Summary 2010-02-23 OpenFormula teleconference

["David A. Wheeler" <dwheeler@dwheeler.com>]

Summary 2010-02-23

DON'T FORGET TO VOTE!!

(As always, please reply-all with corrections.)

Attendees:
David A. Wheeler
Eike Rathke
Andreas Guelzow
Eric Patterson
Michael Brauer
Patrick Durusau
Rob Weir
Dennis Hamilton


Discussion:

============

TOPIC: Voting on the CD.

The current openformula draft, cleaned up to remove comments, is up for
a vote as a committee draft (CD):
http://www.oasis-open.org/apps/org/workgroup/office/ballot.php?id=1841
From there, look at the "Referenced items", click on "Details", and select
and link on revision #2 (the latest one).

Rob Weir: Every document has two URLs, one that requires authentication and the
other does not, so some URLs won't work for everyone.

Wheeler: In the future, every change (even for cleaning) should have a different
filename, so that we can tell what we have on our local systems.

Dennis H.: The ballot (link) is fine, so TC members can see it fine, and once it's
approved others will be able to see it.

Dennis: A CD is a more durable thing that lets people have a "more stable" version
to look at, even though more changes will come.

Rob: A CD is a way to break deliverables up & be more discete.

Dennis: Helps with outsider (review).  It also lets others know that a milestone
has been achieved.

Wheeler: Please vote, and if you vote no, please say why.  Indeed, feel free to
comment no matter what.

Dennis: For *this* CD, this is the first shot at seeing what it looks like
WITHOUT all the comments, and that's a big deal.

Dennis: You can always change your vote, so vote now so you won't forget.

Rob & David have already voted.

Wheeler: PLEASE DON'T FORGET TO VOTE.

=====
TOPIC: Process for catching "proposed but not applied".

(JIRA failed at OASIS at this point.)

Patrick D.: I posted a search to show "proposed but not applied".
Does that catch them all?  Do we have a consensus?

Eike: Many of those comments have no marking for 1.2 or anything else.
I will be updating them to 1.2 so that they will be caught,
will do that today.

Eike: One trouble is that once something is marked "Resolved", it can't be
unmarked as "Resolved" (even if it's "open", it stays "resolved").
Perhaps the solution is to search for "Resolved and not open".

Eric: If Patrick can't apply, then re-assign back.

Eike: I don't like assigning back and forth.
The workflow is that we assigned to particular people.

Rob: Can we get "marked as resolved, but blank proposal field"?

Patrick: That probably wouldn't help.
Sometimes there's a clear resolution, but in some cases it's not
clear if we have a common resolution or not.

Wheeler: Rob, can you take a look at Patrick's query?
Rob: Sure.


======
TOPIC: Unassigned JIRA issues for CD2.

At this point, JIRA came back up.

We looked at the dashboard ("OASIS Technical Committees Issue Tracker"):
http://tools.oasis-open.org/issues/secure/Dashboard.jspa
and assigned anything in "Unresolved OpenFormula ODF 1.2 part2 CD2",
focusing primarily on "Major" (and then on "Minor").
We also discussed issues with them, to help the person who took on the assignment.

* OFFICE-2345 (Error) - Dennis will take up
* OFFICE-2347 (Definition of array, conformance clause)
  Note: The original text permitted non-rectangular arrays,
  but no one does that.  Arrays with some empty cells must be okay;
  need to make sure that's clear.  Dennis will take up.
* OFFICE-2372 - we have a resolution.  Assign to Patrick.
* OFFICE-2374 - It's NOT any subpart.
  Wheeler: It has to be at the left, not "any part", right?
  Eike: Any subpart.
  Eike: Will pick up.
* OFFICE-2379: Dennis will pick up.
* OFFICE-2381: Eric P. will pick up; Dennis and Eike will help.
* OFFICE-2385: Dennis will pick up.
  Equality is *hard*, unless it is of integers where all calculations
  are within range of exact calculation.  Different implementations handle
  "close" values differently.

Note: Spreadsheet evaluators may set a setting where
"the cell value displayed is what will be used for future values".
Wheeler and Eike believed this was in part 1, but Eric wanted to double-check
to make sure the setting is somewhere.  Eric checked, found that it's there,
it's calculation setting "table:precision-as-shown".  So there was
no need to open a new JIRA issue on this.

* OFFICE-2399: Dennis will pick up.
  Rob: We don't say anything about 2/3; why bother with this?
  Wheeler: Implementations can implement using exact fractions, and obviously
  they CANNOT do "nearest representatable Number type value" for PI() in that case.
  Rob: I suggest looking at what C does.

* OFFICE-2452: Assigned to Patrick.


--- David A. Wheeler