Summary 2010-05-04 of OpenFormula meeting

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

Summary 2010-05-04 of OpenFormula meeting

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

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

(Michael B. won't be joining us.)

Note: For our current status, see the dashboard:
http://tools.oasis-open.org/issues/secure/Dashboard.jspa?selectPageId=10056


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


Topic: Summary last week (from Rob)

Last week, we "untargeted" a number of comments;
some of them might become "ODF-Next".
A few were identified as "must-fix" for CD3.

Rob: I found that about 10 issues were assigned to me, quickly resolved.

Rob: We should discuss the 2 open issues
"Unresolved OpenFormula ODF 1.2 part2 CD3"
Which are items 2210 and 2206

Topic: 2210 (Calculation Settings)
http://tools.oasis-open.org/issues/browse/OFFICE-2210

This connects part 1 and part 2.

Some things are implementation-defined, others are provided by part 1.

Rob discussed various options; I had trouble keeping up writing them down.

Rob: We could add introspection functions.
Wheeler: Too late to do that, and unlikely to be used anyway.
Dennis: I favor "option 2" - make the ambient properties host-defined,
 then let part 1 decide which ones are implementation-defined or not.

Rob: So please take a look at this.

Eric P: First, it sounds like we're talking about different things:
1. How part 1 and part 2 are connected
2. Introspection/calculation settings.
I think we're too late for #2, but we need #1.

Rob: I agree.  So what's the minimum number of things that must be
defined for part 1?

Dennis: It's crucial.
There are some bidirectional issues?

Rob: Did we add anything for leap-year?
Wheeler: No.  Epoch is there.

The current material is in section 2.4; take a look at that.

Eric: Still, the calculation settings are at the document level.

Dennis: Key issue - a formula is not always inside a cell.




Topic: 2226 (draw prefix [N 1309])
http://tools.oasis-open.org/issues/browse/OFFICE-2226


Rob: It really *does* matter if it's units of a 100, which is bizarre.

Wheeler: Eric, can you provide more light on this?
Eric: I'm not sure what else I can provide.
Wheeler: Could you confirm that the linked-to formula is what Excel does? Is it correct?
Eric: Yes.  I'll look for another link as well.
Andreas: This is supposed to be the inverse of ODDFPRICE; could plug in to double-check.
Rob: I could try that!
Eric: You can assign that one to me.

Eike: Can we resolve that within 1-2 days, for CD03?
Eric: I'll let you know by tomorrow.

Rob: It wouldn't kill us if not, but it'd be nice if it was.

Dennis: Where's CD02?
Rob: Michael was assigned to do that.



Topic: 1874
http://tools.oasis-open.org/issues/browse/OFFICE-1874

We agree that it's not well-defined.
Is it correct for negative numbers?

Eric: I'll work on a proposal.
Eike: Also check ROUNDDOWN.
Dennis: Need to clarify what happens with negative numbers.
Wheeler: Eike, please post the negative number results.
Need 4 values: +/- for both argument and Digits.
Excel, Gnumeric, rounds away from zero.
(OOo probably does too.)

Thus, the specification is probably WRONG and needs to be fixed -
applications appear to round away from zero instead.

Eric is still the assignee on this one.

?: David, you still have 2 issues.
Wheeler: They're minor issues, I'll try to get to them today, but
 even they aren't addressed it is not a big deal.

Rob: If anyone makes changes now, they're late, but we can probably
handle changes today and tomorrow... but not after that.

Rob: At some point we need to treat the parts as a unit.

Eric: I'm having trouble opening the latest file from Patrick D.
Eike: I'm not having trouble.
Eric: I've tried it 3 times, hard to open.
Eike: That was because N() and T() were indexed, but it caused
 *every* n and *every* t (including the one in "the") to be indexed.
Patrick: We're going to index it in a different way, eliminating that.



--- David A. Wheeler