[ibis-macro] Re: Status of IBIS-ISS Draft Revisions
- From: "Walter Katz" <wkatz@xxxxxxxxxx>
- To: <Arpad_Muranyi@xxxxxxxxxx>, "'IBIS-ATM'" <ibis-macro@xxxxxxxxxxxxx>
- Date: Fri, 18 Jun 2010 11:24:20 -0400
Mike, Arpad,
I have not seen the discussion that led to this e-mail, but I am surmising
that there is a proposal to adding NEW "spice" keywords, like .title and
.version to the ISS specification. If these are true HSPICE keywords, then
OK, but if these are truly "NEW", then they must be included as "Comments
Only", e.g.
*ISS Title
*ISS Version
Walter
-----Original Message-----
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Muranyi, Arpad
Sent: Friday, June 18, 2010 10:36 AM
To: IBIS-ATM
Subject: [ibis-macro] Re: Status of IBIS-ISS Draft Revisions
Mike,
Thanks for the update on your progress of the ISS
editorial work. Here are my answers to your
recommendations:
0) I agree about your conclusion that we need a
definition or glossary section.
1) Adding .title and .end may not be a good idea
because this is a "subcircuit specification", not
a complete netlist specification. An ISS file
would most likely be included into a netlist, and
therefore an .end statement would not work.
Even though a .title statement would technically
work, it doesn't seem to make sense, for the same
reason.
2) I think making the parameters all local is a
good suggestion.
3) Since this is a subcircuit specification, these
files will be included into bigger netlists. For this
reason we cannot make the assumption about the first
line being a title, because in the big netlist that
line will never be the first line after the inclusion.
4) We should support abbreviated versions of all
dot statement the same way as HSPICE does it.
5) I think a .version keyword may be safer, especially
considering that this is a subcircuit spec, not a
complete netlisting spec. But we will have to implement
this carefully because these subcircuits might coexist
with other subcircuits having a different version number.
6) Again, this is a subcircuit spec, not a complete
netlisting spec. These files will always start with
a .subcircuit and a .ends statement. What you are
talking about is not applicable to this, since
subcircuits can be included multiple times already.
Thanks,
Arpad
========================================================
_____
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Mirmak, Michael
Sent: Wednesday, June 16, 2010 9:00 PM
To: IBIS-ATM
Subject: [ibis-macro] Status of IBIS-ISS Draft Revisions
As several people have asked, here's a more formal status report on the
editing of the IBIS-ISS draft document. As I will be unable to attend
Tuesday's meeting, I thought I would "kick off" some discussions and get
resolutions on key issues via e-mail.
* MM
6/16/2010 - IBIS-ISS Draft Update Status
The document is unfortunately not readily consumable by the wider industry
as written today, and would, without major changes, not be a useful
reference for users who do not have extensive experience SPICE experience.
The first round of edits concentrate on making the document better organized
, mostly for simplification and clarity. Edits focus on word choice,
compliance with industry standards writing practices and self-consistency.
Many technical concerns have arisen, which are detailed below. A "sense of
ATM Task Group" for each of these issues and proposals is requested.
Many edits are motivated by undefined terms, terms which are used only once
or twice, or terms which change meaning in context within the document.
Adding a glossary of terms is strongly suggested. Terms that cause
definition or usage issues include:
- delimiter
- token
- command
- keyword
- statement
- card
A revision 0.3 draft and list of changes will be made available for formal
review shortly.
-------------------------------------------------------------
Proposals, Recommendations and Calls for Comment
1) Only a few "dot" statements are now permitted in the specification.
These are:
.ends
.include
.model (only used for W-element and S-element definitions)
.parameter
.subckt
This excludes several common "dot" statements, including:
.title
.lib
.macro
.end
Recommendation: maintain exclusion of .macro and .lib. Add .title and .end
(particularly the latter). Approved?
2) Proposal (related to libraries/.lib): all parameters now have local scope
and must be explicitly passed into subcircuit instantiations. Approved?
3) Older SPICE variants assume that the first line of an input file is a
comment, regardless of the presence of a comment character. Recommendation:
add support for this assumption. Approved?
4) SPICE variants commonly support ".inc" as well as ".include" and ".param"
in addition to ".parameter". The current document is inconsistent on
whether these two abbreviations are legal or not.
Recommendation: support ".inc" and ".include"; support ".param" and
".parameter".
5) The current document contains no provision for recognizing the version of
ISS supported. Recommendation: assuming that ISS will be revised, a
.version or first-line (comment) version identifier should be required.
Approved?
6) Some SPICE variants support input files containing multiple netlists, so
long as these are separated by .end. Recommendation: preserve this
functionality. Approved?
Other related posts:
