ODFF: list of suggestions

["David King" <openoffice@drking.plus.com>]

ODFF draft spec: 

I contributed some suggestions/bug reports here at the end of 
last year. I know it's unlikely that anyone will get round to 
doing much about it (my previous suggestions have still not made 
it onto the to-do list as far as I can tell), but nevertheless 
I've collected these suggestions/bugs from reading the Dec07 
draft:


Function descriptions:

LEN
LEN(A1) where A1 contains a number - it's not specified whether 
to return the count of the displayed digits or internal storage 
or what. Calc converts to a number with max decimals (not the 
default format), and counts that. 

PERMUT
The formula uses P<sub>n,k</sub> which is not the common 
convention (actually incorrect I think).

PERMUTATIONA 
The formula (not given) is simply n^k - is this function really 
needed or is it bloat? Excel doesn't have it I think?

LEFT,  RIGHT
The 'see also' list should include MID

PERCENTRANK 
"as a percentage of the data set" - this phrase is surely 
meaningless - needs a rigorous description.

GROWTH
misleading to call the parameter 'allowOffset' as this is not an 
offset but a multiplier.

HARMEAN 
there is no N in the formula

CONVERT
Care must be taken with prefixes for areas and volumes. 'km2' 
means '[kilometres] squared'  NOT '1000 [metres squared]' - this 
is potentially misleading, and the draft spec does not say

LINEST 
needs a rewrite into clear English - it's actually wrong about 
multiple x values as well

MUNIT
"Summary: Creates a unity matrix of a specified dimension N" - 
should read 'unit' not unity

INDEX page 243
*"this is essentially a 2D version of CHOOSE" - no it's not - 
CHOOSE cannot accept a range parameter.
*"which looks as if Row is taken as the column offset into the 
vector" - needs to be put more clearly.
*"If DataSource is a one-dimensional row vector" - this needs 
rewording to be clear about whether or not this applies to 
multiranges. Clearly not in this phrase, yet area number may 
still be specified?

STDEVA
"Estimates the standard deviation" - no, it calculates the 
standard deviation. The standard deviation *is* an estimate (of 
spread).

COUNTA, AVERAGEA, xxxA....
Reword these xxxA functions with consistency eg treatment of text

MATCH page 246
*Doesn't specify that array must be single row or column
*Constraints not specific enough - eg currently allows 
Matchtype=0.5

RANK  page 369
*Typo: 'Order' has no ascending option (both are descending)
*Should say that the number to be ranked must exist in numberlist 
(otherwise would be sensible too)

FIND 
Needs examples with more than one character in 'find' parameter, 
to show how it matches.

SEARCH 
The examples should make it clear when regexp / wildcards are on. 
Also show example to match literal question mark / dot etc.? 

SHEETS
* =SHEETS([.B7]) does return 1 but I misunderstood - explain that 
B7 is on a single sheet? Contrast with sheet.b7:sheet3.c9?
* "referenc" -> 'reference' (typo)

FIXED
"If the optional parameter OmitSeparators (which defaults to 
False) is True, then decimal separators are not included in the 
resulting string." Couldn't this be re-worded? - it's rather open 
to mistakes, and shouldn't it be 'thousands', not 'decimals'? My 
offering: "If the optional parameter OmitSeparators is True, then 
thousands separators are omitted from the resulting string. 
Thousands separators are included in the absence of this 
parameter."

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

Other issues:

'DateSequence' is undefined.

'NumberSequence' is not properly defined - it does not specify 
what happens in an array, nor what happens for a rectangular 
range (more than one row / column). There is no section actually 
defining it - merely a 'Conversion to...' section.

page 41
* reword 1.2 and 1.3 - both are 'References'. 
* Actually the whole section could be better worded (what's a 
'transpose approach'?)
*  =A1:C4 in a cell does not use intersection in 'non-array' 
mode. This could be better defined - it could be more clearly 
said that intersection is only defined for function arguments.

page 42
"=MID("abcd";{1;2};{1;2;3})	=> {"a";"bc";#N/A}"
no - returns {..;..;#VALUE!}, because MID("abcd";#N/A;3) returns 
#VALUE!

page 427
"Year 1683"  ->1583 (typo)

page 428 
* "Number of either side" - > on either side (typo)
* "= i = COMPLEX(1;0)" - should be COMPLEX(0;1)


David King