[prev in list] [next in list] [prev in thread] [next in thread] 

List:       koffice-devel
Subject:    Re: KSpread formulas
From:       John Dailey <dailey () vt ! edu>
Date:       2002-05-21 17:25:40
[Download RAW message or body]

I agree with everything here, the question is what to prioritize.  How much 
can we accomplish before early July?  In my mind, the most important thing to 
do is to separate out the KSpreadValue class (or adjust the KSValue class for 
this purpose).  Right now there are several bugs in the system where there 
are rounding errors, and I would like to see us use something like the gnu MP 
library.

I had already started to separate the value out of the cell class, but stopped 
because I'm not sure how we want to handle numeric data.  Do we want to keep 
using double precision floats?  Or try to use some kind of high precision 
library?  What opinions do people have?

-John

> (1) making a cell value stored in KSpreadValue class. KSpreadCell is quite
> a mess with value handling and so on. and we don't need integer, all
> number are just numbers (should be double), including date/time. this
> value can holds error from evaluating the expression, incl. error code and
> optionally message. when we move to rich-text (KoTextParag), it should be
> easy to extend this class.
>
> (2) define 'format', i.e how the cell value should be displayed, e.g
> precision, currency, and so on. should be put in class KSpreadFormat. and
> this is likely to be shared among. cells will "apply" a prefined/custom
> format. improving cell formatting with Excel would be possible. OTOH it
> also reduces KSpreadCell again.
>
> (3) define 'layout', which somehow already there (KSpreadLayout). layout
> defines apperance of the cell, e.g border, color, font, etc. think of
> value vs format+layout like document vs view.
>
> (4) if we have (2) and (3), it should be difficult to have 'style'.
>
> (5) define 'expression' (class KSpreadExpression) will be used to parse
> and evaluate formula/expression.  again, it can be shared among cells for
> good (copy formula from one cell to another(s) and you only have one
> instance of KSpreadExpression). for starting, KSpreadExpression might be
> simply a wrapper/adapter for KSInterpreter.  when we move on, it might be
> interesting to parse expression once and stored in Polish notation (ala
> Lotus 1-2-3), then with a simple stack-machine, executing expression
> should be faster. add GNU MP library and KSpread becomes the most precise
> spreadsheet ever :-) its lexical analyzer is also useful to create
> color-syntax-highlighted formula editor.
>
> Ideas ? Comments ? Flames ?
>
> Regards,
>
> Ariya
>
>
>
> _______________________________________________
> koffice-devel mailing list
> koffice-devel@mail.kde.org
> http://mail.kde.org/mailman/listinfo/koffice-devel

_______________________________________________
koffice-devel mailing list
koffice-devel@mail.kde.org
http://mail.kde.org/mailman/listinfo/koffice-devel
[prev in list] [next in list] [prev in thread] [next in thread]