Summary of 2009-12-08 teleconference

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

Here's my summary of the 2009-12-08 teleconference.

Present: Eric, Eike, Rob Weir, Patrick D., Dennis, myself.

David A. Wheeler noted that he'd *like* to resolve the rest of the comments by the end of the month.  This may not be realistic, but it's worth trying.  At some point, if no one's interested in working an issue, to Wheeler that suggests that it's not an issue worth working on.

There was a brief discussion on EUROCONVERT.  Every time a country enters the EU, a new fixed conversion should be added.  Rob Weir noted that this is essentially a registrar-type function, and so we should be pointing to an authoritative registrar.  That way, when a country enters the EU, implementers can immediately add the currency instead of waiting for an OpenFormula update.  Eike will find out what the "incantation language" is to refer to the registrar.

There was then a *long* discussion about "portable document" and related terms.  To be honest, I don't think I can summarize this well at all, there was a LOT of discussion.  Here are SOME of the discussion items (I'm certain I omitted some key points):
* The problem with the term "portable document" is that the spec is really only about formulas, not about everything else.
* The problem with the term "portable formula" is that a formula by itself isn't necessarily portable, even if it's syntactically valid.  E.G., "=SUM([.A1:.A2]" is not portable if A1 contains a logical value (e.g., "=1=1"), because SUM skips non-numeric values in ranges, and logical values may be (but are not necessarily) numeric.  In addition, formulas like "=ORG.OPENOFFICE.EASTERSUNDAY(1999)" conform with the spec, but aren't portable.
* The problem with just removing the term "portable document" is that users *do* want to know if their file (or whatever) will work on any conforming application.  Thus, although we could switch to a simple non-normative appendix with suggestions for portability, that will leave users without the ability to specify it as a requirement.  (And it won't be easy to re-add back in; the term "portable document" is all over the specification.)
* Thus, perhaps the way forward is to use another term, and flesh out its definition at the front of the specification.  What the new term should be isn't clear (usage? context? hosting?).  It should include not just the formula, but the underlying environment (with table contents, including the cell types).
We ran out of time without resolving how to improve things, so we agreed to continue the discussion on the mailing list.

My apologies for having trouble summarizing this one.  Please "reply" with important clarifications.... or even better, please "reply" with proposed improvements to the current "portable document" scheme.

--- David A. Wheeler