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: Finish in July, Statistical function names, YEARFRAC


As was discussed in an earlier OpenDocument conference call, we need to finish the OpenFormula specification in July (this month).  Frankly, at this point we're in surprisingly good shape.  We've removed the stuff that needed to be removed, and it looks like in many cases the test cases influenced text writing so much that we're not as bad off as I thought.  Nevertheless, we could very much use another volunteer or two to see what text needs updating.

I'm still not entirely happy with our set of function names for the statistical functions (e.g., LEGACY.*, esp. when there ISN'T a non-legacy function).  Can anyone suggest some alternative names?  Granted, quibbling over names is a good sign, but this has been raised several times and we need to wrap that up.  And don't take that as a zap on the statistical work; my thanks to all (including Andreas J Guelzow) who have done such great work on that section.

FYI: IBM has created a massive set of test cases and expected results for YEARFRAC, and submitted them (thanks!!).  I have already begun writing a program (in Python) that reads that datafile in and calculates the actual value according to various documented algorithms.  Once we've confirmed that the expected values match the actual values of given algorithms, we can document the algorithm.  We're doing this because there are several documented algorithms, but nowhere is it documented WHICH algorithm is WHICH.  I'm using Python because it's a particularly easy-to-read language - the code can be more-or-less cut & pasted into the specification as executable pseudocode.  (Is it pseudocode if you can run it?).  This process should eliminate the "YEARFRAC/Basis is not fully defined" comments.  If anyone wants to do this Python implementation instead of me, I'd love to release it, but no one else jumped up & I want to make sure we resolve this issue (the incomplete definition of Basis/YEARFRAC is the known outstanding technical issue.  Granted, our information is already better and more detailed than anyone else, but we have a much higher standard for ourselves).

--- David A. Wheeler


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