OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

office-formula message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: "Large" function set


We now have a working "small" function set, and nobody's
complained about the "medium" set.
I am now working to flesh out the "large" function set.
(the old "level 3" set).  I propose as a starting point we use
the "union of the Excel and OpenOffice.org" function set
as the "large" set... any objection to that as the starting point?

Why the union? Because that would mean that any implementation
of the "large" function set could handle the functions of any formula saved
by the current version of either Excel or OpenOffice.org 2
(modulo some fine print that will affect few users).
I think that is very important for migration - we need to
give people a "safe" place to migrate to.

These two programs have very similar sets of functions;
adding OpenOffice.org only adds 25 new functions + 10 "compatibility"
functions (which we'llneed to examine).  And if we can handle
either of them, in practice we can handle
nearly all spreadsheets from anyone; most documents
emphasize the common, not the exotic.

To answer the obvious question: Why not just "union" in
Gnumeric, or Corel Quattro Pro?  The answer is that those
two implementations have HUGE libraries, MUCH larger
than Excel or OpenOffice.org.  If there are a few "gotta have"
functions in them that would enable transition from them too,
then I think we should entertain them,
but not hundreds at the "large" level.
I do think we should add GAMMA, because FACT cannot
portably do the job and it's not exactly easy to simulate
with the existing functions.  Other plausible functions:

Lots of research into the functions supported by various
implementations is here:
  http://www.openformula.org/Research_Data
For example, here's the complete list of all OpenOffice.org
functions that are NOT in Excel at all (the _ADD functions
indicate a semantic variance):
ACOT ACOTH ARABIC B BASE
COMBINA COT COTH CUMIPMT_ADD CUMPRINC_ADD CURRENT
DAYS DDE DECIMAL DURATION_ADD
EFFECTIVE EFFECT_ADD ERRORTYPE
FORMULA GAUSS GCD_ADD
ISEVEN_ADD ISFORMULA ISODD_ADD
LCM_ADD MUNIT NOMINAL_ADD
PERMUTATIONA PHI PRICEDISC
RRI SHEET SHEETS STYLE WEEKNUM_ADD

--- David A. Wheeler 


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]