Phone Conference P1450.1 Working Doc Subgroup
Thurs Jan 31, 10:00 am PST
---------------------------------------------
Attendees:
----------
Jim Teisher
Tony Taylor
Greg Maston
Peter Wohl

Agenda
------
- Continue resolution of internal review cycle and finalize draft 13 
  document containing reviewer resolutions. Focus on red text in
  resolution document (indicating open issues).

- Discuss recommended usage doc and dot0 clarification doc as
  distributed by Greg just before the meeting.

Issues
------
Started with item DM-1, bcause Tony identified additional constructs
necessary to support multiple instances of hierarchical scanchains and
this was a change to the resolution of the last meeting. Started
discussion of hierarchy/instance-name of scancells, using example
Annex M, page 84-85. Proposed solution presented a slightly different
mechanism to define an instance name. After some discussion, Peter
proposed a change to just the ScanCell syntax, from:

  ScanCells { !; a1."A"; a2."A"; !; a3."B"; a4."A"; !; a5."B"; }

to:

  ScanCells { !; LSSD_MSS->"A"->a1; LSSD_MSS->"A"->a2; ... }

Discussion about the ordering and semantic intent at each level of
this construct: scanstruct -> chainname ->
instancename. Accepted this ordering and semantic interpretation of
each level of this statement.

Discussion about needing more levels. Since each scanchain description
is complete (and resolvable down to a set of uniquely-named cells) in
it's own hierachical definition, any higher level only needs to
reference that chain (with an instancename) and still get unique
scancells, so additional levels in this statement are not necessary.

The WG accepted this proposal, with the following changes to the
document ([AI0]):
(a) remove the ScanModule or any other referencing-statement-construct
from the language. These are unnneed when the referencing is performed
explicitly in the statement.
(b) remove from the language any reference to a common namespace for
scancells and scanchains. This is an unncessary restriction when the
scanchains are explicitly referenced in the construct above. This
makes dot0-compliant files compatible with dot1 again.
(c) add the '->' operator to language constructs (table 2)
(d) add the scanstructure_name->scanchainname->isntancename; statement
to the ScanCells definition.

Recommended Usage Doc
---------------------
Skipped to item DS15 for a discussion about the recommended usage
doc. Brief discussion about whether this document should be made
generally available on the website (as has been done with the "sample
grammer" for dot0), or balloted as a recommended usage doc. 
[AI1] to Tony to talk to Adam about whether we can put recommended
usage doc on the web directly.

Greg took [AI2] to make the following changes to this doc (from Tony):
- Add bullet to reference the "sample grammer" doc in this document.
- Add references to all dot-efforts here.

Discussed the issue that Tony has had no response from the IEEE about
placing text documents of the 1450-1999 examples on the website yet.

Tony raised a discussion about a potential solution to the lack of a
double-quote internal to a string. Greg took [AI3] to present one
'recommended mechanism' to indicate a double-quote, by dotting around
an empty string, to the WG. Again this is just a recommendation.

JC2: reviewed text in resolution and WG agreed to it.

DS4: still open ([AI4])

DS6: decision to NOT use commas is still agreed to. Added comment to
resolutions doc that D13 has been updated to remove commas as a
defined construct in the language.

DS7: still open ([AI5])

DS14: still open ([AI6])

New Items
---------
pg 28 parallelpatlist issues:

editorial suggestion from Tony: move clause 12.3 into clause 12.1 (6)
since it discusses patset behavior only. Agreed to by WG.

technical discussion from Tony, on parallelpatlist constructs. Consider design where two
subcomponents access the same top-level signal.

Tony proposes: it is ok to have signal overlap, but the data behind each
signal shall be identical.

Clarification Doc
-----------------
everyone review references 9-17


Next Meeting
------------
in two weeks; Feb 14, 2002.

AIs
---
[AI0] Tony - change scanstruct hierarchy constructs as identified.
[AI1] Tony - talk to Adam about placing recommended usage doc on web.
[AI2] Greg - make identified changes to the recommended usage doc.
[AI4] Greg to rewite last paragraph of section 6.1
[AI5] Doug to provide expression-type-clarification diagram or text.
[AI6] Tony to put Doug's X-stmnt-position-add't into resolution doc.