used the web issue numbers which
only match for the first few items - so it could be easily confused. I only have an old
laptop at work for looking at Excel files and it is pretty difficult to see all the
relevent columns at the same time.
Kev.
PS: My priority order (highest first) by web issue # -
0007:For:Essential for general adoption in SOC design flow.
0002:For:Essential for back-annotation to work.
0003:For:Makes module A/D interface insertion independent of hierarchy.
0001:For:Gives most accurate timing, least difference when swapping modules from analog to digital
0004:For:Lets auto-inserted A/Ds check there context.
0013:For:Protects against propagation of bad data for poor arithmetic.
0014:For:Needed if you can't modify the source modules being connected.
0019:For:Provides an explicit method for defining the primitives, and hence makes netlists more portable.
0017:For:(See Issue #0006 too) Avoids arguments about exact translaton of "Spice" - translation becomes a user issue.
0018:For:Essential for behavioral/PLI access from digital to analog (modules may be swapped).
0016:For:Essential for getting valid digital states for initialization.
0005:For:More abstract, less prone to patent issues. Better if using many kernels.
0006:For:Mechanism allows support of new "primitives" (analog built-in models).
0804:--:No proposal
0801:--:No proposal
0802:No:Should reconcile with Verilog-2k
0803:--:No proposal
0015:For:Makes it consistent with C/C++ approach and helps users do redefinition.
0805:For:I assume the proposal is that 'real nets' be removed.
0806:For:The pre-processor directives need fixed.
0810:For:See Issue #5 && <a href="../misc/schd_sm.pdf">Mixed Signal Scheduling Semantics PDF</a>
0819:For:But, need definition of "successful".
0820:For:Agree with Antrim that unnecessary restrictions should be removed.
0824:For:See Issue #5 && <a href="../misc/schd_sm.pdf">Mixed Signal Scheduling Semantics PDF</a>
0825:No:I think $table should be handled through the rep-stop & external module mechanisms.
0828:For:See the rep-stop & external module/filter mechanism proposals.
0829:For:
0901:For:
0902:For:
0903:--:No proposal
0904:No:OOMRs are like ports in the referer and consequently may need disciplines defined. The deficiency is on the target end of the OOMR.
0905:For:Never made sense.
0907:No:Use NaN, 0.0 will look like ground.
0908:No:Should just remove "real nets"
0909:For:
0911:For:
0915:No:Proposal needs clarification - example doesn't appear to be a switch branch.
0917:For:
0918:For:
0919:For:
0922:No:Analog should be able to set X/Z on a digital signal - maybe just need clarification.
0923:No:Needs to be considered with discipline 'binding' and consequences of getting the wrong discipline.
0924:No:Need extended syntax for assigning disciplines to OOMR targets.
0925:No:The drivers of a signal and the receivers are seperate, i.e. the reg/expression is not re-written, only the value received by processes accessing it.
0926:No:Should use a "modified" assign with a different strength level.
0928:For:Start analog after all time-zero digital events.
0929:No:Leave for later.
0930:No:Needs carefull description as solver will re-run code repearedly for any solution.
0932:For:
0933:No:Keyword scope should be limited to blocks where they are used for backward compatibility.
0935:For:
0936:For:
0939:--:Consider with rep-stops.
0940:For:
0941:No:Consider later, OOMR and defparam sufficent for now.
1201:For:Make correction
1202:For:Make correction
1203:For:But use alternative description.
1204:For:
0807:--:No proposal
0808:--:No proposal
0809:--:No proposal
0811:--:No proposal
0812:--:No proposal
0813:--:No proposal
0814:--:No proposal
0815:--:No proposal
0816:--:No proposal
0817:--:No proposal
0818:--:No proposal
0821:--:No proposal
0822:--:No proposal
0823:--:No proposal
0826:--:No proposal
0827:--:No proposal
0912:--:No proposal
1205:--:
0008:--:Split
0009:--:Split
0012:--:Split
0010:--:
0011:--:
0934:--:
0937:--:
0938:--:
0916:--:
0920:--:
0921:--:
0906:--:
0910:--:
0913:--:
0914:--:
0927:--:
0931:--:
This archive was generated by hypermail 2b28 : Wed Aug 15 2001 - 20:56:50 PDT