XML accepts most versions of line-break on input... but it is generally
considered polite for serializers to write out whatever version of newline
is considered native for the platform they're running on, to make the
output easy to process further/examine using tools native to that platform.
"Same output on all platforms" depends on what you mean by "same" -- from
the user's point of view, native line-break is closer to being "the same"
than the same bytes would be, since it's the same semantics.
Sothe right default mapping is probably to keep it as it is now... which I
think means Java's definition of the system property line.separator.
But I can see an argument for letting folks override that default if they
know the next stage of processing doesn't care, or explicitly wants a
particular version of line-break. I'm not sure whether we should tell them
to set that property or if we'd need to get our own code involved.
I do have one hesitation: the special-casing to deal with recognizing and
translating the value might cost us some cycles, so this may not actually
be a performance gain... and if it isn't, then we have to think about what
it's worth to us as a convenience.
______________________________________
Joe Kesselman, IBM Next-Generation Web Technologies: XML, XSL and more.
"The world changed profoundly and unpredictably the day Tim Berners Lee
got bitten by a radioactive spider." -- Rafe Culpin, in r.m.filk
