Kohsuke Kawaguchi wrote:
> Paul Sandoz wrote:
>
>> - We can have other methods for int[], float[] etc. All optimized data
>> types will be arrays even if only one value is present, will this be
>> an issue?
>
>
> Do we expect a lot of FI documents to have more than one ints, floats,
> and etc?
>
> Right now, when we tokenize CharSequence, we always treat that as a
> string, so any advantage of having typed information is lost right there.
>
> If we expect FI users to take advantage of compact int/float/... list,
> maybe I should do something smart there.
>
Scientific data will usually contain lots of arrrays of ints/floats and
this is one use case for FI.
That is why i thought that additional methods on the XmlVisitor would be
more efficient it is possible to match the parser and XmlVisitor
requirements before the parser reports an 'optimized' event.
Also attributes can also contain optimized data.
Paul.
--
| ? + ? = To question
----------------\
Paul Sandoz
x38109
+33-4-76188109