Phone Conference P1450.1 Working Doc Subgroup
Thurs. July 18, 2002 10:00 am PDT

Attendees:
Doug Sprague
Greg Maston
Peter Wohl
Tony Taylor
Jim Teisher
Dave Kellerman

Documents
p1450.1 D14 draft July 16.
D14 review-resolution document July 16.

Agenda
Starting with the agenda mailed by Tony; Peter requested a different
order (move BistStructures discussion up).

Jim identified the following additional concerns to add to the agenda:
- not all statements have keywords-first
- unclear definition of expression types
- should the integer-list construct be generalized?

Discussion

TT-1: any issues with new section 1 present in this doc? Greg moved to
accept this new section and close the AI as resolved, Doug seconded,
no discussion, approved unanimously.

TT-2: removal of BistStructures. Doug asked if there has been any
explicit objections to this section. The CTL group (.6) and several
other people have raised concerns that the constructs are not
sufficiently general (note GR-5). Some discussion about the loss of
this information and potential impact to applications; identification
that all data present here is not necessary to execute the
patterns. Recommendation that the removal of this information be
distributed to a wider set of STIL peoples - Tony took [AI1] to send
out email on STIL reflector about this proposal to remove
BistStructures from dot1.

GR-7: new Variables section. General agreement that this is a better
structure; less confusing than mixing them under the Signals. There is
a "Variable" block in the Spec, some discussion about the
differentiation of the 's'. Consensus in this meeting in support of
the separate block, name of block is in question. Also Doug asked
about the use of a ';' statement in the enumeration values; Greg
indicated that this was consistent with the language and the LAST time
we DIDN'T do this (in ScanCells) it forced additions in .1 to support
this. However this particular issue is still open.

DM-5: Issue resolved.

DM-3: To resolve this issue, Tony moved annex K into section 13.4 and
made the syntax consistent. Issue resolved pending further review by
everyone. Some discussion about Jason's Pattern Merging and whether
that proposal was part of this synchronization issue, not closed
pending participation from Jason.

GM-6: Lockstep. This section to be reviewed by the Working
Group. [AI2] to Tony to change the example on pg 32
'ParallelPatternList' to 'ParallelPatList' and add missing LockStep
declaration.

GM-4: Issue resolved.

DM-1: \m, \j clarifications. Agreed on change; implementation of the
agreement (text in section 16.2) to be reviewed.

GM-7,8: Clarifications and change to behavior of Shift and Loop Data;
no issues raised about text in section 17.6.

... Start of discussion of the questions raised by Jim:

Leading-keyword discussion. the specific sections in question are the
statements in the Variables block and the SignalVariables declarations
in SignalGroups. Discussed the presence of the SignalVariables under
SignalGroups and need for domain-restrictions; identified that the
Variables block (now separate) should follow similar
domain-restrictions and support domain-referencing akin to
SignalGroups. As a result of this discussion, proposed the following
changes:
1. pull SignalVariables from SignalGroups and place in Variables
block.
2. put type of each declaration first in statement.
3. Add a Variables reference statement to the PatternBurst and Timing
blocks to support domain-named referencing.
[AI3] to Tony.

The remaining two issues from Jim to be discussed in the next meeting;
[AI4] and [AI5] for next meeting
 

Next Meeting
same time, in two weeks; Aug 1, 2002.

Meeting was ajourned at 11:30 PDT.

Greg proposed [AI6] in this minutes for consideration by Tony: move
the Variables clause in the document AFTER all existing-clause
modifications sections. New clauses have been placed at the end of
this document (Environment and previously the BistStructures block),
to reduce confusion by readers over what is a modified-clause section
and what is a totally-new-clause section.

AIs

[AI1]: Tony email reflector about BistStructure removal from .1.
[AI2]: Tony Lockstep example pg 32 change to ParallelPatList LockStep
[AI3]: Tony put SignalVars into Signals, change statements to
keyword-first and add domain-reference statements to burst and
timing.
[AI4]: for discussion next meeting: unclear definition of expression types
[AI5]: for discussion next meeting: should the integer-list construct be
generalized?
[AI6]: Tony to consider moving Variables clause AFTER the
modification-to-existing-clauses clauses.