Re: [office-formula] Groups - OpenDocument-formula-201001018(OpenDocument-formula-20100118.odt) uploaded

[Eike Rathke <erack@sun.com>]

Hi Patrick,

On Monday, 2010-01-18 23:17:23 +0000, Patrick Durusau wrote:

> The latest draft of OpenFormula is attached. Heavy editing in chapters 1-4,
> with more to come. Comments and suggestions welcome. 

You entirely removed sections 2.1 to 2.6, that is unacceptable. 


In section "3 Typers" you added the TODO
|| General question: For the basic types, is there a reason why we don't
|| simply cite the XML Schema datatypes? I understand that reference and
|| reference list and others we must define but what of the others? 

We already had that disussion a few months ago, why bring it up again?
The outcome was that XML datatypes are not identical, for example the
lexical representation of double in XML allows the special values INF,
-INF and NaN.


4.2 Basic Expressions
|| Question: Why is this here? Important but that isn't my question. First,
|| the syntax has yet to say anything about “literal” numbers or strings,
|| whatever those are. Second, we haven't discussed logical constants so
|| can't have said anything about them. 
| The syntax defines rules for literal numbers and literal strings, but it
| does not define a rule for literal logical constants. For logical
| constants, TRUE() and FALSE() are used.

I think we can remove that paragraph.


4.3 Constant Numbers
|| Sorry, “constant” numbers? Do you mean decimal numbers? And floating
|| point numbers? Or are you trying to lump the two together?

Yes. A literal number constant.

|| Suggest losing the “constant” numbers and have decimal and floating
|| point numbers. All that requires is citing the W3C. 

No. See above.

|| Err, did we mean to call this “StandardNumber?” This is the only
|| reference to the “standardNumber” format. Or the Number format. This
|| looks like a mistake. 

Take a look at the EBNF expression. StandardNumber is a decimal or
floating point number, we may call that another name if that helps.


4.4 Constant Strings
|| I assume we mean “strings.?” I don't know what “constant” string means. 

A literal string constant, e.g. "this is a string constant", it is
constant because it is not the result of an evaluation.


4.5 Operators
|| And we define these under 5.4. Let's do one or the other. I suggest that
|| we define associativity as part of the syntax and define the operators
|| under 5.4. 

Er.. we did that, didn't we?


4.6 Functions and Function Parameters
| Function names are case-insensitive.. [... deleted that implementations
| should write function names all uppercase ...]
|| I deleted the statement about implementations as we only have evaluators
|| and expressions. 

And how do we express that evaluators should write function names
uppercase?


4.7 Nonstandard Function Names
You deleted the entire section, that is not acceptable. It defines how
functions that are not defined in the standard are to be named when
written.


4.8 References
You deleted the one-sentence paragraph about whole row and column
references. Why?


5.2 Common Template for Functions and Operators
|| I am not sure what the “Hear {..} means? That the parameters are optional? 

That a list of parameters is optional.

| Components surrounded by [ ... ] are optional, as is any parameter with
| the keyword optional. 
|| Is there more than one way to have an optional parameter? 

{...} is an optional list of parameters, [...] is a single optional
parameter or a combination of optional parameters. I think the "as is
any parameter with the keyword optional" can be removed, IMHO we don't
use that anymore.

| A parameter followed by the = symbol has the default value given after
| equal sign; any such parameter is always optional (even if it is not
| surrounded by [...]). 
|| So there are two or three ways to have optional parameters?

Actually not. Optional parameters _are_ surrounded by [] brackets. That
is to be reformulated as

An optional parameter followed by the = symbol has the default value
given after equal sign.


  Eike

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

PGP signature