Re: before kawa-1.8

[Per Bothner <per at bothner dot com>]

Hoehle, Joerg-Cyril wrote:

> One thing that comes to my mind as a possible improvement is somewhat better integration of Java functions returning java.math.BigInteger.
> All Java functions which manipulate large numbers operate on these, and I've been bitten several times by things which display as a large number in Kawa but are not Kawa objects (i.e. Scheme + or -  obviously don't work on them).
> But I don't know whether it would be a good idea to automatically transform a java.math.BigInteger into a gnu.math.BigInteger when entering Scheme. They may have different semantics (shared or not, immutable or not byte[]).
>
> I vaguely remember mentioning a biginteger problem with JDBC last year, but I don't have access to an Oracle DB anymore for testing.

Integrating BigInteger into Kawa arithmetic is not a change I want to 
work on right now.  Converting BigInteger to IntNum is inefficient, as 
is arithmethic that can add a BigInteger and a IntNum.  The cleanest 
option is to replace IntNum by BigInteger, but it would probably be a 
performance regression:  Now we use IntNum for *all* Scheme integers, 
and we have optimization to do this torerably well.  Also, now all 
Scheme numbers inherit from gnu.math.Numeric, which has builtin support 
for mix-type arithmetic.  java.math and java.lang don't.

So it's tricky.  I have me forced to deal with it - but not for 1.8.

> > One of many function I just haven't gotten around to implementing. 
>
> Oh, I had understood the document in the sense that dm:node-kind is the standard, but that qexo provides fn:node-kind instead. I did not understand that the standard provides two ways of expressing the same thing (does it)?

The document is not a Qexo-specific document.  The difference is the the 
dm: functions are *conceptual* (expository) functions, which only exist 
in the data model, but which you can use in a program, while the fn: 
functions are libary functions that you can use in a program.  (Assuming 
they have been implemented, which is not always the case in Kawa.)
--
	--Per Bothner
per@bothner.com   http://per.bothner.com/