Summary 2010-08-17 of OpenFormula meeting

[Eike Rathke <eike.rathke@oracle.com>]

Hi,

I'm posting this on behalf of David, posting to the list isn't working for him.

  Eike

----- Forwarded message from "David A. Wheeler" <dwheeler@dwheeler.com> -----

Date: Tue, 17 Aug 2010 12:09:48 -0400 (EDT)


Summary 2010-08-17 of OpenFormula meeting

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

Attendees:
David A. Wheeler
Andreas G.
Patrick D.
Eric Patterson
Dennis Hamilton
Eike Rathke


Not present: Rob Weir.

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



TOPICS:

NEEDS-DISCUSSION & Unassigned

* OFFICE-2718

Dennis: We don't need to assign it here.
It's part 1, already been discussed in the TC.  It's really up to the TC for who to assigned.

* OFFICE-3340
http://tools.oasis-open.org/issues/browse/OFFICE-3340

Eric P. will pick this up.

Dennis: There are a number of places in part 1 where this isn't in a sheet;
they may need different description.

Wheeler: Dennis - please add comments to this, since it relates to your work.


* OFFICE-3345 (IMPOWER)

OOo takes Number as second argument (at first we thought it only took Integer -
  documentation is wrong, implementation supports Number).
Excel takes Number as second argument.
Gnumeric takes Complex as second argument.

FYI, C++ accepts complex as second argument for pow.

Changing it gives backward compatibility problems to implementations that don't support it.

"It's better to have implementations do this; it's hard to emulate if you need Complex, Complex".

Wheeler: Like "at least number in 2nd argument, implementations MAY implement Complex in second arg."
Eric: Leave as-is (Number), and permit Complex.
Eike: Do we really intend Number, and not Integer?

Hamilton: Could make it "Complex" and make that rock-solid.

Andreas: We'd need to define it for Complex, *if* you support that.

Hamilton: Need to make sure that negative first argument is handled correctly.

Eike: Need to state that it returns the principle value, in the case where multiple values can result.

Wheeler: Isn't that the result with the smallest non-negative angle?

Andreas: Yes.

Wheeler: So, Number is required as the second argument.
Complex *may* be supported as the second argument.
It returns the principle result, which is the result with the smallest non-negative angle.



* OFFICE-3342

http://tools.oasis-open.org/issues/browse/OFFICE-3342

Dennis: Don't do anything yet.  This is being discussed in TC; once resolved in TC,
we should match.

Every IRI is a URI, and every URL is a URI.
IRI is an interpretation of % encoding that limits % encoding.

So we'll just wait.


* OFFICE-3042
http://tools.oasis-open.org/issues/browse/OFFICE-3042

* Need to change "Should" to "Shall".

* Otherwise, all agree with the text except new 5.7.1.

Eric: Need this to deal with compatibility backward to older versions.

* Big discussion about 5.7.1 (Future OpenFormula Function names).
Should we add a prefix?  When?  What should that prefix be?

Wheeler: The original idea was to use namespaces to identify version numbers.

?: But the namespaces so far don't change when the version changes.

Wheeler: It'd make more sense to add the prefix when writing only in cases
where there was originally no prefix *and* it's not a known function
(so it *must* be a function "from the future").

* This needs to make sense for 1.2.  We are making requirements on the
implementors of the *current* spec, not just of future specs.


--- David A. Wheeler 

----- End forwarded message -----

-- 
Automatic string conversions considered dangerous. They are the GOTO statements
of spreadsheets.  --Robert Weir on the OpenDocument formula subcommittee's list.

PGP signature