xml-dev - Re: Possible changes for XML 2nd Edition

["David Brownell" <david-b@pacbell.net>]

> Earlier today David Brownell wrote:

Make that yesterday ... ;-)

> > (*) Sun's not the vendor to blame for those speed issues;
> >     it's easy to implement a java.io.Reader class that
> >     does the decodings with  very low overhead, but the folk
> >     who did that subsysteim in JDK 1.1 didn't choose such
> >     a natural approach.
> 
> Just to clarify, are you saying that Sun outsourced the implementation
> of java.io.Reader?  To whom?  Or are you instead saying that it's
> inappropriate to blame Sun even if they did the implementation
> themselves, because it would be so easy to override the behavior of the
> class?

Without Naming Names, another company provided the
decoder framework, with performance impact that was
quite visible the last several times I checked ...
though perhaps HotSpot (and hence JDK 1.3) reduces
it by some of its magic inlining.

Readers delegate to the decoder framework in all
cases; my technical comment is that there are costs
therein, in memory (extra objects) as well as in
method invocations (with a framework API that's perhaps
suboptimal).

Re blame, one can't on one hand insist Sun open up
to more partners, and on the other hand blame them
for the work of those partners ... that's all I meant
to say.

- Dave



***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/
***************************************************************************